Increasing diversity as an HRMchange strategy

Ellen Ernst KossekMichigan State University East Lansing Michigan USA

Karen S MarkelOakland University Rochester Michigan USA and

Patrick P McHughThe George Washington University Washington DC USA

Keywords Human resource management Diversity Change management

Abstract In order to manage strategic demographic change in economic and labor markets acommon human resource (HR) change strategy is to increase the diversity of the workforcethrough hiring over time This study examined department level consensus and valence regardingan organizational HR strategy to shift demography toward greater diversity in race and sexcomposition over an eight-year period Though the organization had experienced signiregcantchange in organizational demography an increase in the overall representation of white women(36 percent) and minorities (41 percent) over time work group members in units with the greatestchange did not necessarily agree nor hold positive perceptions regarding these HR changes Theresults show that HR strategies that focus on structural change without working to developsupportive group norms and positive climate may be inadequate change strategies

Managing growth in workforce diversity and increasing the representationof women and minorities throughout the organization is a critical strategichuman resource (HR) management issue for most organizations (Thomasand Ely 1996) In order to manage demographic change in economic andlabor markets a common HR change strategy is to increase the diversity ofthe work force through hiring over time Though most HR strategies tomanage diversity are conceived at the regrm level they are often socially andpractically enacted at the work group level of analysis (Larkey 1996) Sinceemployees are typically hired (and expected to assimilate) into departmentsit is important to understand how demographic changes are experienced inthis context Departmental groups are the receptacles of organizationaldispersion of HR strategies to manage diversity Individuals are embeddedin departmental work groups which provide a context shaping the socialmeaning of HR strategies to shift organizational demography Research ismixed on whether HR strategies to increase minority representation results

The Emerald Research Register for this journal is available at The current issue and full text archive of this journal is available at

httpwwwemeraldinsightcomresearchregister httpwwwemeraldinsightcom0953-4814htm

This research was partially funded by a doctoral assistantship provided by the School of Laborand Industrial Relations at Michigan State University An earlier version of this paper waspresented at the 2002 Society of Industrial Organizational Psychology International MeetingsToronto Canada

JOCM163

328

Received 15 March 2002Revised 26 September2002Accepted 5 December2002

Journal of Organizational ChangeManagementVol 16 No 3 2003pp 328-352

q MCB UP Limited0953-4814DOI 10110809534810310475532

in positive or negative group processes depending on whether one takes asocial contact (Allport 1954) or resource competition view (Blalock 1967)We designed this study to develop and assess climate constructs groundedin these prevailing literature themes consensus the degree to which groupmembers held common perceptions of diversity climate and valence thedirection of those perceptions The study also was designed to explore thewell grounded but mixed theoretical stances in the literature and toconsider how the magnitude of organizational diversity change strategiesmay be diluted at the work group level

The perceived magnitude of change at the group level is important toconsider because organizational demographic shifts are likely to trickle downin a fashion that dilutes saliency and increases variance in the interpretation oforganizational change at the group level of analysis Positive social climatemay not necessarily occur when organizational changes are diluted inenactment within work groups Rationale for the view that group climate maynot necessarily be improved by minor demographic shifts is provided byKanterrsquos (1977) work on tokenism and tipping points within groups She arguedthat depending on proportional representation levels members experienceothers as dominant token tilted or balanced She held that negative socialpsychological processes such as subtle discrimination is minimized whereminority representation reaches a critical mass referred to as tipping pointSuch negative dynamics may occur in groups where there are tokens deregnedas15 percent or less and are not likely to signiregcantly lessen in exaggeratedimpact until groups are tilted where minorities comprise at least 35 percent Ifdemographic shifts are not enough to alter tipping points in speciregc workgroups then increased consensus and valence regarding the social climate fordiversity may not necessarily occur Despite the signiregcance of conductinganalysis of group level reactions to diversity change little research has beenconducted on this issue

1 Research objectivesThis study examined department level consensus and valence regarding anorganizational HR strategy to shift demography toward greater diversity inrace and sex composition over an eight-year period Our regrst research objectivewas to identify and develop new measures of constructs remacrecting a positivesocial construction of change in diversity at the department level Our secondobjective was to examine the question ordfDo work groups with higherdemographic shifts tend to have more positive climatesordm that is higherconsensus and valence toward diversity change for both organizational andgroup level change referents Our third objective was to examine the questionordfDoes increasing the proportion of senior women and minorities at thedepartment level positively relate to group valence and consensusordm

HRM changestrategy

329

11 Developing group level measures of the social construction of change indiversityAddressing our regrst objective we develop theory and measures forunderstanding the social construction of change in diversity at the workgroup level These constructs include work group demographic shift groupclimate (ie consensus and valence) and group and organizational changereferents First it is important to examine structural progress which we callordfwork group demographic shiftordm This is deregned as a work grouprsquos increase inthe proportion of women and racial minorities over time occurring throughchange in member composition If an organization has been successful in hiringstrategies it will increase the number of women and minorities within anorganization as a whole These hiring increases typically create demographicshifts that are cascaded down to the work group level and impact the socialclimate for diversity The study of work group demographic shift raises issuesnot ordinarily addressed in cross-sectional HR management research ondiversity (McGrath et al 1995) Members of two groups with the samecomposition might not necessarily share current perceptions about the climatefor diversity plusmn whether it is a favorable or unfavorable climate The reasons forthis difference may include but not be limited to historical differences inheterogeneity the degree to which diversity is seen as valuable for achievingstrategic goals and the quality of interpersonal relations and resourceallocation between members of different backgrounds As Kanter (1977)argued relative (rather than absolute) numbers or proportional representationshape and tip minority and majority group dynamics surroundingorganizational change Work group demographic shift remacrects contextualand relative experience with a regrmrsquos overall or absolute demographic hiringpractices

The second set of constructs we use to understand demographic change isordfgroup level climateordm the social construction of the demographic shift Climateor prevailing member beliefs about ordfthe way things are around hereordm isincreasingly being studied less in terms of ordfthe climateordm and more in terms ofordfthe climate for somethingordm plusmn a speciregc referent (Schneider and Reichers 1983)Using cross-sectional data Kossek and Zonia (1993) examined how currentdemographic composition related to individual perceptions of diversity climateWe build on this previous research by measuring climate at the group level toassess shared member perceptions of the effects of a demographic shift over asigniregcant time period Our approach is grounded in the suggestion that byderegnition climate may be best measured as a group phenomenon (Reichers andSchneider 1990) Climate is partly a function of structural aspects of the workcontext (Payne and Pugh 1976) Work group members observe structuralfeatures such as small or large work group demographic shifts interpret themin order to make meaning of organizational change The greater the work

JOCM163

330

group shift the greater the implementation of the change strategy which couldlead to a more positive climate since these members experienced the mostchange

We contend group climate may be operationalized by the constructs ofconsensus increased agreement in member perceptions and valence thedirection of agreement Consensus remacrects what Harrison et al (1998) refer toas deep level similarity in attitudes and values Consensus identireges theextent to which there is commonality among the perceptions of unit members(Brown et al 1996) The issue here is variability within a work group Workgroups with low variability (high consensus) would share commonperceptions regarding the social climate for diversity and similarunderstanding of social reality (Festinger 1954) In high consensus groupswe would anticipate that members would experience higher behavioral andattitudinal predictability regarding how to enact change in response to anorganizationally-driven demographic shift This would shape shared basicassumptions (Schein 1988) regarding appropriate behaviors in response toorganizational change (Forsyth 1990) High attitudinal and behavioralpredictability reduces strain on interpersonal relations within groups (Zander1994) and allows for greater ease of communication and less frictionregarding how to enact change (Bleise and Halverson 1998) resulting in amore similarly construed group climate

In order to have a positive climate members must not only agree on theclimate they must also see the climate as favorable Valence measures thedirection of group membersrsquo perceptions akin to what Lindell and Brandt(2000) identify as climate quality The focus here is on group membersrsquo typicalaverage or median response Following work group demographic shifts thosegroups with high valence would have a positive construction of the changeMembers would be expected to perceive that it is important to achieveorganizational goals through implementing diversity activities In contrast lowvalence groups would have negative climates grounded in perceptions ofdiscrimination and harassment toward minority members as well as less thanenthusiastic unit response to strategic diversity initiatives (James et al 1994Schneider et al 2000)

The last set of constructs we identireged as important for understandingdiversity change are ordfgroup and organizational referents of changeordm Like mostHR organizational change strategies diversity change initiatives are enacted atthe group and organizational levels and therefore should have referents at bothlevels It is critical to assess group perceptions of organizational levelphenomenon as groups can vary in the degree to which they perceiveorganizational strategic initiatives as supporting unit objectives Based on areview of the literature we identireged two organizational referents favorableagreement that top management is committed to diversity and the belief that

HRM changestrategy

331

the employer should promote diversity initiatives as a strategic organizationalgoal Many studies conclude that the success of diversity initiatives iscorrelated with member perceptions that management is highly committed tothe active pursuit of diversity policies as an organizational strategic objective(Cox 1993 Kossek and Lobel 1996 Thomas and Ely 1996)

Since the group level of analysis is more proximal to interpersonal socialoutcomes we developed speciregc group indicators for the social construction ofdiversity change that were distinctive from those at the organizational levelMost individuals are sophisticated enough today to know that it is sociallydesirable to state that managing diversity is an important organizational goalas such contentions have become a truism in the management literatureHowever at the group level membersrsquo attitudes are more likely to remacrect howthey see shifting demography affecting them personally in their immediatework group context The assumptions that a work group is viewed as havingwomen and minorities who mix well with white men and that resources areequitably allocated across groups are key in operationalizing diversityparadigms (Kossek and Zonia 1993 Thomas and Ely 1996) Thus groupreferents include favorable agreement that women and minorities are seen asmixing well with group members and that work group resources are allocatedwith equal access to women and men and whites and non-whites Theseperceptions would suggest that the work group has become more multiculturalin a positive way

12 Do work groups with higher demographic shifts have a more positive groupclimateOur second objective was to address the question ordfDo groups with higherdemographic shifts have higher consensus and valence toward diversitychange for both organizational and group referentsordm Competing argumentscould be made that higher demographic shifts could lead to either negative orpositive climates depending on whether one takes a social contact or intergrouptheoretical perspective Under the social contact view work groupdemographic shift to increase the proportion of women and minoritiesenables higher cross-group interaction and mixing which is necessary toimprove perceptions of intergroup relations and social integration and reduceprejudice (Allport 1954 Triandis et al 1994) Because groups with a higherdemographic shift would have more social contact there would be higherconsensus in a positive direction about the change Unfortunately the socialcontact theory may only apply if the demographic shifts are great enough toreduce negative tokenism dynamics (Kanter 1977)

More likely in groups where the demographic shifts are not large enough tocreate a critical mass and tip the demographics to reduce tokenism thealternative perspective grounded in intergroup resource competition theory

JOCM163

332

prevails Here the argument is made that the diversity change is not largeenough to lead to greater consensus and valence The redistribution ofintergroup power through resource reallocation to reduce inequality (Alderferand Smith 1982) is likely to be experienced negatively by both majority andminority group members In groups where demographic shifts are dilutedminority members may still experience negative psychological processes anddiscrimination Majority members may see the slight demographic shifts aschanging the status quo to negatively impact them as they would see this aslosing power

H1 Work groups with higher demographic shifts may not necessarily havemore positive climates that is higher consensus and valence towarddiversity change for organizational and group level change referents

13 Is there a positive senior women and minority proportional effect on changeWe turn to our third objective that investigates ordfDoes increasing theproportion of senior women and minorities at the department level positivelyrelate to favorable group valence and high consensusordm Growing evidencesuggests that the interaction of demography with hierarchical level is critical tounderstanding organizational change processes The manner in which HRstrategies to promote demographic change shifts group demography acrossstatus levels within groups can lead to differing group climates Thus thecurrent race and gender composition of two groups can be identical in collectiverace and sex composition but have markedly different social reactions todynamics due to the saliency of membersrsquo multiple identities in relation to otherkey employee backgrounds such as hierarchical level

By examining the interaction of level (ie status) with race and sexdemography we are able to investigate the theory that it is not just the changein overall proportional demographic representation that effects climate but thecombination of demographic and non-demographic attributes that mightinmacruence the group social construction of diversity strategies in amultiplicative manner (Ely 1995) Demographic variables such as race andgender usually assume greater importance when associated with thedifferences in status (Tsui et al 1992) For example higher status membersmay be more likely to withdraw when groups are regrst integrated by lowerstatus members (Harrison et al 1998) Research on how status relates to groupproportional representation suggests that HR strategies to simply increase theoverall proportion of a minority group do not necessarily result in improvedorganizational effectiveness and increased inclusion of under-representedgroups (Ely 1994 1995) In order to have favorable climate the minority groupalso must be well represented in senior positions

Yet the dilution of organizational strategies across group hierarchiesagain may hamper the achievement of positive change in the status of

HRM changestrategy

333

groups When organizational diversity objectives to increase the upwardmobility of women and minorities are regrst enacted across work groupsstudies show that most organizations are likely to increase the number ofsenior women and racial minorities by only a few in each group (Tsui andGutek 1999) Again relying on Kanterrsquos (1977) theory of tipping points aslight gradual increase may not be strong enough to improve climatecompared to work groups that experienced no change or a decrease inhierarchical representation Recent work by Lau and Murnighan (1998)builds on Kanterrsquos (1977) work and suggests that groups have fault lines orways to align demographic characteristics across level gender and raceIncreasing gender and racial diversity through diversity hiring strategiescan alter the demographic strength and alignment or fault lines of internalsubgroups (eg senior and junior management tenured and junior faculty) Ifthese alignments are weak and work groups do not experienceordfearthquakesordm like the notion of a shift to a critical mass a slight shiftwork group demographic may be even more diluted at the senior levels Thedilution of change may result in senior groups being still largely seen aswhite male having little change impact on diversity climate

Weak demographic shifts create weak group fault lines and consequentlymay do little to change or coalesce the social construction of group climateUnder such conditions intergroup resource competition theories are againlikely to be apt explanations of group processes As a minority group gainsstature and starts to become larger and more of a realistic competitor for scarceresources the majority is more likely to discriminate (Blalock 1967) Gradualproportionate increases in the representation of women and minorities at seniorlevels might lead to greater hostility by majority members because of resourcecontrol issues This would lead to more strained intergroup relations andreduced social interaction among members resulting in greater variation of theclimate for diversity (ie reduced consensus) and more negative perceptions ofclimate

H2 Increasing the proportion of senior women and minorities at thedepartment level may not necessarily positively relate to group valenceand consensus

2 Method21 Organizational settingThe data collection period for this study was designed to mirror the timeperiod of implementation of an organizational change strategy to managediversity at a large public sector university in the USA Over eight yearsthe administration had been actively engaged in HR strategies to fosterorganizational change in diversity These included active recruitment of adiverse workforce and wide dissemination of a document afregrming the

JOCM163

334

universityrsquos commitment to diversity National consulting regrms had beencontracted to provide advice on managing diversity The administrationmade funds readily available (sometimes over the initial salary posting) tohire quality minorities and women even for units in which new tenurestream positions had not been approved Training to promote diversitysensitivity among administrators was initiated Small cash achievementawards were distributed to recognize department activities that promotedmulticulturalism

Using the human information system the researchers collected facultydemographic data by department from records in the HR information system attwo points over an eight-year period to measure change in demography bydepartment We used a stratireged sample of tenure stream faculty with fullselection of the population of women and minorities and random sampling ofwhite males At the end of the eight years a survey was developed to examinethe success of the change efforts The archival data on proportional change indemography by group were used to form the independent variables (changefrom year 1 to year 8) and the survey data collected at the end of the eight yearswere used to form the group climate dependent variables The survey to assessthe success of these change strategies was developed with expert and facultyinput

22 MeasuresGroup dependent variables climate for diversity (consensus and valence) Fivegroup dependent variables were developed to assess the climate for diversityExploratory factor analysis ensured that each of these measures developed forthe study tapped into unique constructs Each of these measures utilized thesame regve-point scale ranging from 1 (strongly disagree) to 5 (strongly agree)Below we regrst describe the regve dependent variables and then we discuss howwe aggregated them to create group level consensus and valence measures

Diversity as an organizational goal We used Kossek and Zoniarsquos (1993) six-item scale (alpha 089) to assess the perceptions of the extent to which membersbelieved that organizational excellence and effective functioning were related tothe recruitment and retention of faculty who are female or minority Sampleitem ordfIf organization X is to remain an excellent institution it must recruit andretain more minority facultyordm

Commitment of management to diversity We developed a regve-item scale(alpha 085) to assess the administrationrsquos commitment to diversity through itssupport of efforts to increase faculty diversity through hiring and relatedpolicies Sample item ordfThe Dean of my college is strongly committed torecruiting more minority facultyordm

Perceived work-group mix We developed a regve-item scale (alpha 072) toassess the degree to which the department had a mix where women and

HRM changestrategy

335

minorities had good representation Sample item ordfMy department has a goodmix of men and women faculty membersordm

Equality of department support of women We used Kossek and Zoniarsquos(1993) three-item scale (alpha 071) to assess the degree to which faculty womenhad the same opportunity to have graduate assistants teaching release andabove average merit increases compared to males

Equality of department support of racial minorities We used Kossek andZoniarsquos (1993) three-item scale (alpha 077) to assess the degree to whichminority faculty had the same opportunity to have graduate assistantsteaching release and above average merit increases compared to non-minorities

Group consensus about climate for diversity These scales measured groupagreement for each of the regve dependent variables assessing climate Weutilized procedures outlined by Bleise and Halverson (1998) to ensure the validaggregation of individual data Group consensus was computed by

(1) determining the mean variance for each scale item per department

(2) calculating the sum of each scalersquos mean item variance

(3) dividing the sum (from step 2) by the number of items to obtain theaverage mean item variance for each scale

(4) multiplying the result from step 3 by plusmn1 to capture the level agreement

The closer the score is to 0 the higher the consensusGroup valence about the climate for diversity This scale assessed the median

score (a robust measure of central tendency) to assess the favorability of thegroup climate for the regve dependent variables Groups whose scores were at orabove the median have a more favorable climate for diversity than thosewhose scores were below the median

Group independent variables All of the independent variables were drawnfrom the universityrsquos HR information systems These included measures ofsize race gender and tenure (level) distribution by department at time 1 andtime 2 eight years later Variables were created to assess proportionaldemographic change for each group comparing the two points in time Forexample a grouprsquos proportion of women at time 1 was subtracted from theproportion of women at time 2 to measure proportional change

Change in department size was entered as a control variable since thedegree to which a department is growing may inmacruence member climateperceptions Drawing on Allportrsquos (1954) theory of social contact the moreopportunity an individual has to interact with members of other social groupsthe more likely they are to disconregrm individually held stereotypes Howeverwith increasing department size members may migrate toward others withfamiliar characteristics

JOCM163

336

3 Results31 Organizational structural demographic changeThe data show considerable success in increasing organizational diversity overthe eight-year hiring period At the organizational level the universityexperienced a 36 percent increase in female faculty from 437 at time 1 (baseline)to 644 at time 2 Similarly there was a 41 percent increase in the number ofracial minority faculty (from 211 at time 1 to 298 at time 2) Overall there were2511 total faculty at time 1 and 2684 at time 2 The net change was from 19percent at baseline to 24 percent eight years later for women and from 8percent to 11 percent for minorities Upward mobility was also improving Atbaseline the data show 9 percent women and 5 percent minority tenuredfaculty which increased to 11 percent (women) and 6 percent (minority) at time2 We took these data and transferred them into the measures described above

32 Survey response rateA total of 1529 individuals received the survey distributed at time 2 and 775were returned plusmn a 51 percent response rate For those surveys in whichdemographic identiregcation was provided analysis showed that response rateswere 47 percent for white men 46 percent for racial minority women 51 percentfor white women and 43 percent for racial minority men Of the 81 departmentsin the original sample 74 had at least three departmental respondents completethe diversity survey which were included in the regnal analysis Analyses weredone to ensure the respondents from the seven omitted departments did notstatistically differ from the 74 included Only tenure stream faculty wereincluded in the regnal dataset since there were relatively few non-tenure streamfaculty

33 Group level descriptive measures of diversity changeMeans standard deviations and correlations of measures are shown in Table IIn support of our regrst research objective Table I shows the descriptivestatistics for our new group level measures of diversity change These included

work group demographic shift

group climate (ie consensus and valence) and

group and organizational change referents

As the change variables show the average work group shift to increase inwomen was 1 percent with a standard deviation of 7 percent The averagework group increase in minorities was 1 percent with a standard deviationof 3 percent The group climate variables for the group referents (eg mixwell equality of department support of women equality of departmentsupport of minorities) and organizational referents (diversity is anorganizational goal commitment of management to diversity) are shownin Table I There was the greatest consensus that departments had equal

HRM changestrategy

337

Var

iable

Mea

nSD

12

34

5

1D

epar

tmen

tsi

zeT

ime

125

52

160

4plusmn

2D

epar

tmen

tsi

zeT

ime

229

06

176

60

89

plusmn3

Pro

por

tion

ofw

hit

esT

ime

10

900

010

210

21plusmn

4P

ropor

tion

ofw

hit

esT

ime

20

810

180

070

120

47

plusmn5

Pro

por

tion

ofw

omen

Tim

e1

019

021

005

002

001

017

plusmn6

Pro

por

tion

ofw

omen

Tim

e2

043

027

006

006

017

035

0

58

7P

ropor

tion

wit

hte

nure

Tim

e1

081

015

010

010

20

152

005

20

098

Pro

por

tion

wit

hT

enure

Tim

e2

084

011

004

003

20

082

018

20

129

Pro

por

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

034

019

20

042

009

20

100

000

76

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

039

018

20

092

010

20

39

20

130

65

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

247

550

40

062

0

59

20

092

013

009

12C

han

ge

inte

nure

tim

e1-

Tim

e2

001

001

20

152

013

20

28

20

120

0613

C

han

ge

ingen

der

Tim

e1-

Tim

e2

001

007

004

20

020

090

100

76

14C

han

ge

inw

hit

esT

ime

1-T

ime

20

010

032

009

20

162

059

2

005

2

008

15D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s2

100

057

001

001

20

192

015

024

16

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s2

099

048

007

014

20

032

006

008

17P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s2

125

047

001

011

002

20

060

1218

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s2

024

016

007

007

015

009

001

19E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s2

025

021

006

011

20

050

142

006

20D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n3

900

382

003

20

052

023

2

007

039

21

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n3

530

402

010

001

20

222

007

029

22P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

314

048

20

212

021

20

112

017

030

23E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n2

450

292

009

20

032

017

20

31

002

24E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

393

050

20

010

020

072

010

20

1625

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

361

049

20

030

012

018

20

050

2026

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

324

057

20

092

003

20

24

001

014

27P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n2

500

372

009

20

112

003

000

027

28

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

200

014

20

060

022

013

20

242

008

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n1

920

152

018

20

020

070

062

010

(con

tinued

)

Table IMeans standardsdeviations andcorrelations ofvariables

JOCM163

338

Var

iable

67

89

1011

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

plusmn7

Pro

por

tion

wit

hte

nure

Tim

e1

20

20plusmn

8P

ropor

tion

wit

hte

nure

Tim

e2

20

190

31plusmn

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

054

2

058

2

029

plusmn

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

048

2

018

20

32

073

plusmn

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

004

009

012

006

003

plusmn12

C

han

ge

inte

nure

Tim

e1-

Tim

e2

20

062

016

022

014

006

008

13C

han

ge

ingen

der

Tim

e1-

Tim

e2

032

2

019

20

130

68

055

0

0514

C

han

ge

inw

hit

esT

ime

1-T

ime

22

018

023

014

001

034

0

1315

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s0

010

020

080

140

110

0816

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s2

013

001

000

009

003

007

17P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s2

008

20

170

070

170

092

011

18E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s2

005

20

162

009

007

20

080

1519

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s2

011

011

008

20

102

023

004

20D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n0

57

20

022

014

036

0

48

014

21C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n0

030

182

001

016

026

0

0122

P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

007

20

040

050

32

030

0

0423

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n2

013

001

012

008

011

005

24E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

20

33

20

052

007

20

072

012

20

0225

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

048

0

022

004

020

027

0

1626

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

20

110

140

010

030

172

003

27P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n0

050

030

190

210

150

0628

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

20

092

004

022

20

012

002

004

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n2

026

2

018

20

102

007

20

162

024

(con

tinued

)

Table I

HRM changestrategy

339

Var

iable

1213

1415

1617

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

7P

ropor

tion

wit

hte

nure

Tim

e1

8P

ropor

tion

wit

hte

nure

Tim

e2

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

12C

han

ge

inte

nure

Tim

e1-

Tim

e2

plusmn13

C

han

ge

ingen

der

Tim

e1-

Tim

e2

001

plusmn14

C

han

ge

inw

hit

esT

ime

1-T

ime

20

162

007

plusmn15

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s0

150

110

14plusmn

16C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s2

013

021

006

20

07plusmn

17P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s0

150

130

060

38

034

plusmn

18E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s0

130

182

001

20

110

45

008

19E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s0

000

040

042

011

042

0

0020

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n0

000

050

150

34

20

32

20

0721

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n2

020

022

022

022

023

002

22P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

010

011

018

017

010

016

23E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n0

022

009

015

20

090

062

007

24E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

001

007

011

20

062

019

007

25D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

005

20

060

100

080

012

010

26C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

20

27

009

021

024

20

030

0227

P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n0

070

040

120

232

003

013

28E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

010

20

190

102

003

20

142

014

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n0

022

002

009

000

001

020

(con

tinued

)

Table I

JOCM163

340

Var

iable

1819

2021

2223

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

7P

ropor

tion

wit

hte

nure

Tim

e1

8P

ropor

tion

wit

hte

nure

Tim

e2

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

12C

han

ge

inte

nure

Tim

e1-

Tim

e2

13C

han

ge

ingen

der

Tim

e1-

Tim

e2

14C

han

ge

inw

hit

esT

ime

1-T

ime

215

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s16

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s17

P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s18

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

splusmn

19E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s0

57

plusmn20

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n2

035

2

027

plusmn

21C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n0

180

180

08plusmn

22P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

010

20

050

150

27

plusmn23

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n0

010

002

008

008

014

plusmn24

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

030

0

34

20

26

026

0

192

021

25D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

20

222

013

083

0

040

082

007

26C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

20

070

020

150

79

018

000

27P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n2

002

20

040

140

120

84

012

28E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

20

052

009

20

020

022

004

086

29

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n0

140

31

20

210

100

042

008

(con

tinued

)

Table I

HRM changestrategy

341

Table I

Var

iable

2425

2627

2829

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

7P

ropor

tion

wit

hte

nure

Tim

e1

8P

ropor

tion

wit

hte

nure

Tim

e2

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

12C

han

ge

inte

nure

Tim

e1-

Tim

e2

13C

han

ge

ingen

der

Tim

e1-

Tim

e2

14C

han

ge

inw

hit

esT

ime

1-T

ime

215

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s16

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s17

P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s18

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s19

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s20

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n21

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n22

P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

23E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n24

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

plusmn25

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

20

20plusmn

26C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

025

002

plusmn27

P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n0

080

010

19plusmn

28E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

20

32

000

20

010

05plusmn

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n0

58

20

240

180

132

003

plusmn

Note

p

001

orgre

ater

JOCM163

342

support of resources for women and men and whites and non-whites asthese measures hovered at plusmn024 and 2 025 or close to 0 which would betotal consensus There was the least consensus that minorities and majoritiesmixed well (plusmn125) with slightly more consensus that diversity should bepursued as organizational (plusmn10) The most positive valence was theperception of equality of department support for minorities (393) and thatdiversity should be pursued as a goal (39)

34 Regression results on demographic shift and climateTo address our second research objective which was to examine whetherwork groups with higher demographic shifts had more positive groupclimates for each of the regve dependent variables separate hierarchicalregressions were run to examine the relation between group demographicchange and group consensus (see Table II) or valence (see Table III) In theregrst step of all regressions proportional change in size was entered as acontrol For the consensus regressions the current mean of each dependentvariable was also entered as a control In step 2 the group proportionalchange in women and the group proportional change in minorities wereentered (In early analyses for the second step of each regression interactionterms change in gender or race by tenure were included following Aikenand Westrsquos (1993) procedures No signiregcant interactions were found) Usingstandardized beta weights as seen in step 2 of Tables II and III ourhypothesis that work groups with higher demographic shifts may havehigher consensus and valence toward diversity change received partialsupport Group referents of diversity climate were unchanged bydemographic shifts remacrecting possible dilution of HR strategiesOrganizational referents of valence were effected but consensus wasreduced in terms of whether the organization should pursue diversity as anorganizational goal Departments with higher shifts in the proportion ofwomen had lower consensus on whether the organization should pursuediversity as an organizational goal Both regressions with organizationalreferents as dependent variables were signiregcant Groups with higherincreases in the proportion of women over time tended to have morefavorable valence regarding the desirability of promoting diversity as anorganizational goal but were less sure of managementrsquos commitment asthey had more experience with the reality of implementing theorganizational change

35 Regression results on positive senior women and minority effect on climateOur third research objective was to examine whether increasing the proportionof senior women and minorities at the department level positively relates togroup valence and consensus In our data analysis the proportional change inoverall tenured faculty was entered in step 2 and then the proportional change

HRM changestrategy

343

Org

aniz

atio

nD

epar

tmen

tD

iver

sity

asor

gan

izat

ional

goa

l

Com

mit

men

tof

man

agem

ent

todiv

ersi

ty

Per

ceiv

edw

ork-g

roup

mix

Equal

ity

ofdep

artm

ent

suppor

tof

wom

enE

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

Ste

p1

Dep

tsi

zeT

ime

2-T

ime

10

020

080

180

000

06M

ean

ofD

V0

35

022

017

002

029

D

R2

012

0

060

060

000

09T

otal

R2

012

0

060

060

000

09Ste

p2

Wom

enT

ime

2-T

ime

12

035

2

017

20

182

007

20

02W

hit

esT

ime

2-T

ime

12

002

20

052

003

004

021

Ten

ure

Tim

e2-

Tim

e1

009

003

022

010

000

DR

20

11

003

008

001

004

Tot

alR

20

23

009

014

001

013

Ste

p3

Whit

ete

nure

dm

ales

Tim

e2-

Tim

e1

20

092

012

005

20

302

031

DR

20

010

010

000

040

05T

otal

R2

024

0

100

140

050

18

Note

p

005

orgre

ater

p0

01or

gre

ater

Table IIRegressionpredictingconsensus onclimate for diversity

JOCM163

344

Org

aniz

atio

nD

epar

tmen

tD

iver

sity

asor

gan

izat

ional

goa

l

Com

mit

men

tof

man

agem

ent

todiv

ersi

ty

Per

ceiv

edw

ork-g

roup

mix

Equal

ity

ofdep

artm

ent

suppor

tof

wom

enE

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

Ste

p1

Dep

tsi

zeT

ime

2-T

ime

10

080

152

006

021

0

05D

R2

001

002

000

004

0

00T

otal

R2

001

002

000

004

0

00Ste

p2

Wom

enT

ime

2-T

ime

10

40

20

38

20

21

20

062

013

Whit

esT

ime

2-T

ime

12

007

019

008

20

172

001

Ten

ure

Tim

e2-

Tim

e1

20

072

013

012

016

019

DR

20

16

017

0

060

070

05T

otal

R2

017

0

19

006

011

0

05Ste

p3

Whit

ete

nure

dm

ales

Tim

e2-

Tim

e1

005

011

000

007

035

D

R2

000

001

000

000

004

T

otal

R2

017

0

20

006

011

009

Note

p

005

orgre

ater

p0

01or

gre

ater

Table IIIRegression

predicting valenceof climate for

diversity

HRM changestrategy

345

in tenured white males was entered in step 3 The results are that the thirdblock was not statistically signiregcant in the regressions showing that smallincreases in the group proportions of senior women and minorities did notincrease group consensus or valence regarding climate for change

4 DiscussionThis study explored links between actual change in organizational structureorganizational demography and work group interpretation of this change Ourresults showed that HR strategies that focus on structural change withoutworking to develop supportive group norms and positive climate areinadequate change strategies Current employer efforts to increase diversity asan isolated HR strategy may not necessarily lead to increased memberagreement that change is favorable Most change efforts like the diversitychange effort we studied are focused at the organizational or individualintervention levels and under-emphasize the work group level and theimportance of group tipping points Our study highlights a disconnect betweenorganizational level diversity practice which often assumes that hiring moreminorities and women will improve climate and what our research found at thegroup level a climate that was mixed and ambivalent at best and sometimesnegative

41 Group level measures of demographic change consensus and valence ondiversity climateThe regrst objective of this study was to develop new theory and measures ofgroup level change in diversity We provided new ways to measure and groupsocial construction of change These included consensus the degree to whichmembers held common perceptions of diversity climate and valence thedirection of those perceptions Our measures identify group and organizationallevel referents of HR strategies to increase demographic diversity Group levelreferents of diversity change were how well members perceived their workgroup mix and the perceived equality of resource allocation acrossdemographic groups Organizational referents were managementcommitment to change and whether diversity should be an organizationalstrategy assessed general support for change Our study highlights howimportant it is for research and practice to not only examine how change indemography relates to differences in individual attitudes but also to groupmember shared perceptions about the direction of the climate for changeUnderstanding linkages between HR changes (in this case group demographywhich raises unique psychological processes related to social identity) andpositive consensus is important because it is relevant to cooperation workexperiences interpersonal conmacrict stress and norms in groups striving tomanage change

JOCM163

346

42 Greater work group demographic shift does not necessarily result infavorable climateOur second objective was to assess whether work groups with higherdemographic shifts tend to have more positive climates that is higherconsensus and valence toward diversity change for both organizational andgroup level change referents Our results only partially supported this beliefthat greater structural demographic shift would positively relate to a morefavorable climate for diversity Although at the organizational level theuniversity had experienced signiregcant change in organizational demographyan increase in the overall representation of white women (36 percent) andminorities (41 percent) over time work group members in units with thegreatest change did not have higher consensus or more favorable valenceWe found more favorable results for organizational than group levelreferents of change Members may regnd it easier to agree with organizationalreferents and more difregcult to agree on group level referents due to theirgreater psychological proximity to group processes that affect how theypersonally experience conmacrict and wellbeing from change in their dailywork Work groups with more demographic change had higher valence butlower consensus on whether the organization should pursue diversity as astrategic goal When organizational change strategies to increase diversityare enacted at the work group level the group demographic shift may bediluted across the organization to only slightly increase the representation ofwomen and minorities within actual groups Due to this dilution effectgroup membersrsquo social construction of the HR strategies may lack consensusor positive valence regarding the organizational diversity change strategiesIf the change is not large enough to create a critical mass to tiprepresentation of minorities to at least 35 percent according to Kanter (1977)negative intergroup dynamics such as increased resource competition andtokenism may still occur

43 Diluting diversity gradually increasing the proportion of senior womenminorities does not necessarily improve group climateOur third objective was to investigate whether increasing the proportion ofsenior women and minorities at the department level positively relates togroup valence and consensus There was no relation between increasing theproportion of senior women and minorities and group valence andconsensus Due to diluted saliency of organizational HR change strategies toincrease diversity at the work group level slight demographic shifts toproportionally increase the representation of senior women and minoritiesdid not relate to higher group valence or consensus This increase did littleto improve group climate and may have even hurt the climate in theshort run

HRM changestrategy

347

5 Implications for future research and practice51 Incremental structural change not sufregcient to improve climateWe hope this study will encourage researchers and practitioners interested indiversity change strategies to augment individual and organizational levels ofintervention to focus on work group climates supporting change Our resultshighlight some disparities between the underlying assumptions of HRstrategies to manage organizational change (increasing structural diversitywill be positive for climate) and reality at the department level (incrementalincreases did not improve climate) Organizational strategists implementingdiversity recruitment initiatives typically assume that change strategies (in thiscase hiring diversity) would improve the group social climate due to a positivechange in member attitudes based on increased group diversity and positivesocial contact Yet our study does not support this view Although theorganization we studied had had signiregcant success in enacting the formal HRstrategy plusmn increasing the representation of women and minorities at theorganizational level plusmn our research shows that gradually shifting thedemography of work groups may not necessarily lead to increased consensusor valence about the social climate within work groups For example it did notlead to improved climate regarding the perceived fairness of resourceallocation or good mixing in social interaction In fact the incremental increasein the proportion of minority groups lowered consensus and agreement in theshort term as subgroups sorted out shifting and socially ambiguous powerrelationships within work units These results suggest the importance ofmeasuring change initiatives impacts at the group level and having group levelinterventions such as team-building and group-focused resource allocation tosupport organizational level diversity strategies

52 Focus change strategies on tipping points within targeted work unitsFuture research should consider how HR strategies to manage organizationaldemographic change are socially constructed and enacted at the work grouplevel It is especially important to identify what constitutes a group leveltipping point or critical mass across hierarchical levels and differentdemographic groups to enable positive consensus supporting changePractice might regnd it more fruitful to focus future HR strategies to increasediversity through hiring to alter tipping points within targeted work groupsThis will dramatically alter the saliency of strategic organizationaldemographic change and intervene to shape the climate of speciregc units Ourstudy suggests that organizations may need demographic earthquakes withinwork groups to effect group consensus and valence on the climate regardingthe HR change (ie managing diversity) Dilution of change at the group levelmay ameliorate intended positive effects unless certain group level tippingpoints such as at least 35 percent of the work group are effected in each unit by

JOCM163

348

the HR strategy After eight years of hiring new faculty in our sample womenhovered at about 15 percent in 89 percent of groups and minorities had amaximum representation of 25 percent The demographic shifts whileincreasing were not dramatic enough to create imbalances that markedlyincreased the salience and power of new demographic subgroups created bythe HR strategies Instead of being balanced many of our groups were stilllargely white male and remain tilted or skewed over time which Kanter (1977)notes perpetuates stereotypical negative dynamics

53 Avoiding backlash weaving tipping points across the hierarchyRecent work at MIT by Bailyn (2000) suggests that backlash and resistance toorganizational change may be occurring among senior white male faculty inresponse to HR diversity strategies Initially women and minorities now enterdepartments with relatively equal resources to white men as supported by theHR system design However by the time women and minorities reach theassociate and full professor levels subtle discrimination processes occur in howHR policies are applied such as lower access to resources promotions andpoor climate Despite the increase in upward mobility the representation ofwomen and minorities at the top is still relatively low Ely (1994) found that ifwomen are under-represented at higher organizational levels relations betweenwomen at lower levels might be of lower quality due to increased competitionand negative gender dynamics in the larger social system If HR changestrategies that alter the distribution of existing resources across organizationaland demographic groups are to be successful they must take into account hownot only tipping points in terms of overall departmental representation butmore importantly how tipping points are distributed in demographies acrossthe hierarchies of work groups or departments

HR diversity change strategies that bring in many new hires of differentbackgrounds may have negative ramiregcations for social functioning in groupsparticularly if new members are not supported by additional HR strategies thatallow them to enter work groups on an equal footing or if their work groups donot have leadership that mirror their demographic identities As Kanter (1983)found identity groups must be equal in their access to power resources ifcontact is to produce improvement in attitudes There must also be time for thenew members to be socialized and integrated into the culture Harrison et al(1998) found that the longer the length of time that members worked togetherthe greater the effects of deep level attitude similarity Organizations areincreasing surface (demographic) and deep level (attitudinal) diversitysimultaneously which is likely to have negative short-term ramiregcations forthe consensus and direction of climate We argue HR strategies must not onlyfocus on the surface level plusmn remacrecting structural and demographic attributes plusmnbut also simultaneously to deep level characteristics remacrecting values within

HRM changestrategy

349

the context of speciregc groups at all organizational levels In sum HR strategistsand change agents should focus on how HR change strategies relate to the jointalignment of group memberrsquos multiple characteristics (eg demographichierarchical status task) in work groups plusmn that is to examine group contextualinmacruences (Wharton 1992)

54 LimitationsOur study while making contributions to organizational research is notwithout limitations Though large in individual respondents by utilizing thegroup as the level of analysis our sample size includes only 74 groups Thiscould reduce the level of statistical power of our regndings Additionally becausethe sample was based on one organization it is important to replicate thisanalysis across multiple organizations to substantiate our results Althoughchanges in department heterogeneity across an eight-year period wereincluded this is not a longitudinal design Future research should employrepeated measures over a length of time across all variables measuringdemographic change in heterogeneity as well as consensus and valence Ourstudy points to some of the challenges in measuring ordfsuccess in increasingdiversityordm as we did not look at annual turnover or the effects of the glassceiling at the group level in this study which is a gap we hope futureresearchers will be motivated to study spurred by our research

Since little published work has been done on group consensus regarding theeffects of HR policies to increase diversity through demographic change weextend many social science literatures ranging from organizationaldevelopment to demography and HR Future research on diversity andorganizational change also should strive to be interdisciplinary as theliterature on HR strategy and change has been largely developed in the HRliterature separate from other social science disciplines reducing theoreticalintegration Without such integration and increased scholarly and practitionerattention to climate and consensus many organizations may fail to supportgroup membersrsquo explicit use of their cultural experiences to advanceorganizational learning from enacted diversity change strategies despite thegroup and intergroup social tensions that will inevitably occur (Thomas andEly 1996)

References

Aiken LA and West SG (1993) Multiple Regression Testing and Interpreting InteractionsSage Publications Newbury Park CA

Alderfer CP and Smith KK (1982) ordfStudying embedded intergroup relations inorganizationsordm Administrative Science Quarterly Vol 27 pp 35-65

Allport G (1954) The Nature of Prejudice Addison Wesley Cambridge MA

Bailyn L (2000) ordfGender and diversity in organizations preconference presentationordm paperpresented at the National Academy of Management meeting Toronto

JOCM163

350

Blalock HM (1967) Toward a Theory of Minority Group Relations Wiley New York NY

Bleise PD and Halverson RR (1998) ordfGroup consensus and psychological well-being a largeregeld studyordm Journal of Applied Social Psychology Vol 28 pp 563-80

Brown KG Kozlowski SWJ and Hattrup K (1996) Theory Issues and Recommendations inConceptualizing Agreement as a Construct in Organizational Research The Search forConsensus Regarding Consensus National Academy of Management Cincinnati OH

Cox T Jr (1993) Cultural Diversity in Organizations Berrett-Koehler San Francisco CA

Ely R (1994) ordfThe effects of organizational demographics and social identity on relationshipsamong professional womenordm Administrative Science Quarterly Vol 39 pp 203-38

Ely R (1995) ordfThe power in demography womenrsquos social constructions of gender identity atworkordm Academy of Management Journal Vol 38 pp 589-634

Festinger L (1954) ordfA theory of social comparison processesordm Human Relations Vol 7pp 117-40

Forsyth DR (1990) Group Dynamics BrooksCole Publishing Paciregc Grove CA

Harrison DA Price KH and Bell MP (1998) ordfBeyond relational demography time and theeffects of surface and deep-level diversity on work group cohesionordm Academy ofManagement Journal Vol 41 pp 96-107

James K Lovato C and Khoo G (1994) ordfSocial identity correlates of minority workersrsquo healthordmAcademy of Management Journal Vol 37 p 383+

Kanter R (1977) ordfSome effects of proportions on group life skewed sex ratios and reactions totoken womenordm American Journal of Sociology Vol 82 pp 965-90

Kanter R (1983) Men and Women of the Corporation Basic Books New York NY

Kossek EE and Lobel SA (1996) ordfIntroduction transforming human resource systems tomanage diversityordm in Managing Diversity Human Resource Strategies for Transforming

the Workplace Blackwell Oxford pp 1-19

Kossek EE and Zonia SC (1993) ordfAssessing diversity climate a regeld study of reactions toemployer efforts to promote diversityordm Journal of Organizational Behavior Vol 14pp 61-81

Larkey LK (1996) ordfToward a theory of communicative interactions in culturally diverseworkgroupsordm Academy of Management Review Vol 21 pp 463-91

Lau DC and Murnighan JK (1998) ordfDemographic diversity and faultlines the compositionaldynamics of organizational groupsordm Academy of Management Review Vol 29 pp 325-40

Lindell MK and Brandt CJ (2000) ordfClimate quality and climate consensus as mediators of therelationship between organizational antecedents and outcomesordm Journal of AppliedPsychology Vol 85 pp 331-48

McGrath JE Berdahl JL and Arrow H (1995) ordfTraits expectations culture and clout thedynamics of diversity in work groupsordm in Jackson SE and Ruderman MN (Eds)Diversity in Work Teams Research Paradigms for a Changing Workplace AmericanPsychological Association Washington DC

Payne RL and Pugh DS (1976) ordfOrganizational structure and climateordm in Dunnette MD(Ed) Handbook of Industrial and Organizational Psychology Rand McNally Chicago ILpp 1125-73

Reichers A and Schneider B (1990) ordfClimate and culture an evolution of constructsordm inSchneider B (Ed) Organizational Climate and Culture Jossey-Bass San Francisco CA

HRM changestrategy

351

11 Developing group level measures of the social construction of change indiversityAddressing our regrst objective we develop theory and measures forunderstanding the social construction of change in diversity at the workgroup level These constructs include work group demographic shift groupclimate (ie consensus and valence) and group and organizational changereferents First it is important to examine structural progress which we callordfwork group demographic shiftordm This is deregned as a work grouprsquos increase inthe proportion of women and racial minorities over time occurring throughchange in member composition If an organization has been successful in hiringstrategies it will increase the number of women and minorities within anorganization as a whole These hiring increases typically create demographicshifts that are cascaded down to the work group level and impact the socialclimate for diversity The study of work group demographic shift raises issuesnot ordinarily addressed in cross-sectional HR management research ondiversity (McGrath et al 1995) Members of two groups with the samecomposition might not necessarily share current perceptions about the climatefor diversity plusmn whether it is a favorable or unfavorable climate The reasons forthis difference may include but not be limited to historical differences inheterogeneity the degree to which diversity is seen as valuable for achievingstrategic goals and the quality of interpersonal relations and resourceallocation between members of different backgrounds As Kanter (1977)argued relative (rather than absolute) numbers or proportional representationshape and tip minority and majority group dynamics surroundingorganizational change Work group demographic shift remacrects contextualand relative experience with a regrmrsquos overall or absolute demographic hiringpractices

The second set of constructs we use to understand demographic change isordfgroup level climateordm the social construction of the demographic shift Climateor prevailing member beliefs about ordfthe way things are around hereordm isincreasingly being studied less in terms of ordfthe climateordm and more in terms ofordfthe climate for somethingordm plusmn a speciregc referent (Schneider and Reichers 1983)Using cross-sectional data Kossek and Zonia (1993) examined how currentdemographic composition related to individual perceptions of diversity climateWe build on this previous research by measuring climate at the group level toassess shared member perceptions of the effects of a demographic shift over asigniregcant time period Our approach is grounded in the suggestion that byderegnition climate may be best measured as a group phenomenon (Reichers andSchneider 1990) Climate is partly a function of structural aspects of the workcontext (Payne and Pugh 1976) Work group members observe structuralfeatures such as small or large work group demographic shifts interpret themin order to make meaning of organizational change The greater the work

JOCM163

330

group shift the greater the implementation of the change strategy which couldlead to a more positive climate since these members experienced the mostchange

We contend group climate may be operationalized by the constructs ofconsensus increased agreement in member perceptions and valence thedirection of agreement Consensus remacrects what Harrison et al (1998) refer toas deep level similarity in attitudes and values Consensus identireges theextent to which there is commonality among the perceptions of unit members(Brown et al 1996) The issue here is variability within a work group Workgroups with low variability (high consensus) would share commonperceptions regarding the social climate for diversity and similarunderstanding of social reality (Festinger 1954) In high consensus groupswe would anticipate that members would experience higher behavioral andattitudinal predictability regarding how to enact change in response to anorganizationally-driven demographic shift This would shape shared basicassumptions (Schein 1988) regarding appropriate behaviors in response toorganizational change (Forsyth 1990) High attitudinal and behavioralpredictability reduces strain on interpersonal relations within groups (Zander1994) and allows for greater ease of communication and less frictionregarding how to enact change (Bleise and Halverson 1998) resulting in amore similarly construed group climate

In order to have a positive climate members must not only agree on theclimate they must also see the climate as favorable Valence measures thedirection of group membersrsquo perceptions akin to what Lindell and Brandt(2000) identify as climate quality The focus here is on group membersrsquo typicalaverage or median response Following work group demographic shifts thosegroups with high valence would have a positive construction of the changeMembers would be expected to perceive that it is important to achieveorganizational goals through implementing diversity activities In contrast lowvalence groups would have negative climates grounded in perceptions ofdiscrimination and harassment toward minority members as well as less thanenthusiastic unit response to strategic diversity initiatives (James et al 1994Schneider et al 2000)

The last set of constructs we identireged as important for understandingdiversity change are ordfgroup and organizational referents of changeordm Like mostHR organizational change strategies diversity change initiatives are enacted atthe group and organizational levels and therefore should have referents at bothlevels It is critical to assess group perceptions of organizational levelphenomenon as groups can vary in the degree to which they perceiveorganizational strategic initiatives as supporting unit objectives Based on areview of the literature we identireged two organizational referents favorableagreement that top management is committed to diversity and the belief that

HRM changestrategy

331

the employer should promote diversity initiatives as a strategic organizationalgoal Many studies conclude that the success of diversity initiatives iscorrelated with member perceptions that management is highly committed tothe active pursuit of diversity policies as an organizational strategic objective(Cox 1993 Kossek and Lobel 1996 Thomas and Ely 1996)

Since the group level of analysis is more proximal to interpersonal socialoutcomes we developed speciregc group indicators for the social construction ofdiversity change that were distinctive from those at the organizational levelMost individuals are sophisticated enough today to know that it is sociallydesirable to state that managing diversity is an important organizational goalas such contentions have become a truism in the management literatureHowever at the group level membersrsquo attitudes are more likely to remacrect howthey see shifting demography affecting them personally in their immediatework group context The assumptions that a work group is viewed as havingwomen and minorities who mix well with white men and that resources areequitably allocated across groups are key in operationalizing diversityparadigms (Kossek and Zonia 1993 Thomas and Ely 1996) Thus groupreferents include favorable agreement that women and minorities are seen asmixing well with group members and that work group resources are allocatedwith equal access to women and men and whites and non-whites Theseperceptions would suggest that the work group has become more multiculturalin a positive way

12 Do work groups with higher demographic shifts have a more positive groupclimateOur second objective was to address the question ordfDo groups with higherdemographic shifts have higher consensus and valence toward diversitychange for both organizational and group referentsordm Competing argumentscould be made that higher demographic shifts could lead to either negative orpositive climates depending on whether one takes a social contact or intergrouptheoretical perspective Under the social contact view work groupdemographic shift to increase the proportion of women and minoritiesenables higher cross-group interaction and mixing which is necessary toimprove perceptions of intergroup relations and social integration and reduceprejudice (Allport 1954 Triandis et al 1994) Because groups with a higherdemographic shift would have more social contact there would be higherconsensus in a positive direction about the change Unfortunately the socialcontact theory may only apply if the demographic shifts are great enough toreduce negative tokenism dynamics (Kanter 1977)

More likely in groups where the demographic shifts are not large enough tocreate a critical mass and tip the demographics to reduce tokenism thealternative perspective grounded in intergroup resource competition theory

JOCM163

332

prevails Here the argument is made that the diversity change is not largeenough to lead to greater consensus and valence The redistribution ofintergroup power through resource reallocation to reduce inequality (Alderferand Smith 1982) is likely to be experienced negatively by both majority andminority group members In groups where demographic shifts are dilutedminority members may still experience negative psychological processes anddiscrimination Majority members may see the slight demographic shifts aschanging the status quo to negatively impact them as they would see this aslosing power

H1 Work groups with higher demographic shifts may not necessarily havemore positive climates that is higher consensus and valence towarddiversity change for organizational and group level change referents

13 Is there a positive senior women and minority proportional effect on changeWe turn to our third objective that investigates ordfDoes increasing theproportion of senior women and minorities at the department level positivelyrelate to favorable group valence and high consensusordm Growing evidencesuggests that the interaction of demography with hierarchical level is critical tounderstanding organizational change processes The manner in which HRstrategies to promote demographic change shifts group demography acrossstatus levels within groups can lead to differing group climates Thus thecurrent race and gender composition of two groups can be identical in collectiverace and sex composition but have markedly different social reactions todynamics due to the saliency of membersrsquo multiple identities in relation to otherkey employee backgrounds such as hierarchical level

By examining the interaction of level (ie status) with race and sexdemography we are able to investigate the theory that it is not just the changein overall proportional demographic representation that effects climate but thecombination of demographic and non-demographic attributes that mightinmacruence the group social construction of diversity strategies in amultiplicative manner (Ely 1995) Demographic variables such as race andgender usually assume greater importance when associated with thedifferences in status (Tsui et al 1992) For example higher status membersmay be more likely to withdraw when groups are regrst integrated by lowerstatus members (Harrison et al 1998) Research on how status relates to groupproportional representation suggests that HR strategies to simply increase theoverall proportion of a minority group do not necessarily result in improvedorganizational effectiveness and increased inclusion of under-representedgroups (Ely 1994 1995) In order to have favorable climate the minority groupalso must be well represented in senior positions

Yet the dilution of organizational strategies across group hierarchiesagain may hamper the achievement of positive change in the status of

HRM changestrategy

333

groups When organizational diversity objectives to increase the upwardmobility of women and minorities are regrst enacted across work groupsstudies show that most organizations are likely to increase the number ofsenior women and racial minorities by only a few in each group (Tsui andGutek 1999) Again relying on Kanterrsquos (1977) theory of tipping points aslight gradual increase may not be strong enough to improve climatecompared to work groups that experienced no change or a decrease inhierarchical representation Recent work by Lau and Murnighan (1998)builds on Kanterrsquos (1977) work and suggests that groups have fault lines orways to align demographic characteristics across level gender and raceIncreasing gender and racial diversity through diversity hiring strategiescan alter the demographic strength and alignment or fault lines of internalsubgroups (eg senior and junior management tenured and junior faculty) Ifthese alignments are weak and work groups do not experienceordfearthquakesordm like the notion of a shift to a critical mass a slight shiftwork group demographic may be even more diluted at the senior levels Thedilution of change may result in senior groups being still largely seen aswhite male having little change impact on diversity climate

Weak demographic shifts create weak group fault lines and consequentlymay do little to change or coalesce the social construction of group climateUnder such conditions intergroup resource competition theories are againlikely to be apt explanations of group processes As a minority group gainsstature and starts to become larger and more of a realistic competitor for scarceresources the majority is more likely to discriminate (Blalock 1967) Gradualproportionate increases in the representation of women and minorities at seniorlevels might lead to greater hostility by majority members because of resourcecontrol issues This would lead to more strained intergroup relations andreduced social interaction among members resulting in greater variation of theclimate for diversity (ie reduced consensus) and more negative perceptions ofclimate

H2 Increasing the proportion of senior women and minorities at thedepartment level may not necessarily positively relate to group valenceand consensus

2 Method21 Organizational settingThe data collection period for this study was designed to mirror the timeperiod of implementation of an organizational change strategy to managediversity at a large public sector university in the USA Over eight yearsthe administration had been actively engaged in HR strategies to fosterorganizational change in diversity These included active recruitment of adiverse workforce and wide dissemination of a document afregrming the

JOCM163

334

universityrsquos commitment to diversity National consulting regrms had beencontracted to provide advice on managing diversity The administrationmade funds readily available (sometimes over the initial salary posting) tohire quality minorities and women even for units in which new tenurestream positions had not been approved Training to promote diversitysensitivity among administrators was initiated Small cash achievementawards were distributed to recognize department activities that promotedmulticulturalism

Using the human information system the researchers collected facultydemographic data by department from records in the HR information system attwo points over an eight-year period to measure change in demography bydepartment We used a stratireged sample of tenure stream faculty with fullselection of the population of women and minorities and random sampling ofwhite males At the end of the eight years a survey was developed to examinethe success of the change efforts The archival data on proportional change indemography by group were used to form the independent variables (changefrom year 1 to year 8) and the survey data collected at the end of the eight yearswere used to form the group climate dependent variables The survey to assessthe success of these change strategies was developed with expert and facultyinput

22 MeasuresGroup dependent variables climate for diversity (consensus and valence) Fivegroup dependent variables were developed to assess the climate for diversityExploratory factor analysis ensured that each of these measures developed forthe study tapped into unique constructs Each of these measures utilized thesame regve-point scale ranging from 1 (strongly disagree) to 5 (strongly agree)Below we regrst describe the regve dependent variables and then we discuss howwe aggregated them to create group level consensus and valence measures

Diversity as an organizational goal We used Kossek and Zoniarsquos (1993) six-item scale (alpha 089) to assess the perceptions of the extent to which membersbelieved that organizational excellence and effective functioning were related tothe recruitment and retention of faculty who are female or minority Sampleitem ordfIf organization X is to remain an excellent institution it must recruit andretain more minority facultyordm

Commitment of management to diversity We developed a regve-item scale(alpha 085) to assess the administrationrsquos commitment to diversity through itssupport of efforts to increase faculty diversity through hiring and relatedpolicies Sample item ordfThe Dean of my college is strongly committed torecruiting more minority facultyordm

Perceived work-group mix We developed a regve-item scale (alpha 072) toassess the degree to which the department had a mix where women and

HRM changestrategy

335

minorities had good representation Sample item ordfMy department has a goodmix of men and women faculty membersordm

Equality of department support of women We used Kossek and Zoniarsquos(1993) three-item scale (alpha 071) to assess the degree to which faculty womenhad the same opportunity to have graduate assistants teaching release andabove average merit increases compared to males

Equality of department support of racial minorities We used Kossek andZoniarsquos (1993) three-item scale (alpha 077) to assess the degree to whichminority faculty had the same opportunity to have graduate assistantsteaching release and above average merit increases compared to non-minorities

Group consensus about climate for diversity These scales measured groupagreement for each of the regve dependent variables assessing climate Weutilized procedures outlined by Bleise and Halverson (1998) to ensure the validaggregation of individual data Group consensus was computed by

(1) determining the mean variance for each scale item per department

(2) calculating the sum of each scalersquos mean item variance

(3) dividing the sum (from step 2) by the number of items to obtain theaverage mean item variance for each scale

(4) multiplying the result from step 3 by plusmn1 to capture the level agreement

The closer the score is to 0 the higher the consensusGroup valence about the climate for diversity This scale assessed the median

score (a robust measure of central tendency) to assess the favorability of thegroup climate for the regve dependent variables Groups whose scores were at orabove the median have a more favorable climate for diversity than thosewhose scores were below the median

Group independent variables All of the independent variables were drawnfrom the universityrsquos HR information systems These included measures ofsize race gender and tenure (level) distribution by department at time 1 andtime 2 eight years later Variables were created to assess proportionaldemographic change for each group comparing the two points in time Forexample a grouprsquos proportion of women at time 1 was subtracted from theproportion of women at time 2 to measure proportional change

Change in department size was entered as a control variable since thedegree to which a department is growing may inmacruence member climateperceptions Drawing on Allportrsquos (1954) theory of social contact the moreopportunity an individual has to interact with members of other social groupsthe more likely they are to disconregrm individually held stereotypes Howeverwith increasing department size members may migrate toward others withfamiliar characteristics

JOCM163

336

3 Results31 Organizational structural demographic changeThe data show considerable success in increasing organizational diversity overthe eight-year hiring period At the organizational level the universityexperienced a 36 percent increase in female faculty from 437 at time 1 (baseline)to 644 at time 2 Similarly there was a 41 percent increase in the number ofracial minority faculty (from 211 at time 1 to 298 at time 2) Overall there were2511 total faculty at time 1 and 2684 at time 2 The net change was from 19percent at baseline to 24 percent eight years later for women and from 8percent to 11 percent for minorities Upward mobility was also improving Atbaseline the data show 9 percent women and 5 percent minority tenuredfaculty which increased to 11 percent (women) and 6 percent (minority) at time2 We took these data and transferred them into the measures described above

32 Survey response rateA total of 1529 individuals received the survey distributed at time 2 and 775were returned plusmn a 51 percent response rate For those surveys in whichdemographic identiregcation was provided analysis showed that response rateswere 47 percent for white men 46 percent for racial minority women 51 percentfor white women and 43 percent for racial minority men Of the 81 departmentsin the original sample 74 had at least three departmental respondents completethe diversity survey which were included in the regnal analysis Analyses weredone to ensure the respondents from the seven omitted departments did notstatistically differ from the 74 included Only tenure stream faculty wereincluded in the regnal dataset since there were relatively few non-tenure streamfaculty

33 Group level descriptive measures of diversity changeMeans standard deviations and correlations of measures are shown in Table IIn support of our regrst research objective Table I shows the descriptivestatistics for our new group level measures of diversity change These included

work group demographic shift

group climate (ie consensus and valence) and

group and organizational change referents

As the change variables show the average work group shift to increase inwomen was 1 percent with a standard deviation of 7 percent The averagework group increase in minorities was 1 percent with a standard deviationof 3 percent The group climate variables for the group referents (eg mixwell equality of department support of women equality of departmentsupport of minorities) and organizational referents (diversity is anorganizational goal commitment of management to diversity) are shownin Table I There was the greatest consensus that departments had equal

HRM changestrategy

337

Var

iable

Mea

nSD

12

34

5

1D

epar

tmen

tsi

zeT

ime

125

52

160

4plusmn

2D

epar

tmen

tsi

zeT

ime

229

06

176

60

89

plusmn3

Pro

por

tion

ofw

hit

esT

ime

10

900

010

210

21plusmn

4P

ropor

tion

ofw

hit

esT

ime

20

810

180

070

120

47

plusmn5

Pro

por

tion

ofw

omen

Tim

e1

019

021

005

002

001

017

plusmn6

Pro

por

tion

ofw

omen

Tim

e2

043

027

006

006

017

035

0

58

7P

ropor

tion

wit

hte

nure

Tim

e1

081

015

010

010

20

152

005

20

098

Pro

por

tion

wit

hT

enure

Tim

e2

084

011

004

003

20

082

018

20

129

Pro

por

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

034

019

20

042

009

20

100

000

76

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

039

018

20

092

010

20

39

20

130

65

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

247

550

40

062

0

59

20

092

013

009

12C

han

ge

inte

nure

tim

e1-

Tim

e2

001

001

20

152

013

20

28

20

120

0613

C

han

ge

ingen

der

Tim

e1-

Tim

e2

001

007

004

20

020

090

100

76

14C

han

ge

inw

hit

esT

ime

1-T

ime

20

010

032

009

20

162

059

2

005

2

008

15D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s2

100

057

001

001

20

192

015

024

16

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s2

099

048

007

014

20

032

006

008

17P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s2

125

047

001

011

002

20

060

1218

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s2

024

016

007

007

015

009

001

19E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s2

025

021

006

011

20

050

142

006

20D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n3

900

382

003

20

052

023

2

007

039

21

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n3

530

402

010

001

20

222

007

029

22P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

314

048

20

212

021

20

112

017

030

23E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n2

450

292

009

20

032

017

20

31

002

24E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

393

050

20

010

020

072

010

20

1625

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

361

049

20

030

012

018

20

050

2026

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

324

057

20

092

003

20

24

001

014

27P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n2

500

372

009

20

112

003

000

027

28

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

200

014

20

060

022

013

20

242

008

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n1

920

152

018

20

020

070

062

010

(con

tinued

)

Table IMeans standardsdeviations andcorrelations ofvariables

JOCM163

338

Var

iable

67

89

1011

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

plusmn7

Pro

por

tion

wit

hte

nure

Tim

e1

20

20plusmn

8P

ropor

tion

wit

hte

nure

Tim

e2

20

190

31plusmn

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

054

2

058

2

029

plusmn

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

048

2

018

20

32

073

plusmn

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

004

009

012

006

003

plusmn12

C

han

ge

inte

nure

Tim

e1-

Tim

e2

20

062

016

022

014

006

008

13C

han

ge

ingen

der

Tim

e1-

Tim

e2

032

2

019

20

130

68

055

0

0514

C

han

ge

inw

hit

esT

ime

1-T

ime

22

018

023

014

001

034

0

1315

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s0

010

020

080

140

110

0816

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s2

013

001

000

009

003

007

17P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s2

008

20

170

070

170

092

011

18E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s2

005

20

162

009

007

20

080

1519

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s2

011

011

008

20

102

023

004

20D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n0

57

20

022

014

036

0

48

014

21C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n0

030

182

001

016

026

0

0122

P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

007

20

040

050

32

030

0

0423

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n2

013

001

012

008

011

005

24E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

20

33

20

052

007

20

072

012

20

0225

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

048

0

022

004

020

027

0

1626

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

20

110

140

010

030

172

003

27P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n0

050

030

190

210

150

0628

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

20

092

004

022

20

012

002

004

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n2

026

2

018

20

102

007

20

162

024

(con

tinued

)

Table I

HRM changestrategy

339

Var

iable

1213

1415

1617

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

7P

ropor

tion

wit

hte

nure

Tim

e1

8P

ropor

tion

wit

hte

nure

Tim

e2

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

12C

han

ge

inte

nure

Tim

e1-

Tim

e2

plusmn13

C

han

ge

ingen

der

Tim

e1-

Tim

e2

001

plusmn14

C

han

ge

inw

hit

esT

ime

1-T

ime

20

162

007

plusmn15

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s0

150

110

14plusmn

16C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s2

013

021

006

20

07plusmn

17P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s0

150

130

060

38

034

plusmn

18E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s0

130

182

001

20

110

45

008

19E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s0

000

040

042

011

042

0

0020

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n0

000

050

150

34

20

32

20

0721

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n2

020

022

022

022

023

002

22P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

010

011

018

017

010

016

23E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n0

022

009

015

20

090

062

007

24E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

001

007

011

20

062

019

007

25D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

005

20

060

100

080

012

010

26C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

20

27

009

021

024

20

030

0227

P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n0

070

040

120

232

003

013

28E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

010

20

190

102

003

20

142

014

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n0

022

002

009

000

001

020

(con

tinued

)

Table I

JOCM163

340

Var

iable

1819

2021

2223

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

7P

ropor

tion

wit

hte

nure

Tim

e1

8P

ropor

tion

wit

hte

nure

Tim

e2

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

12C

han

ge

inte

nure

Tim

e1-

Tim

e2

13C

han

ge

ingen

der

Tim

e1-

Tim

e2

14C

han

ge

inw

hit

esT

ime

1-T

ime

215

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s16

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s17

P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s18

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

splusmn

19E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s0

57

plusmn20

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n2

035

2

027

plusmn

21C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n0

180

180

08plusmn

22P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

010

20

050

150

27

plusmn23

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n0

010

002

008

008

014

plusmn24

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

030

0

34

20

26

026

0

192

021

25D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

20

222

013

083

0

040

082

007

26C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

20

070

020

150

79

018

000

27P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n2

002

20

040

140

120

84

012

28E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

20

052

009

20

020

022

004

086

29

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n0

140

31

20

210

100

042

008

(con

tinued

)

Table I

HRM changestrategy

341

Table I

Var

iable

2425

2627

2829

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

7P

ropor

tion

wit

hte

nure

Tim

e1

8P

ropor

tion

wit

hte

nure

Tim

e2

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

12C

han

ge

inte

nure

Tim

e1-

Tim

e2

13C

han

ge

ingen

der

Tim

e1-

Tim

e2

14C

han

ge

inw

hit

esT

ime

1-T

ime

215

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s16

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s17

P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s18

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s19

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s20

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n21

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n22

P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

23E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n24

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

plusmn25

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

20

20plusmn

26C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

025

002

plusmn27

P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n0

080

010

19plusmn

28E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

20

32

000

20

010

05plusmn

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n0

58

20

240

180

132

003

plusmn

Note

p

001

orgre

ater

JOCM163

342

support of resources for women and men and whites and non-whites asthese measures hovered at plusmn024 and 2 025 or close to 0 which would betotal consensus There was the least consensus that minorities and majoritiesmixed well (plusmn125) with slightly more consensus that diversity should bepursued as organizational (plusmn10) The most positive valence was theperception of equality of department support for minorities (393) and thatdiversity should be pursued as a goal (39)

34 Regression results on demographic shift and climateTo address our second research objective which was to examine whetherwork groups with higher demographic shifts had more positive groupclimates for each of the regve dependent variables separate hierarchicalregressions were run to examine the relation between group demographicchange and group consensus (see Table II) or valence (see Table III) In theregrst step of all regressions proportional change in size was entered as acontrol For the consensus regressions the current mean of each dependentvariable was also entered as a control In step 2 the group proportionalchange in women and the group proportional change in minorities wereentered (In early analyses for the second step of each regression interactionterms change in gender or race by tenure were included following Aikenand Westrsquos (1993) procedures No signiregcant interactions were found) Usingstandardized beta weights as seen in step 2 of Tables II and III ourhypothesis that work groups with higher demographic shifts may havehigher consensus and valence toward diversity change received partialsupport Group referents of diversity climate were unchanged bydemographic shifts remacrecting possible dilution of HR strategiesOrganizational referents of valence were effected but consensus wasreduced in terms of whether the organization should pursue diversity as anorganizational goal Departments with higher shifts in the proportion ofwomen had lower consensus on whether the organization should pursuediversity as an organizational goal Both regressions with organizationalreferents as dependent variables were signiregcant Groups with higherincreases in the proportion of women over time tended to have morefavorable valence regarding the desirability of promoting diversity as anorganizational goal but were less sure of managementrsquos commitment asthey had more experience with the reality of implementing theorganizational change

35 Regression results on positive senior women and minority effect on climateOur third research objective was to examine whether increasing the proportionof senior women and minorities at the department level positively relates togroup valence and consensus In our data analysis the proportional change inoverall tenured faculty was entered in step 2 and then the proportional change

HRM changestrategy

343

Org

aniz

atio

nD

epar

tmen

tD

iver

sity

asor

gan

izat

ional

goa

l

Com

mit

men

tof

man

agem

ent

todiv

ersi

ty

Per

ceiv

edw

ork-g

roup

mix

Equal

ity

ofdep

artm

ent

suppor

tof

wom

enE

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

Ste

p1

Dep

tsi

zeT

ime

2-T

ime

10

020

080

180

000

06M

ean

ofD

V0

35

022

017

002

029

D

R2

012

0

060

060

000

09T

otal

R2

012

0

060

060

000

09Ste

p2

Wom

enT

ime

2-T

ime

12

035

2

017

20

182

007

20

02W

hit

esT

ime

2-T

ime

12

002

20

052

003

004

021

Ten

ure

Tim

e2-

Tim

e1

009

003

022

010

000

DR

20

11

003

008

001

004

Tot

alR

20

23

009

014

001

013

Ste

p3

Whit

ete

nure

dm

ales

Tim

e2-

Tim

e1

20

092

012

005

20

302

031

DR

20

010

010

000

040

05T

otal

R2

024

0

100

140

050

18

Note

p

005

orgre

ater

p0

01or

gre

ater

Table IIRegressionpredictingconsensus onclimate for diversity

JOCM163

344

Org

aniz

atio

nD

epar

tmen

tD

iver

sity

asor

gan

izat

ional

goa

l

Com

mit

men

tof

man

agem

ent

todiv

ersi

ty

Per

ceiv

edw

ork-g

roup

mix

Equal

ity

ofdep

artm

ent

suppor

tof

wom

enE

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

Ste

p1

Dep

tsi

zeT

ime

2-T

ime

10

080

152

006

021

0

05D

R2

001

002

000

004

0

00T

otal

R2

001

002

000

004

0

00Ste

p2

Wom

enT

ime

2-T

ime

10

40

20

38

20

21

20

062

013

Whit

esT

ime

2-T

ime

12

007

019

008

20

172

001

Ten

ure

Tim

e2-

Tim

e1

20

072

013

012

016

019

DR

20

16

017

0

060

070

05T

otal

R2

017

0

19

006

011

0

05Ste

p3

Whit

ete

nure

dm

ales

Tim

e2-

Tim

e1

005

011

000

007

035

D

R2

000

001

000

000

004

T

otal

R2

017

0

20

006

011

009

Note

p

005

orgre

ater

p0

01or

gre

ater

Table IIIRegression

predicting valenceof climate for

diversity

HRM changestrategy

345

in tenured white males was entered in step 3 The results are that the thirdblock was not statistically signiregcant in the regressions showing that smallincreases in the group proportions of senior women and minorities did notincrease group consensus or valence regarding climate for change

4 DiscussionThis study explored links between actual change in organizational structureorganizational demography and work group interpretation of this change Ourresults showed that HR strategies that focus on structural change withoutworking to develop supportive group norms and positive climate areinadequate change strategies Current employer efforts to increase diversity asan isolated HR strategy may not necessarily lead to increased memberagreement that change is favorable Most change efforts like the diversitychange effort we studied are focused at the organizational or individualintervention levels and under-emphasize the work group level and theimportance of group tipping points Our study highlights a disconnect betweenorganizational level diversity practice which often assumes that hiring moreminorities and women will improve climate and what our research found at thegroup level a climate that was mixed and ambivalent at best and sometimesnegative

41 Group level measures of demographic change consensus and valence ondiversity climateThe regrst objective of this study was to develop new theory and measures ofgroup level change in diversity We provided new ways to measure and groupsocial construction of change These included consensus the degree to whichmembers held common perceptions of diversity climate and valence thedirection of those perceptions Our measures identify group and organizationallevel referents of HR strategies to increase demographic diversity Group levelreferents of diversity change were how well members perceived their workgroup mix and the perceived equality of resource allocation acrossdemographic groups Organizational referents were managementcommitment to change and whether diversity should be an organizationalstrategy assessed general support for change Our study highlights howimportant it is for research and practice to not only examine how change indemography relates to differences in individual attitudes but also to groupmember shared perceptions about the direction of the climate for changeUnderstanding linkages between HR changes (in this case group demographywhich raises unique psychological processes related to social identity) andpositive consensus is important because it is relevant to cooperation workexperiences interpersonal conmacrict stress and norms in groups striving tomanage change

JOCM163

346

42 Greater work group demographic shift does not necessarily result infavorable climateOur second objective was to assess whether work groups with higherdemographic shifts tend to have more positive climates that is higherconsensus and valence toward diversity change for both organizational andgroup level change referents Our results only partially supported this beliefthat greater structural demographic shift would positively relate to a morefavorable climate for diversity Although at the organizational level theuniversity had experienced signiregcant change in organizational demographyan increase in the overall representation of white women (36 percent) andminorities (41 percent) over time work group members in units with thegreatest change did not have higher consensus or more favorable valenceWe found more favorable results for organizational than group levelreferents of change Members may regnd it easier to agree with organizationalreferents and more difregcult to agree on group level referents due to theirgreater psychological proximity to group processes that affect how theypersonally experience conmacrict and wellbeing from change in their dailywork Work groups with more demographic change had higher valence butlower consensus on whether the organization should pursue diversity as astrategic goal When organizational change strategies to increase diversityare enacted at the work group level the group demographic shift may bediluted across the organization to only slightly increase the representation ofwomen and minorities within actual groups Due to this dilution effectgroup membersrsquo social construction of the HR strategies may lack consensusor positive valence regarding the organizational diversity change strategiesIf the change is not large enough to create a critical mass to tiprepresentation of minorities to at least 35 percent according to Kanter (1977)negative intergroup dynamics such as increased resource competition andtokenism may still occur

43 Diluting diversity gradually increasing the proportion of senior womenminorities does not necessarily improve group climateOur third objective was to investigate whether increasing the proportion ofsenior women and minorities at the department level positively relates togroup valence and consensus There was no relation between increasing theproportion of senior women and minorities and group valence andconsensus Due to diluted saliency of organizational HR change strategies toincrease diversity at the work group level slight demographic shifts toproportionally increase the representation of senior women and minoritiesdid not relate to higher group valence or consensus This increase did littleto improve group climate and may have even hurt the climate in theshort run

HRM changestrategy

347

5 Implications for future research and practice51 Incremental structural change not sufregcient to improve climateWe hope this study will encourage researchers and practitioners interested indiversity change strategies to augment individual and organizational levels ofintervention to focus on work group climates supporting change Our resultshighlight some disparities between the underlying assumptions of HRstrategies to manage organizational change (increasing structural diversitywill be positive for climate) and reality at the department level (incrementalincreases did not improve climate) Organizational strategists implementingdiversity recruitment initiatives typically assume that change strategies (in thiscase hiring diversity) would improve the group social climate due to a positivechange in member attitudes based on increased group diversity and positivesocial contact Yet our study does not support this view Although theorganization we studied had had signiregcant success in enacting the formal HRstrategy plusmn increasing the representation of women and minorities at theorganizational level plusmn our research shows that gradually shifting thedemography of work groups may not necessarily lead to increased consensusor valence about the social climate within work groups For example it did notlead to improved climate regarding the perceived fairness of resourceallocation or good mixing in social interaction In fact the incremental increasein the proportion of minority groups lowered consensus and agreement in theshort term as subgroups sorted out shifting and socially ambiguous powerrelationships within work units These results suggest the importance ofmeasuring change initiatives impacts at the group level and having group levelinterventions such as team-building and group-focused resource allocation tosupport organizational level diversity strategies

52 Focus change strategies on tipping points within targeted work unitsFuture research should consider how HR strategies to manage organizationaldemographic change are socially constructed and enacted at the work grouplevel It is especially important to identify what constitutes a group leveltipping point or critical mass across hierarchical levels and differentdemographic groups to enable positive consensus supporting changePractice might regnd it more fruitful to focus future HR strategies to increasediversity through hiring to alter tipping points within targeted work groupsThis will dramatically alter the saliency of strategic organizationaldemographic change and intervene to shape the climate of speciregc units Ourstudy suggests that organizations may need demographic earthquakes withinwork groups to effect group consensus and valence on the climate regardingthe HR change (ie managing diversity) Dilution of change at the group levelmay ameliorate intended positive effects unless certain group level tippingpoints such as at least 35 percent of the work group are effected in each unit by

JOCM163

348

the HR strategy After eight years of hiring new faculty in our sample womenhovered at about 15 percent in 89 percent of groups and minorities had amaximum representation of 25 percent The demographic shifts whileincreasing were not dramatic enough to create imbalances that markedlyincreased the salience and power of new demographic subgroups created bythe HR strategies Instead of being balanced many of our groups were stilllargely white male and remain tilted or skewed over time which Kanter (1977)notes perpetuates stereotypical negative dynamics

53 Avoiding backlash weaving tipping points across the hierarchyRecent work at MIT by Bailyn (2000) suggests that backlash and resistance toorganizational change may be occurring among senior white male faculty inresponse to HR diversity strategies Initially women and minorities now enterdepartments with relatively equal resources to white men as supported by theHR system design However by the time women and minorities reach theassociate and full professor levels subtle discrimination processes occur in howHR policies are applied such as lower access to resources promotions andpoor climate Despite the increase in upward mobility the representation ofwomen and minorities at the top is still relatively low Ely (1994) found that ifwomen are under-represented at higher organizational levels relations betweenwomen at lower levels might be of lower quality due to increased competitionand negative gender dynamics in the larger social system If HR changestrategies that alter the distribution of existing resources across organizationaland demographic groups are to be successful they must take into account hownot only tipping points in terms of overall departmental representation butmore importantly how tipping points are distributed in demographies acrossthe hierarchies of work groups or departments

HR diversity change strategies that bring in many new hires of differentbackgrounds may have negative ramiregcations for social functioning in groupsparticularly if new members are not supported by additional HR strategies thatallow them to enter work groups on an equal footing or if their work groups donot have leadership that mirror their demographic identities As Kanter (1983)found identity groups must be equal in their access to power resources ifcontact is to produce improvement in attitudes There must also be time for thenew members to be socialized and integrated into the culture Harrison et al(1998) found that the longer the length of time that members worked togetherthe greater the effects of deep level attitude similarity Organizations areincreasing surface (demographic) and deep level (attitudinal) diversitysimultaneously which is likely to have negative short-term ramiregcations forthe consensus and direction of climate We argue HR strategies must not onlyfocus on the surface level plusmn remacrecting structural and demographic attributes plusmnbut also simultaneously to deep level characteristics remacrecting values within

HRM changestrategy

349

the context of speciregc groups at all organizational levels In sum HR strategistsand change agents should focus on how HR change strategies relate to the jointalignment of group memberrsquos multiple characteristics (eg demographichierarchical status task) in work groups plusmn that is to examine group contextualinmacruences (Wharton 1992)

54 LimitationsOur study while making contributions to organizational research is notwithout limitations Though large in individual respondents by utilizing thegroup as the level of analysis our sample size includes only 74 groups Thiscould reduce the level of statistical power of our regndings Additionally becausethe sample was based on one organization it is important to replicate thisanalysis across multiple organizations to substantiate our results Althoughchanges in department heterogeneity across an eight-year period wereincluded this is not a longitudinal design Future research should employrepeated measures over a length of time across all variables measuringdemographic change in heterogeneity as well as consensus and valence Ourstudy points to some of the challenges in measuring ordfsuccess in increasingdiversityordm as we did not look at annual turnover or the effects of the glassceiling at the group level in this study which is a gap we hope futureresearchers will be motivated to study spurred by our research

Since little published work has been done on group consensus regarding theeffects of HR policies to increase diversity through demographic change weextend many social science literatures ranging from organizationaldevelopment to demography and HR Future research on diversity andorganizational change also should strive to be interdisciplinary as theliterature on HR strategy and change has been largely developed in the HRliterature separate from other social science disciplines reducing theoreticalintegration Without such integration and increased scholarly and practitionerattention to climate and consensus many organizations may fail to supportgroup membersrsquo explicit use of their cultural experiences to advanceorganizational learning from enacted diversity change strategies despite thegroup and intergroup social tensions that will inevitably occur (Thomas andEly 1996)

References

Aiken LA and West SG (1993) Multiple Regression Testing and Interpreting InteractionsSage Publications Newbury Park CA

Alderfer CP and Smith KK (1982) ordfStudying embedded intergroup relations inorganizationsordm Administrative Science Quarterly Vol 27 pp 35-65

Allport G (1954) The Nature of Prejudice Addison Wesley Cambridge MA

Bailyn L (2000) ordfGender and diversity in organizations preconference presentationordm paperpresented at the National Academy of Management meeting Toronto

JOCM163

350

Blalock HM (1967) Toward a Theory of Minority Group Relations Wiley New York NY

Bleise PD and Halverson RR (1998) ordfGroup consensus and psychological well-being a largeregeld studyordm Journal of Applied Social Psychology Vol 28 pp 563-80

Brown KG Kozlowski SWJ and Hattrup K (1996) Theory Issues and Recommendations inConceptualizing Agreement as a Construct in Organizational Research The Search forConsensus Regarding Consensus National Academy of Management Cincinnati OH

Cox T Jr (1993) Cultural Diversity in Organizations Berrett-Koehler San Francisco CA

Ely R (1994) ordfThe effects of organizational demographics and social identity on relationshipsamong professional womenordm Administrative Science Quarterly Vol 39 pp 203-38

Ely R (1995) ordfThe power in demography womenrsquos social constructions of gender identity atworkordm Academy of Management Journal Vol 38 pp 589-634

Festinger L (1954) ordfA theory of social comparison processesordm Human Relations Vol 7pp 117-40

Forsyth DR (1990) Group Dynamics BrooksCole Publishing Paciregc Grove CA

Harrison DA Price KH and Bell MP (1998) ordfBeyond relational demography time and theeffects of surface and deep-level diversity on work group cohesionordm Academy ofManagement Journal Vol 41 pp 96-107

James K Lovato C and Khoo G (1994) ordfSocial identity correlates of minority workersrsquo healthordmAcademy of Management Journal Vol 37 p 383+

Kanter R (1977) ordfSome effects of proportions on group life skewed sex ratios and reactions totoken womenordm American Journal of Sociology Vol 82 pp 965-90

Kanter R (1983) Men and Women of the Corporation Basic Books New York NY

Kossek EE and Lobel SA (1996) ordfIntroduction transforming human resource systems tomanage diversityordm in Managing Diversity Human Resource Strategies for Transforming

the Workplace Blackwell Oxford pp 1-19

Kossek EE and Zonia SC (1993) ordfAssessing diversity climate a regeld study of reactions toemployer efforts to promote diversityordm Journal of Organizational Behavior Vol 14pp 61-81

Larkey LK (1996) ordfToward a theory of communicative interactions in culturally diverseworkgroupsordm Academy of Management Review Vol 21 pp 463-91

Lau DC and Murnighan JK (1998) ordfDemographic diversity and faultlines the compositionaldynamics of organizational groupsordm Academy of Management Review Vol 29 pp 325-40

Lindell MK and Brandt CJ (2000) ordfClimate quality and climate consensus as mediators of therelationship between organizational antecedents and outcomesordm Journal of AppliedPsychology Vol 85 pp 331-48

McGrath JE Berdahl JL and Arrow H (1995) ordfTraits expectations culture and clout thedynamics of diversity in work groupsordm in Jackson SE and Ruderman MN (Eds)Diversity in Work Teams Research Paradigms for a Changing Workplace AmericanPsychological Association Washington DC

Payne RL and Pugh DS (1976) ordfOrganizational structure and climateordm in Dunnette MD(Ed) Handbook of Industrial and Organizational Psychology Rand McNally Chicago ILpp 1125-73

Reichers A and Schneider B (1990) ordfClimate and culture an evolution of constructsordm inSchneider B (Ed) Organizational Climate and Culture Jossey-Bass San Francisco CA

HRM changestrategy

351

group shift the greater the implementation of the change strategy which couldlead to a more positive climate since these members experienced the mostchange

We contend group climate may be operationalized by the constructs ofconsensus increased agreement in member perceptions and valence thedirection of agreement Consensus remacrects what Harrison et al (1998) refer toas deep level similarity in attitudes and values Consensus identireges theextent to which there is commonality among the perceptions of unit members(Brown et al 1996) The issue here is variability within a work group Workgroups with low variability (high consensus) would share commonperceptions regarding the social climate for diversity and similarunderstanding of social reality (Festinger 1954) In high consensus groupswe would anticipate that members would experience higher behavioral andattitudinal predictability regarding how to enact change in response to anorganizationally-driven demographic shift This would shape shared basicassumptions (Schein 1988) regarding appropriate behaviors in response toorganizational change (Forsyth 1990) High attitudinal and behavioralpredictability reduces strain on interpersonal relations within groups (Zander1994) and allows for greater ease of communication and less frictionregarding how to enact change (Bleise and Halverson 1998) resulting in amore similarly construed group climate

In order to have a positive climate members must not only agree on theclimate they must also see the climate as favorable Valence measures thedirection of group membersrsquo perceptions akin to what Lindell and Brandt(2000) identify as climate quality The focus here is on group membersrsquo typicalaverage or median response Following work group demographic shifts thosegroups with high valence would have a positive construction of the changeMembers would be expected to perceive that it is important to achieveorganizational goals through implementing diversity activities In contrast lowvalence groups would have negative climates grounded in perceptions ofdiscrimination and harassment toward minority members as well as less thanenthusiastic unit response to strategic diversity initiatives (James et al 1994Schneider et al 2000)

The last set of constructs we identireged as important for understandingdiversity change are ordfgroup and organizational referents of changeordm Like mostHR organizational change strategies diversity change initiatives are enacted atthe group and organizational levels and therefore should have referents at bothlevels It is critical to assess group perceptions of organizational levelphenomenon as groups can vary in the degree to which they perceiveorganizational strategic initiatives as supporting unit objectives Based on areview of the literature we identireged two organizational referents favorableagreement that top management is committed to diversity and the belief that

HRM changestrategy

331

the employer should promote diversity initiatives as a strategic organizationalgoal Many studies conclude that the success of diversity initiatives iscorrelated with member perceptions that management is highly committed tothe active pursuit of diversity policies as an organizational strategic objective(Cox 1993 Kossek and Lobel 1996 Thomas and Ely 1996)

Since the group level of analysis is more proximal to interpersonal socialoutcomes we developed speciregc group indicators for the social construction ofdiversity change that were distinctive from those at the organizational levelMost individuals are sophisticated enough today to know that it is sociallydesirable to state that managing diversity is an important organizational goalas such contentions have become a truism in the management literatureHowever at the group level membersrsquo attitudes are more likely to remacrect howthey see shifting demography affecting them personally in their immediatework group context The assumptions that a work group is viewed as havingwomen and minorities who mix well with white men and that resources areequitably allocated across groups are key in operationalizing diversityparadigms (Kossek and Zonia 1993 Thomas and Ely 1996) Thus groupreferents include favorable agreement that women and minorities are seen asmixing well with group members and that work group resources are allocatedwith equal access to women and men and whites and non-whites Theseperceptions would suggest that the work group has become more multiculturalin a positive way

12 Do work groups with higher demographic shifts have a more positive groupclimateOur second objective was to address the question ordfDo groups with higherdemographic shifts have higher consensus and valence toward diversitychange for both organizational and group referentsordm Competing argumentscould be made that higher demographic shifts could lead to either negative orpositive climates depending on whether one takes a social contact or intergrouptheoretical perspective Under the social contact view work groupdemographic shift to increase the proportion of women and minoritiesenables higher cross-group interaction and mixing which is necessary toimprove perceptions of intergroup relations and social integration and reduceprejudice (Allport 1954 Triandis et al 1994) Because groups with a higherdemographic shift would have more social contact there would be higherconsensus in a positive direction about the change Unfortunately the socialcontact theory may only apply if the demographic shifts are great enough toreduce negative tokenism dynamics (Kanter 1977)

More likely in groups where the demographic shifts are not large enough tocreate a critical mass and tip the demographics to reduce tokenism thealternative perspective grounded in intergroup resource competition theory

JOCM163

332

prevails Here the argument is made that the diversity change is not largeenough to lead to greater consensus and valence The redistribution ofintergroup power through resource reallocation to reduce inequality (Alderferand Smith 1982) is likely to be experienced negatively by both majority andminority group members In groups where demographic shifts are dilutedminority members may still experience negative psychological processes anddiscrimination Majority members may see the slight demographic shifts aschanging the status quo to negatively impact them as they would see this aslosing power

H1 Work groups with higher demographic shifts may not necessarily havemore positive climates that is higher consensus and valence towarddiversity change for organizational and group level change referents

13 Is there a positive senior women and minority proportional effect on changeWe turn to our third objective that investigates ordfDoes increasing theproportion of senior women and minorities at the department level positivelyrelate to favorable group valence and high consensusordm Growing evidencesuggests that the interaction of demography with hierarchical level is critical tounderstanding organizational change processes The manner in which HRstrategies to promote demographic change shifts group demography acrossstatus levels within groups can lead to differing group climates Thus thecurrent race and gender composition of two groups can be identical in collectiverace and sex composition but have markedly different social reactions todynamics due to the saliency of membersrsquo multiple identities in relation to otherkey employee backgrounds such as hierarchical level

By examining the interaction of level (ie status) with race and sexdemography we are able to investigate the theory that it is not just the changein overall proportional demographic representation that effects climate but thecombination of demographic and non-demographic attributes that mightinmacruence the group social construction of diversity strategies in amultiplicative manner (Ely 1995) Demographic variables such as race andgender usually assume greater importance when associated with thedifferences in status (Tsui et al 1992) For example higher status membersmay be more likely to withdraw when groups are regrst integrated by lowerstatus members (Harrison et al 1998) Research on how status relates to groupproportional representation suggests that HR strategies to simply increase theoverall proportion of a minority group do not necessarily result in improvedorganizational effectiveness and increased inclusion of under-representedgroups (Ely 1994 1995) In order to have favorable climate the minority groupalso must be well represented in senior positions

Yet the dilution of organizational strategies across group hierarchiesagain may hamper the achievement of positive change in the status of

HRM changestrategy

333

groups When organizational diversity objectives to increase the upwardmobility of women and minorities are regrst enacted across work groupsstudies show that most organizations are likely to increase the number ofsenior women and racial minorities by only a few in each group (Tsui andGutek 1999) Again relying on Kanterrsquos (1977) theory of tipping points aslight gradual increase may not be strong enough to improve climatecompared to work groups that experienced no change or a decrease inhierarchical representation Recent work by Lau and Murnighan (1998)builds on Kanterrsquos (1977) work and suggests that groups have fault lines orways to align demographic characteristics across level gender and raceIncreasing gender and racial diversity through diversity hiring strategiescan alter the demographic strength and alignment or fault lines of internalsubgroups (eg senior and junior management tenured and junior faculty) Ifthese alignments are weak and work groups do not experienceordfearthquakesordm like the notion of a shift to a critical mass a slight shiftwork group demographic may be even more diluted at the senior levels Thedilution of change may result in senior groups being still largely seen aswhite male having little change impact on diversity climate

Weak demographic shifts create weak group fault lines and consequentlymay do little to change or coalesce the social construction of group climateUnder such conditions intergroup resource competition theories are againlikely to be apt explanations of group processes As a minority group gainsstature and starts to become larger and more of a realistic competitor for scarceresources the majority is more likely to discriminate (Blalock 1967) Gradualproportionate increases in the representation of women and minorities at seniorlevels might lead to greater hostility by majority members because of resourcecontrol issues This would lead to more strained intergroup relations andreduced social interaction among members resulting in greater variation of theclimate for diversity (ie reduced consensus) and more negative perceptions ofclimate

H2 Increasing the proportion of senior women and minorities at thedepartment level may not necessarily positively relate to group valenceand consensus

2 Method21 Organizational settingThe data collection period for this study was designed to mirror the timeperiod of implementation of an organizational change strategy to managediversity at a large public sector university in the USA Over eight yearsthe administration had been actively engaged in HR strategies to fosterorganizational change in diversity These included active recruitment of adiverse workforce and wide dissemination of a document afregrming the

JOCM163

334

universityrsquos commitment to diversity National consulting regrms had beencontracted to provide advice on managing diversity The administrationmade funds readily available (sometimes over the initial salary posting) tohire quality minorities and women even for units in which new tenurestream positions had not been approved Training to promote diversitysensitivity among administrators was initiated Small cash achievementawards were distributed to recognize department activities that promotedmulticulturalism

Using the human information system the researchers collected facultydemographic data by department from records in the HR information system attwo points over an eight-year period to measure change in demography bydepartment We used a stratireged sample of tenure stream faculty with fullselection of the population of women and minorities and random sampling ofwhite males At the end of the eight years a survey was developed to examinethe success of the change efforts The archival data on proportional change indemography by group were used to form the independent variables (changefrom year 1 to year 8) and the survey data collected at the end of the eight yearswere used to form the group climate dependent variables The survey to assessthe success of these change strategies was developed with expert and facultyinput

22 MeasuresGroup dependent variables climate for diversity (consensus and valence) Fivegroup dependent variables were developed to assess the climate for diversityExploratory factor analysis ensured that each of these measures developed forthe study tapped into unique constructs Each of these measures utilized thesame regve-point scale ranging from 1 (strongly disagree) to 5 (strongly agree)Below we regrst describe the regve dependent variables and then we discuss howwe aggregated them to create group level consensus and valence measures

Diversity as an organizational goal We used Kossek and Zoniarsquos (1993) six-item scale (alpha 089) to assess the perceptions of the extent to which membersbelieved that organizational excellence and effective functioning were related tothe recruitment and retention of faculty who are female or minority Sampleitem ordfIf organization X is to remain an excellent institution it must recruit andretain more minority facultyordm

Commitment of management to diversity We developed a regve-item scale(alpha 085) to assess the administrationrsquos commitment to diversity through itssupport of efforts to increase faculty diversity through hiring and relatedpolicies Sample item ordfThe Dean of my college is strongly committed torecruiting more minority facultyordm

Perceived work-group mix We developed a regve-item scale (alpha 072) toassess the degree to which the department had a mix where women and

HRM changestrategy

335

minorities had good representation Sample item ordfMy department has a goodmix of men and women faculty membersordm

Equality of department support of women We used Kossek and Zoniarsquos(1993) three-item scale (alpha 071) to assess the degree to which faculty womenhad the same opportunity to have graduate assistants teaching release andabove average merit increases compared to males

Equality of department support of racial minorities We used Kossek andZoniarsquos (1993) three-item scale (alpha 077) to assess the degree to whichminority faculty had the same opportunity to have graduate assistantsteaching release and above average merit increases compared to non-minorities

Group consensus about climate for diversity These scales measured groupagreement for each of the regve dependent variables assessing climate Weutilized procedures outlined by Bleise and Halverson (1998) to ensure the validaggregation of individual data Group consensus was computed by

(1) determining the mean variance for each scale item per department

(2) calculating the sum of each scalersquos mean item variance

(3) dividing the sum (from step 2) by the number of items to obtain theaverage mean item variance for each scale

(4) multiplying the result from step 3 by plusmn1 to capture the level agreement

The closer the score is to 0 the higher the consensusGroup valence about the climate for diversity This scale assessed the median

score (a robust measure of central tendency) to assess the favorability of thegroup climate for the regve dependent variables Groups whose scores were at orabove the median have a more favorable climate for diversity than thosewhose scores were below the median

Group independent variables All of the independent variables were drawnfrom the universityrsquos HR information systems These included measures ofsize race gender and tenure (level) distribution by department at time 1 andtime 2 eight years later Variables were created to assess proportionaldemographic change for each group comparing the two points in time Forexample a grouprsquos proportion of women at time 1 was subtracted from theproportion of women at time 2 to measure proportional change

Change in department size was entered as a control variable since thedegree to which a department is growing may inmacruence member climateperceptions Drawing on Allportrsquos (1954) theory of social contact the moreopportunity an individual has to interact with members of other social groupsthe more likely they are to disconregrm individually held stereotypes Howeverwith increasing department size members may migrate toward others withfamiliar characteristics

JOCM163

336

3 Results31 Organizational structural demographic changeThe data show considerable success in increasing organizational diversity overthe eight-year hiring period At the organizational level the universityexperienced a 36 percent increase in female faculty from 437 at time 1 (baseline)to 644 at time 2 Similarly there was a 41 percent increase in the number ofracial minority faculty (from 211 at time 1 to 298 at time 2) Overall there were2511 total faculty at time 1 and 2684 at time 2 The net change was from 19percent at baseline to 24 percent eight years later for women and from 8percent to 11 percent for minorities Upward mobility was also improving Atbaseline the data show 9 percent women and 5 percent minority tenuredfaculty which increased to 11 percent (women) and 6 percent (minority) at time2 We took these data and transferred them into the measures described above

32 Survey response rateA total of 1529 individuals received the survey distributed at time 2 and 775were returned plusmn a 51 percent response rate For those surveys in whichdemographic identiregcation was provided analysis showed that response rateswere 47 percent for white men 46 percent for racial minority women 51 percentfor white women and 43 percent for racial minority men Of the 81 departmentsin the original sample 74 had at least three departmental respondents completethe diversity survey which were included in the regnal analysis Analyses weredone to ensure the respondents from the seven omitted departments did notstatistically differ from the 74 included Only tenure stream faculty wereincluded in the regnal dataset since there were relatively few non-tenure streamfaculty

33 Group level descriptive measures of diversity changeMeans standard deviations and correlations of measures are shown in Table IIn support of our regrst research objective Table I shows the descriptivestatistics for our new group level measures of diversity change These included

work group demographic shift

group climate (ie consensus and valence) and

group and organizational change referents

As the change variables show the average work group shift to increase inwomen was 1 percent with a standard deviation of 7 percent The averagework group increase in minorities was 1 percent with a standard deviationof 3 percent The group climate variables for the group referents (eg mixwell equality of department support of women equality of departmentsupport of minorities) and organizational referents (diversity is anorganizational goal commitment of management to diversity) are shownin Table I There was the greatest consensus that departments had equal

HRM changestrategy

337

Var

iable

Mea

nSD

12

34

5

1D

epar

tmen

tsi

zeT

ime

125

52

160

4plusmn

2D

epar

tmen

tsi

zeT

ime

229

06

176

60

89

plusmn3

Pro

por

tion

ofw

hit

esT

ime

10

900

010

210

21plusmn

4P

ropor

tion

ofw

hit

esT

ime

20

810

180

070

120

47

plusmn5

Pro

por

tion

ofw

omen

Tim

e1

019

021

005

002

001

017

plusmn6

Pro

por

tion

ofw

omen

Tim

e2

043

027

006

006

017

035

0

58

7P

ropor

tion

wit

hte

nure

Tim

e1

081

015

010

010

20

152

005

20

098

Pro

por

tion

wit

hT

enure

Tim

e2

084

011

004

003

20

082

018

20

129

Pro

por

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

034

019

20

042

009

20

100

000

76

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

039

018

20

092

010

20

39

20

130

65

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

247

550

40

062

0

59

20

092

013

009

12C

han

ge

inte

nure

tim

e1-

Tim

e2

001

001

20

152

013

20

28

20

120

0613

C

han

ge

ingen

der

Tim

e1-

Tim

e2

001

007

004

20

020

090

100

76

14C

han

ge

inw

hit

esT

ime

1-T

ime

20

010

032

009

20

162

059

2

005

2

008

15D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s2

100

057

001

001

20

192

015

024

16

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s2

099

048

007

014

20

032

006

008

17P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s2

125

047

001

011

002

20

060

1218

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s2

024

016

007

007

015

009

001

19E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s2

025

021

006

011

20

050

142

006

20D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n3

900

382

003

20

052

023

2

007

039

21

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n3

530

402

010

001

20

222

007

029

22P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

314

048

20

212

021

20

112

017

030

23E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n2

450

292

009

20

032

017

20

31

002

24E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

393

050

20

010

020

072

010

20

1625

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

361

049

20

030

012

018

20

050

2026

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

324

057

20

092

003

20

24

001

014

27P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n2

500

372

009

20

112

003

000

027

28

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

200

014

20

060

022

013

20

242

008

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n1

920

152

018

20

020

070

062

010

(con

tinued

)

Table IMeans standardsdeviations andcorrelations ofvariables

JOCM163

338

Var

iable

67

89

1011

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

plusmn7

Pro

por

tion

wit

hte

nure

Tim

e1

20

20plusmn

8P

ropor

tion

wit

hte

nure

Tim

e2

20

190

31plusmn

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

054

2

058

2

029

plusmn

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

048

2

018

20

32

073

plusmn

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

004

009

012

006

003

plusmn12

C

han

ge

inte

nure

Tim

e1-

Tim

e2

20

062

016

022

014

006

008

13C

han

ge

ingen

der

Tim

e1-

Tim

e2

032

2

019

20

130

68

055

0

0514

C

han

ge

inw

hit

esT

ime

1-T

ime

22

018

023

014

001

034

0

1315

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s0

010

020

080

140

110

0816

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s2

013

001

000

009

003

007

17P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s2

008

20

170

070

170

092

011

18E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s2

005

20

162

009

007

20

080

1519

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s2

011

011

008

20

102

023

004

20D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n0

57

20

022

014

036

0

48

014

21C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n0

030

182

001

016

026

0

0122

P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

007

20

040

050

32

030

0

0423

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n2

013

001

012

008

011

005

24E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

20

33

20

052

007

20

072

012

20

0225

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

048

0

022

004

020

027

0

1626

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

20

110

140

010

030

172

003

27P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n0

050

030

190

210

150

0628

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

20

092

004

022

20

012

002

004

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n2

026

2

018

20

102

007

20

162

024

(con

tinued

)

Table I

HRM changestrategy

339

Var

iable

1213

1415

1617

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

7P

ropor

tion

wit

hte

nure

Tim

e1

8P

ropor

tion

wit

hte

nure

Tim

e2

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

12C

han

ge

inte

nure

Tim

e1-

Tim

e2

plusmn13

C

han

ge

ingen

der

Tim

e1-

Tim

e2

001

plusmn14

C

han

ge

inw

hit

esT

ime

1-T

ime

20

162

007

plusmn15

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s0

150

110

14plusmn

16C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s2

013

021

006

20

07plusmn

17P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s0

150

130

060

38

034

plusmn

18E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s0

130

182

001

20

110

45

008

19E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s0

000

040

042

011

042

0

0020

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n0

000

050

150

34

20

32

20

0721

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n2

020

022

022

022

023

002

22P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

010

011

018

017

010

016

23E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n0

022

009

015

20

090

062

007

24E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

001

007

011

20

062

019

007

25D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

005

20

060

100

080

012

010

26C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

20

27

009

021

024

20

030

0227

P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n0

070

040

120

232

003

013

28E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

010

20

190

102

003

20

142

014

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n0

022

002

009

000

001

020

(con

tinued

)

Table I

JOCM163

340

Var

iable

1819

2021

2223

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

7P

ropor

tion

wit

hte

nure

Tim

e1

8P

ropor

tion

wit

hte

nure

Tim

e2

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

12C

han

ge

inte

nure

Tim

e1-

Tim

e2

13C

han

ge

ingen

der

Tim

e1-

Tim

e2

14C

han

ge

inw

hit

esT

ime

1-T

ime

215

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s16

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s17

P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s18

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

splusmn

19E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s0

57

plusmn20

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n2

035

2

027

plusmn

21C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n0

180

180

08plusmn

22P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

010

20

050

150

27

plusmn23

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n0

010

002

008

008

014

plusmn24

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

030

0

34

20

26

026

0

192

021

25D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

20

222

013

083

0

040

082

007

26C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

20

070

020

150

79

018

000

27P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n2

002

20

040

140

120

84

012

28E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

20

052

009

20

020

022

004

086

29

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n0

140

31

20

210

100

042

008

(con

tinued

)

Table I

HRM changestrategy

341

Table I

Var

iable

2425

2627

2829

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

7P

ropor

tion

wit

hte

nure

Tim

e1

8P

ropor

tion

wit

hte

nure

Tim

e2

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

12C

han

ge

inte

nure

Tim

e1-

Tim

e2

13C

han

ge

ingen

der

Tim

e1-

Tim

e2

14C

han

ge

inw

hit

esT

ime

1-T

ime

215

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s16

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s17

P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s18

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s19

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s20

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n21

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n22

P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

23E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n24

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

plusmn25

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

20

20plusmn

26C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

025

002

plusmn27

P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n0

080

010

19plusmn

28E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

20

32

000

20

010

05plusmn

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n0

58

20

240

180

132

003

plusmn

Note

p

001

orgre

ater

JOCM163

342

support of resources for women and men and whites and non-whites asthese measures hovered at plusmn024 and 2 025 or close to 0 which would betotal consensus There was the least consensus that minorities and majoritiesmixed well (plusmn125) with slightly more consensus that diversity should bepursued as organizational (plusmn10) The most positive valence was theperception of equality of department support for minorities (393) and thatdiversity should be pursued as a goal (39)

34 Regression results on demographic shift and climateTo address our second research objective which was to examine whetherwork groups with higher demographic shifts had more positive groupclimates for each of the regve dependent variables separate hierarchicalregressions were run to examine the relation between group demographicchange and group consensus (see Table II) or valence (see Table III) In theregrst step of all regressions proportional change in size was entered as acontrol For the consensus regressions the current mean of each dependentvariable was also entered as a control In step 2 the group proportionalchange in women and the group proportional change in minorities wereentered (In early analyses for the second step of each regression interactionterms change in gender or race by tenure were included following Aikenand Westrsquos (1993) procedures No signiregcant interactions were found) Usingstandardized beta weights as seen in step 2 of Tables II and III ourhypothesis that work groups with higher demographic shifts may havehigher consensus and valence toward diversity change received partialsupport Group referents of diversity climate were unchanged bydemographic shifts remacrecting possible dilution of HR strategiesOrganizational referents of valence were effected but consensus wasreduced in terms of whether the organization should pursue diversity as anorganizational goal Departments with higher shifts in the proportion ofwomen had lower consensus on whether the organization should pursuediversity as an organizational goal Both regressions with organizationalreferents as dependent variables were signiregcant Groups with higherincreases in the proportion of women over time tended to have morefavorable valence regarding the desirability of promoting diversity as anorganizational goal but were less sure of managementrsquos commitment asthey had more experience with the reality of implementing theorganizational change

35 Regression results on positive senior women and minority effect on climateOur third research objective was to examine whether increasing the proportionof senior women and minorities at the department level positively relates togroup valence and consensus In our data analysis the proportional change inoverall tenured faculty was entered in step 2 and then the proportional change

HRM changestrategy

343

Org

aniz

atio

nD

epar

tmen

tD

iver

sity

asor

gan

izat

ional

goa

l

Com

mit

men

tof

man

agem

ent

todiv

ersi

ty

Per

ceiv

edw

ork-g

roup

mix

Equal

ity

ofdep

artm

ent

suppor

tof

wom

enE

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

Ste

p1

Dep

tsi

zeT

ime

2-T

ime

10

020

080

180

000

06M

ean

ofD

V0

35

022

017

002

029

D

R2

012

0

060

060

000

09T

otal

R2

012

0

060

060

000

09Ste

p2

Wom

enT

ime

2-T

ime

12

035

2

017

20

182

007

20

02W

hit

esT

ime

2-T

ime

12

002

20

052

003

004

021

Ten

ure

Tim

e2-

Tim

e1

009

003

022

010

000

DR

20

11

003

008

001

004

Tot

alR

20

23

009

014

001

013

Ste

p3

Whit

ete

nure

dm

ales

Tim

e2-

Tim

e1

20

092

012

005

20

302

031

DR

20

010

010

000

040

05T

otal

R2

024

0

100

140

050

18

Note

p

005

orgre

ater

p0

01or

gre

ater

Table IIRegressionpredictingconsensus onclimate for diversity

JOCM163

344

Org

aniz

atio

nD

epar

tmen

tD

iver

sity

asor

gan

izat

ional

goa

l

Com

mit

men

tof

man

agem

ent

todiv

ersi

ty

Per

ceiv

edw

ork-g

roup

mix

Equal

ity

ofdep

artm

ent

suppor

tof

wom

enE

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

Ste

p1

Dep

tsi

zeT

ime

2-T

ime

10

080

152

006

021

0

05D

R2

001

002

000

004

0

00T

otal

R2

001

002

000

004

0

00Ste

p2

Wom

enT

ime

2-T

ime

10

40

20

38

20

21

20

062

013

Whit

esT

ime

2-T

ime

12

007

019

008

20

172

001

Ten

ure

Tim

e2-

Tim

e1

20

072

013

012

016

019

DR

20

16

017

0

060

070

05T

otal

R2

017

0

19

006

011

0

05Ste

p3

Whit

ete

nure

dm

ales

Tim

e2-

Tim

e1

005

011

000

007

035

D

R2

000

001

000

000

004

T

otal

R2

017

0

20

006

011

009

Note

p

005

orgre

ater

p0

01or

gre

ater

Table IIIRegression

predicting valenceof climate for

diversity

HRM changestrategy

345

in tenured white males was entered in step 3 The results are that the thirdblock was not statistically signiregcant in the regressions showing that smallincreases in the group proportions of senior women and minorities did notincrease group consensus or valence regarding climate for change

4 DiscussionThis study explored links between actual change in organizational structureorganizational demography and work group interpretation of this change Ourresults showed that HR strategies that focus on structural change withoutworking to develop supportive group norms and positive climate areinadequate change strategies Current employer efforts to increase diversity asan isolated HR strategy may not necessarily lead to increased memberagreement that change is favorable Most change efforts like the diversitychange effort we studied are focused at the organizational or individualintervention levels and under-emphasize the work group level and theimportance of group tipping points Our study highlights a disconnect betweenorganizational level diversity practice which often assumes that hiring moreminorities and women will improve climate and what our research found at thegroup level a climate that was mixed and ambivalent at best and sometimesnegative

41 Group level measures of demographic change consensus and valence ondiversity climateThe regrst objective of this study was to develop new theory and measures ofgroup level change in diversity We provided new ways to measure and groupsocial construction of change These included consensus the degree to whichmembers held common perceptions of diversity climate and valence thedirection of those perceptions Our measures identify group and organizationallevel referents of HR strategies to increase demographic diversity Group levelreferents of diversity change were how well members perceived their workgroup mix and the perceived equality of resource allocation acrossdemographic groups Organizational referents were managementcommitment to change and whether diversity should be an organizationalstrategy assessed general support for change Our study highlights howimportant it is for research and practice to not only examine how change indemography relates to differences in individual attitudes but also to groupmember shared perceptions about the direction of the climate for changeUnderstanding linkages between HR changes (in this case group demographywhich raises unique psychological processes related to social identity) andpositive consensus is important because it is relevant to cooperation workexperiences interpersonal conmacrict stress and norms in groups striving tomanage change

JOCM163

346

42 Greater work group demographic shift does not necessarily result infavorable climateOur second objective was to assess whether work groups with higherdemographic shifts tend to have more positive climates that is higherconsensus and valence toward diversity change for both organizational andgroup level change referents Our results only partially supported this beliefthat greater structural demographic shift would positively relate to a morefavorable climate for diversity Although at the organizational level theuniversity had experienced signiregcant change in organizational demographyan increase in the overall representation of white women (36 percent) andminorities (41 percent) over time work group members in units with thegreatest change did not have higher consensus or more favorable valenceWe found more favorable results for organizational than group levelreferents of change Members may regnd it easier to agree with organizationalreferents and more difregcult to agree on group level referents due to theirgreater psychological proximity to group processes that affect how theypersonally experience conmacrict and wellbeing from change in their dailywork Work groups with more demographic change had higher valence butlower consensus on whether the organization should pursue diversity as astrategic goal When organizational change strategies to increase diversityare enacted at the work group level the group demographic shift may bediluted across the organization to only slightly increase the representation ofwomen and minorities within actual groups Due to this dilution effectgroup membersrsquo social construction of the HR strategies may lack consensusor positive valence regarding the organizational diversity change strategiesIf the change is not large enough to create a critical mass to tiprepresentation of minorities to at least 35 percent according to Kanter (1977)negative intergroup dynamics such as increased resource competition andtokenism may still occur

43 Diluting diversity gradually increasing the proportion of senior womenminorities does not necessarily improve group climateOur third objective was to investigate whether increasing the proportion ofsenior women and minorities at the department level positively relates togroup valence and consensus There was no relation between increasing theproportion of senior women and minorities and group valence andconsensus Due to diluted saliency of organizational HR change strategies toincrease diversity at the work group level slight demographic shifts toproportionally increase the representation of senior women and minoritiesdid not relate to higher group valence or consensus This increase did littleto improve group climate and may have even hurt the climate in theshort run

HRM changestrategy

347

5 Implications for future research and practice51 Incremental structural change not sufregcient to improve climateWe hope this study will encourage researchers and practitioners interested indiversity change strategies to augment individual and organizational levels ofintervention to focus on work group climates supporting change Our resultshighlight some disparities between the underlying assumptions of HRstrategies to manage organizational change (increasing structural diversitywill be positive for climate) and reality at the department level (incrementalincreases did not improve climate) Organizational strategists implementingdiversity recruitment initiatives typically assume that change strategies (in thiscase hiring diversity) would improve the group social climate due to a positivechange in member attitudes based on increased group diversity and positivesocial contact Yet our study does not support this view Although theorganization we studied had had signiregcant success in enacting the formal HRstrategy plusmn increasing the representation of women and minorities at theorganizational level plusmn our research shows that gradually shifting thedemography of work groups may not necessarily lead to increased consensusor valence about the social climate within work groups For example it did notlead to improved climate regarding the perceived fairness of resourceallocation or good mixing in social interaction In fact the incremental increasein the proportion of minority groups lowered consensus and agreement in theshort term as subgroups sorted out shifting and socially ambiguous powerrelationships within work units These results suggest the importance ofmeasuring change initiatives impacts at the group level and having group levelinterventions such as team-building and group-focused resource allocation tosupport organizational level diversity strategies

52 Focus change strategies on tipping points within targeted work unitsFuture research should consider how HR strategies to manage organizationaldemographic change are socially constructed and enacted at the work grouplevel It is especially important to identify what constitutes a group leveltipping point or critical mass across hierarchical levels and differentdemographic groups to enable positive consensus supporting changePractice might regnd it more fruitful to focus future HR strategies to increasediversity through hiring to alter tipping points within targeted work groupsThis will dramatically alter the saliency of strategic organizationaldemographic change and intervene to shape the climate of speciregc units Ourstudy suggests that organizations may need demographic earthquakes withinwork groups to effect group consensus and valence on the climate regardingthe HR change (ie managing diversity) Dilution of change at the group levelmay ameliorate intended positive effects unless certain group level tippingpoints such as at least 35 percent of the work group are effected in each unit by

JOCM163

348

the HR strategy After eight years of hiring new faculty in our sample womenhovered at about 15 percent in 89 percent of groups and minorities had amaximum representation of 25 percent The demographic shifts whileincreasing were not dramatic enough to create imbalances that markedlyincreased the salience and power of new demographic subgroups created bythe HR strategies Instead of being balanced many of our groups were stilllargely white male and remain tilted or skewed over time which Kanter (1977)notes perpetuates stereotypical negative dynamics

53 Avoiding backlash weaving tipping points across the hierarchyRecent work at MIT by Bailyn (2000) suggests that backlash and resistance toorganizational change may be occurring among senior white male faculty inresponse to HR diversity strategies Initially women and minorities now enterdepartments with relatively equal resources to white men as supported by theHR system design However by the time women and minorities reach theassociate and full professor levels subtle discrimination processes occur in howHR policies are applied such as lower access to resources promotions andpoor climate Despite the increase in upward mobility the representation ofwomen and minorities at the top is still relatively low Ely (1994) found that ifwomen are under-represented at higher organizational levels relations betweenwomen at lower levels might be of lower quality due to increased competitionand negative gender dynamics in the larger social system If HR changestrategies that alter the distribution of existing resources across organizationaland demographic groups are to be successful they must take into account hownot only tipping points in terms of overall departmental representation butmore importantly how tipping points are distributed in demographies acrossthe hierarchies of work groups or departments

HR diversity change strategies that bring in many new hires of differentbackgrounds may have negative ramiregcations for social functioning in groupsparticularly if new members are not supported by additional HR strategies thatallow them to enter work groups on an equal footing or if their work groups donot have leadership that mirror their demographic identities As Kanter (1983)found identity groups must be equal in their access to power resources ifcontact is to produce improvement in attitudes There must also be time for thenew members to be socialized and integrated into the culture Harrison et al(1998) found that the longer the length of time that members worked togetherthe greater the effects of deep level attitude similarity Organizations areincreasing surface (demographic) and deep level (attitudinal) diversitysimultaneously which is likely to have negative short-term ramiregcations forthe consensus and direction of climate We argue HR strategies must not onlyfocus on the surface level plusmn remacrecting structural and demographic attributes plusmnbut also simultaneously to deep level characteristics remacrecting values within

HRM changestrategy

349

the context of speciregc groups at all organizational levels In sum HR strategistsand change agents should focus on how HR change strategies relate to the jointalignment of group memberrsquos multiple characteristics (eg demographichierarchical status task) in work groups plusmn that is to examine group contextualinmacruences (Wharton 1992)

54 LimitationsOur study while making contributions to organizational research is notwithout limitations Though large in individual respondents by utilizing thegroup as the level of analysis our sample size includes only 74 groups Thiscould reduce the level of statistical power of our regndings Additionally becausethe sample was based on one organization it is important to replicate thisanalysis across multiple organizations to substantiate our results Althoughchanges in department heterogeneity across an eight-year period wereincluded this is not a longitudinal design Future research should employrepeated measures over a length of time across all variables measuringdemographic change in heterogeneity as well as consensus and valence Ourstudy points to some of the challenges in measuring ordfsuccess in increasingdiversityordm as we did not look at annual turnover or the effects of the glassceiling at the group level in this study which is a gap we hope futureresearchers will be motivated to study spurred by our research

Since little published work has been done on group consensus regarding theeffects of HR policies to increase diversity through demographic change weextend many social science literatures ranging from organizationaldevelopment to demography and HR Future research on diversity andorganizational change also should strive to be interdisciplinary as theliterature on HR strategy and change has been largely developed in the HRliterature separate from other social science disciplines reducing theoreticalintegration Without such integration and increased scholarly and practitionerattention to climate and consensus many organizations may fail to supportgroup membersrsquo explicit use of their cultural experiences to advanceorganizational learning from enacted diversity change strategies despite thegroup and intergroup social tensions that will inevitably occur (Thomas andEly 1996)

References

Aiken LA and West SG (1993) Multiple Regression Testing and Interpreting InteractionsSage Publications Newbury Park CA

Alderfer CP and Smith KK (1982) ordfStudying embedded intergroup relations inorganizationsordm Administrative Science Quarterly Vol 27 pp 35-65

Allport G (1954) The Nature of Prejudice Addison Wesley Cambridge MA

Bailyn L (2000) ordfGender and diversity in organizations preconference presentationordm paperpresented at the National Academy of Management meeting Toronto

JOCM163

350

Blalock HM (1967) Toward a Theory of Minority Group Relations Wiley New York NY

Bleise PD and Halverson RR (1998) ordfGroup consensus and psychological well-being a largeregeld studyordm Journal of Applied Social Psychology Vol 28 pp 563-80

Brown KG Kozlowski SWJ and Hattrup K (1996) Theory Issues and Recommendations inConceptualizing Agreement as a Construct in Organizational Research The Search forConsensus Regarding Consensus National Academy of Management Cincinnati OH

Cox T Jr (1993) Cultural Diversity in Organizations Berrett-Koehler San Francisco CA

Ely R (1994) ordfThe effects of organizational demographics and social identity on relationshipsamong professional womenordm Administrative Science Quarterly Vol 39 pp 203-38

Ely R (1995) ordfThe power in demography womenrsquos social constructions of gender identity atworkordm Academy of Management Journal Vol 38 pp 589-634

Festinger L (1954) ordfA theory of social comparison processesordm Human Relations Vol 7pp 117-40

Forsyth DR (1990) Group Dynamics BrooksCole Publishing Paciregc Grove CA

Harrison DA Price KH and Bell MP (1998) ordfBeyond relational demography time and theeffects of surface and deep-level diversity on work group cohesionordm Academy ofManagement Journal Vol 41 pp 96-107

James K Lovato C and Khoo G (1994) ordfSocial identity correlates of minority workersrsquo healthordmAcademy of Management Journal Vol 37 p 383+

Kanter R (1977) ordfSome effects of proportions on group life skewed sex ratios and reactions totoken womenordm American Journal of Sociology Vol 82 pp 965-90

Kanter R (1983) Men and Women of the Corporation Basic Books New York NY

Kossek EE and Lobel SA (1996) ordfIntroduction transforming human resource systems tomanage diversityordm in Managing Diversity Human Resource Strategies for Transforming

the Workplace Blackwell Oxford pp 1-19

Kossek EE and Zonia SC (1993) ordfAssessing diversity climate a regeld study of reactions toemployer efforts to promote diversityordm Journal of Organizational Behavior Vol 14pp 61-81

Larkey LK (1996) ordfToward a theory of communicative interactions in culturally diverseworkgroupsordm Academy of Management Review Vol 21 pp 463-91

Lau DC and Murnighan JK (1998) ordfDemographic diversity and faultlines the compositionaldynamics of organizational groupsordm Academy of Management Review Vol 29 pp 325-40

Lindell MK and Brandt CJ (2000) ordfClimate quality and climate consensus as mediators of therelationship between organizational antecedents and outcomesordm Journal of AppliedPsychology Vol 85 pp 331-48

McGrath JE Berdahl JL and Arrow H (1995) ordfTraits expectations culture and clout thedynamics of diversity in work groupsordm in Jackson SE and Ruderman MN (Eds)Diversity in Work Teams Research Paradigms for a Changing Workplace AmericanPsychological Association Washington DC

Payne RL and Pugh DS (1976) ordfOrganizational structure and climateordm in Dunnette MD(Ed) Handbook of Industrial and Organizational Psychology Rand McNally Chicago ILpp 1125-73

Reichers A and Schneider B (1990) ordfClimate and culture an evolution of constructsordm inSchneider B (Ed) Organizational Climate and Culture Jossey-Bass San Francisco CA

HRM changestrategy

351

the employer should promote diversity initiatives as a strategic organizationalgoal Many studies conclude that the success of diversity initiatives iscorrelated with member perceptions that management is highly committed tothe active pursuit of diversity policies as an organizational strategic objective(Cox 1993 Kossek and Lobel 1996 Thomas and Ely 1996)

Since the group level of analysis is more proximal to interpersonal socialoutcomes we developed speciregc group indicators for the social construction ofdiversity change that were distinctive from those at the organizational levelMost individuals are sophisticated enough today to know that it is sociallydesirable to state that managing diversity is an important organizational goalas such contentions have become a truism in the management literatureHowever at the group level membersrsquo attitudes are more likely to remacrect howthey see shifting demography affecting them personally in their immediatework group context The assumptions that a work group is viewed as havingwomen and minorities who mix well with white men and that resources areequitably allocated across groups are key in operationalizing diversityparadigms (Kossek and Zonia 1993 Thomas and Ely 1996) Thus groupreferents include favorable agreement that women and minorities are seen asmixing well with group members and that work group resources are allocatedwith equal access to women and men and whites and non-whites Theseperceptions would suggest that the work group has become more multiculturalin a positive way

12 Do work groups with higher demographic shifts have a more positive groupclimateOur second objective was to address the question ordfDo groups with higherdemographic shifts have higher consensus and valence toward diversitychange for both organizational and group referentsordm Competing argumentscould be made that higher demographic shifts could lead to either negative orpositive climates depending on whether one takes a social contact or intergrouptheoretical perspective Under the social contact view work groupdemographic shift to increase the proportion of women and minoritiesenables higher cross-group interaction and mixing which is necessary toimprove perceptions of intergroup relations and social integration and reduceprejudice (Allport 1954 Triandis et al 1994) Because groups with a higherdemographic shift would have more social contact there would be higherconsensus in a positive direction about the change Unfortunately the socialcontact theory may only apply if the demographic shifts are great enough toreduce negative tokenism dynamics (Kanter 1977)

More likely in groups where the demographic shifts are not large enough tocreate a critical mass and tip the demographics to reduce tokenism thealternative perspective grounded in intergroup resource competition theory

JOCM163

332

prevails Here the argument is made that the diversity change is not largeenough to lead to greater consensus and valence The redistribution ofintergroup power through resource reallocation to reduce inequality (Alderferand Smith 1982) is likely to be experienced negatively by both majority andminority group members In groups where demographic shifts are dilutedminority members may still experience negative psychological processes anddiscrimination Majority members may see the slight demographic shifts aschanging the status quo to negatively impact them as they would see this aslosing power

H1 Work groups with higher demographic shifts may not necessarily havemore positive climates that is higher consensus and valence towarddiversity change for organizational and group level change referents

13 Is there a positive senior women and minority proportional effect on changeWe turn to our third objective that investigates ordfDoes increasing theproportion of senior women and minorities at the department level positivelyrelate to favorable group valence and high consensusordm Growing evidencesuggests that the interaction of demography with hierarchical level is critical tounderstanding organizational change processes The manner in which HRstrategies to promote demographic change shifts group demography acrossstatus levels within groups can lead to differing group climates Thus thecurrent race and gender composition of two groups can be identical in collectiverace and sex composition but have markedly different social reactions todynamics due to the saliency of membersrsquo multiple identities in relation to otherkey employee backgrounds such as hierarchical level

By examining the interaction of level (ie status) with race and sexdemography we are able to investigate the theory that it is not just the changein overall proportional demographic representation that effects climate but thecombination of demographic and non-demographic attributes that mightinmacruence the group social construction of diversity strategies in amultiplicative manner (Ely 1995) Demographic variables such as race andgender usually assume greater importance when associated with thedifferences in status (Tsui et al 1992) For example higher status membersmay be more likely to withdraw when groups are regrst integrated by lowerstatus members (Harrison et al 1998) Research on how status relates to groupproportional representation suggests that HR strategies to simply increase theoverall proportion of a minority group do not necessarily result in improvedorganizational effectiveness and increased inclusion of under-representedgroups (Ely 1994 1995) In order to have favorable climate the minority groupalso must be well represented in senior positions

Yet the dilution of organizational strategies across group hierarchiesagain may hamper the achievement of positive change in the status of

HRM changestrategy

333

groups When organizational diversity objectives to increase the upwardmobility of women and minorities are regrst enacted across work groupsstudies show that most organizations are likely to increase the number ofsenior women and racial minorities by only a few in each group (Tsui andGutek 1999) Again relying on Kanterrsquos (1977) theory of tipping points aslight gradual increase may not be strong enough to improve climatecompared to work groups that experienced no change or a decrease inhierarchical representation Recent work by Lau and Murnighan (1998)builds on Kanterrsquos (1977) work and suggests that groups have fault lines orways to align demographic characteristics across level gender and raceIncreasing gender and racial diversity through diversity hiring strategiescan alter the demographic strength and alignment or fault lines of internalsubgroups (eg senior and junior management tenured and junior faculty) Ifthese alignments are weak and work groups do not experienceordfearthquakesordm like the notion of a shift to a critical mass a slight shiftwork group demographic may be even more diluted at the senior levels Thedilution of change may result in senior groups being still largely seen aswhite male having little change impact on diversity climate

Weak demographic shifts create weak group fault lines and consequentlymay do little to change or coalesce the social construction of group climateUnder such conditions intergroup resource competition theories are againlikely to be apt explanations of group processes As a minority group gainsstature and starts to become larger and more of a realistic competitor for scarceresources the majority is more likely to discriminate (Blalock 1967) Gradualproportionate increases in the representation of women and minorities at seniorlevels might lead to greater hostility by majority members because of resourcecontrol issues This would lead to more strained intergroup relations andreduced social interaction among members resulting in greater variation of theclimate for diversity (ie reduced consensus) and more negative perceptions ofclimate

H2 Increasing the proportion of senior women and minorities at thedepartment level may not necessarily positively relate to group valenceand consensus

2 Method21 Organizational settingThe data collection period for this study was designed to mirror the timeperiod of implementation of an organizational change strategy to managediversity at a large public sector university in the USA Over eight yearsthe administration had been actively engaged in HR strategies to fosterorganizational change in diversity These included active recruitment of adiverse workforce and wide dissemination of a document afregrming the

JOCM163

334

universityrsquos commitment to diversity National consulting regrms had beencontracted to provide advice on managing diversity The administrationmade funds readily available (sometimes over the initial salary posting) tohire quality minorities and women even for units in which new tenurestream positions had not been approved Training to promote diversitysensitivity among administrators was initiated Small cash achievementawards were distributed to recognize department activities that promotedmulticulturalism

Using the human information system the researchers collected facultydemographic data by department from records in the HR information system attwo points over an eight-year period to measure change in demography bydepartment We used a stratireged sample of tenure stream faculty with fullselection of the population of women and minorities and random sampling ofwhite males At the end of the eight years a survey was developed to examinethe success of the change efforts The archival data on proportional change indemography by group were used to form the independent variables (changefrom year 1 to year 8) and the survey data collected at the end of the eight yearswere used to form the group climate dependent variables The survey to assessthe success of these change strategies was developed with expert and facultyinput

22 MeasuresGroup dependent variables climate for diversity (consensus and valence) Fivegroup dependent variables were developed to assess the climate for diversityExploratory factor analysis ensured that each of these measures developed forthe study tapped into unique constructs Each of these measures utilized thesame regve-point scale ranging from 1 (strongly disagree) to 5 (strongly agree)Below we regrst describe the regve dependent variables and then we discuss howwe aggregated them to create group level consensus and valence measures

Diversity as an organizational goal We used Kossek and Zoniarsquos (1993) six-item scale (alpha 089) to assess the perceptions of the extent to which membersbelieved that organizational excellence and effective functioning were related tothe recruitment and retention of faculty who are female or minority Sampleitem ordfIf organization X is to remain an excellent institution it must recruit andretain more minority facultyordm

Commitment of management to diversity We developed a regve-item scale(alpha 085) to assess the administrationrsquos commitment to diversity through itssupport of efforts to increase faculty diversity through hiring and relatedpolicies Sample item ordfThe Dean of my college is strongly committed torecruiting more minority facultyordm

Perceived work-group mix We developed a regve-item scale (alpha 072) toassess the degree to which the department had a mix where women and

HRM changestrategy

335

minorities had good representation Sample item ordfMy department has a goodmix of men and women faculty membersordm

Equality of department support of women We used Kossek and Zoniarsquos(1993) three-item scale (alpha 071) to assess the degree to which faculty womenhad the same opportunity to have graduate assistants teaching release andabove average merit increases compared to males

Equality of department support of racial minorities We used Kossek andZoniarsquos (1993) three-item scale (alpha 077) to assess the degree to whichminority faculty had the same opportunity to have graduate assistantsteaching release and above average merit increases compared to non-minorities

Group consensus about climate for diversity These scales measured groupagreement for each of the regve dependent variables assessing climate Weutilized procedures outlined by Bleise and Halverson (1998) to ensure the validaggregation of individual data Group consensus was computed by

(1) determining the mean variance for each scale item per department

(2) calculating the sum of each scalersquos mean item variance

(3) dividing the sum (from step 2) by the number of items to obtain theaverage mean item variance for each scale

(4) multiplying the result from step 3 by plusmn1 to capture the level agreement

The closer the score is to 0 the higher the consensusGroup valence about the climate for diversity This scale assessed the median

score (a robust measure of central tendency) to assess the favorability of thegroup climate for the regve dependent variables Groups whose scores were at orabove the median have a more favorable climate for diversity than thosewhose scores were below the median

Group independent variables All of the independent variables were drawnfrom the universityrsquos HR information systems These included measures ofsize race gender and tenure (level) distribution by department at time 1 andtime 2 eight years later Variables were created to assess proportionaldemographic change for each group comparing the two points in time Forexample a grouprsquos proportion of women at time 1 was subtracted from theproportion of women at time 2 to measure proportional change

Change in department size was entered as a control variable since thedegree to which a department is growing may inmacruence member climateperceptions Drawing on Allportrsquos (1954) theory of social contact the moreopportunity an individual has to interact with members of other social groupsthe more likely they are to disconregrm individually held stereotypes Howeverwith increasing department size members may migrate toward others withfamiliar characteristics

JOCM163

336

3 Results31 Organizational structural demographic changeThe data show considerable success in increasing organizational diversity overthe eight-year hiring period At the organizational level the universityexperienced a 36 percent increase in female faculty from 437 at time 1 (baseline)to 644 at time 2 Similarly there was a 41 percent increase in the number ofracial minority faculty (from 211 at time 1 to 298 at time 2) Overall there were2511 total faculty at time 1 and 2684 at time 2 The net change was from 19percent at baseline to 24 percent eight years later for women and from 8percent to 11 percent for minorities Upward mobility was also improving Atbaseline the data show 9 percent women and 5 percent minority tenuredfaculty which increased to 11 percent (women) and 6 percent (minority) at time2 We took these data and transferred them into the measures described above

32 Survey response rateA total of 1529 individuals received the survey distributed at time 2 and 775were returned plusmn a 51 percent response rate For those surveys in whichdemographic identiregcation was provided analysis showed that response rateswere 47 percent for white men 46 percent for racial minority women 51 percentfor white women and 43 percent for racial minority men Of the 81 departmentsin the original sample 74 had at least three departmental respondents completethe diversity survey which were included in the regnal analysis Analyses weredone to ensure the respondents from the seven omitted departments did notstatistically differ from the 74 included Only tenure stream faculty wereincluded in the regnal dataset since there were relatively few non-tenure streamfaculty

33 Group level descriptive measures of diversity changeMeans standard deviations and correlations of measures are shown in Table IIn support of our regrst research objective Table I shows the descriptivestatistics for our new group level measures of diversity change These included

work group demographic shift

group climate (ie consensus and valence) and

group and organizational change referents

As the change variables show the average work group shift to increase inwomen was 1 percent with a standard deviation of 7 percent The averagework group increase in minorities was 1 percent with a standard deviationof 3 percent The group climate variables for the group referents (eg mixwell equality of department support of women equality of departmentsupport of minorities) and organizational referents (diversity is anorganizational goal commitment of management to diversity) are shownin Table I There was the greatest consensus that departments had equal

HRM changestrategy

337

Var

iable

Mea

nSD

12

34

5

1D

epar

tmen

tsi

zeT

ime

125

52

160

4plusmn

2D

epar

tmen

tsi

zeT

ime

229

06

176

60

89

plusmn3

Pro

por

tion

ofw

hit

esT

ime

10

900

010

210

21plusmn

4P

ropor

tion

ofw

hit

esT

ime

20

810

180

070

120

47

plusmn5

Pro

por

tion

ofw

omen

Tim

e1

019

021

005

002

001

017

plusmn6

Pro

por

tion

ofw

omen

Tim

e2

043

027

006

006

017

035

0

58

7P

ropor

tion

wit

hte

nure

Tim

e1

081

015

010

010

20

152

005

20

098

Pro

por

tion

wit

hT

enure

Tim

e2

084

011

004

003

20

082

018

20

129

Pro

por

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

034

019

20

042

009

20

100

000

76

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

039

018

20

092

010

20

39

20

130

65

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

247

550

40

062

0

59

20

092

013

009

12C

han

ge

inte

nure

tim

e1-

Tim

e2

001

001

20

152

013

20

28

20

120

0613

C

han

ge

ingen

der

Tim

e1-

Tim

e2

001

007

004

20

020

090

100

76

14C

han

ge

inw

hit

esT

ime

1-T

ime

20

010

032

009

20

162

059

2

005

2

008

15D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s2

100

057

001

001

20

192

015

024

16

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s2

099

048

007

014

20

032

006

008

17P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s2

125

047

001

011

002

20

060

1218

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s2

024

016

007

007

015

009

001

19E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s2

025

021

006

011

20

050

142

006

20D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n3

900

382

003

20

052

023

2

007

039

21

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n3

530

402

010

001

20

222

007

029

22P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

314

048

20

212

021

20

112

017

030

23E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n2

450

292

009

20

032

017

20

31

002

24E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

393

050

20

010

020

072

010

20

1625

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

361

049

20

030

012

018

20

050

2026

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

324

057

20

092

003

20

24

001

014

27P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n2

500

372

009

20

112

003

000

027

28

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

200

014

20

060

022

013

20

242

008

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n1

920

152

018

20

020

070

062

010

(con

tinued

)

Table IMeans standardsdeviations andcorrelations ofvariables

JOCM163

338

Var

iable

67

89

1011

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

plusmn7

Pro

por

tion

wit

hte

nure

Tim

e1

20

20plusmn

8P

ropor

tion

wit

hte

nure

Tim

e2

20

190

31plusmn

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

054

2

058

2

029

plusmn

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

048

2

018

20

32

073

plusmn

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

004

009

012

006

003

plusmn12

C

han

ge

inte

nure

Tim

e1-

Tim

e2

20

062

016

022

014

006

008

13C

han

ge

ingen

der

Tim

e1-

Tim

e2

032

2

019

20

130

68

055

0

0514

C

han

ge

inw

hit

esT

ime

1-T

ime

22

018

023

014

001

034

0

1315

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s0

010

020

080

140

110

0816

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s2

013

001

000

009

003

007

17P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s2

008

20

170

070

170

092

011

18E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s2

005

20

162

009

007

20

080

1519

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s2

011

011

008

20

102

023

004

20D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n0

57

20

022

014

036

0

48

014

21C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n0

030

182

001

016

026

0

0122

P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

007

20

040

050

32

030

0

0423

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n2

013

001

012

008

011

005

24E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

20

33

20

052

007

20

072

012

20

0225

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

048

0

022

004

020

027

0

1626

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

20

110

140

010

030

172

003

27P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n0

050

030

190

210

150

0628

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

20

092

004

022

20

012

002

004

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n2

026

2

018

20

102

007

20

162

024

(con

tinued

)

Table I

HRM changestrategy

339

Var

iable

1213

1415

1617

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

7P

ropor

tion

wit

hte

nure

Tim

e1

8P

ropor

tion

wit

hte

nure

Tim

e2

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

12C

han

ge

inte

nure

Tim

e1-

Tim

e2

plusmn13

C

han

ge

ingen

der

Tim

e1-

Tim

e2

001

plusmn14

C

han

ge

inw

hit

esT

ime

1-T

ime

20

162

007

plusmn15

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s0

150

110

14plusmn

16C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s2

013

021

006

20

07plusmn

17P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s0

150

130

060

38

034

plusmn

18E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s0

130

182

001

20

110

45

008

19E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s0

000

040

042

011

042

0

0020

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n0

000

050

150

34

20

32

20

0721

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n2

020

022

022

022

023

002

22P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

010

011

018

017

010

016

23E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n0

022

009

015

20

090

062

007

24E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

001

007

011

20

062

019

007

25D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

005

20

060

100

080

012

010

26C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

20

27

009

021

024

20

030

0227

P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n0

070

040

120

232

003

013

28E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

010

20

190

102

003

20

142

014

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n0

022

002

009

000

001

020

(con

tinued

)

Table I

JOCM163

340

Var

iable

1819

2021

2223

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

7P

ropor

tion

wit

hte

nure

Tim

e1

8P

ropor

tion

wit

hte

nure

Tim

e2

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

12C

han

ge

inte

nure

Tim

e1-

Tim

e2

13C

han

ge

ingen

der

Tim

e1-

Tim

e2

14C

han

ge

inw

hit

esT

ime

1-T

ime

215

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s16

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s17

P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s18

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

splusmn

19E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s0

57

plusmn20

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n2

035

2

027

plusmn

21C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n0

180

180

08plusmn

22P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

010

20

050

150

27

plusmn23

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n0

010

002

008

008

014

plusmn24

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

030

0

34

20

26

026

0

192

021

25D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

20

222

013

083

0

040

082

007

26C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

20

070

020

150

79

018

000

27P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n2

002

20

040

140

120

84

012

28E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

20

052

009

20

020

022

004

086

29

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n0

140

31

20

210

100

042

008

(con

tinued

)

Table I

HRM changestrategy

341

Table I

Var

iable

2425

2627

2829

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

7P

ropor

tion

wit

hte

nure

Tim

e1

8P

ropor

tion

wit

hte

nure

Tim

e2

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

12C

han

ge

inte

nure

Tim

e1-

Tim

e2

13C

han

ge

ingen

der

Tim

e1-

Tim

e2

14C

han

ge

inw

hit

esT

ime

1-T

ime

215

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s16

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s17

P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s18

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s19

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s20

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n21

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n22

P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

23E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n24

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

plusmn25

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

20

20plusmn

26C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

025

002

plusmn27

P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n0

080

010

19plusmn

28E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

20

32

000

20

010

05plusmn

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n0

58

20

240

180

132

003

plusmn

Note

p

001

orgre

ater

JOCM163

342

support of resources for women and men and whites and non-whites asthese measures hovered at plusmn024 and 2 025 or close to 0 which would betotal consensus There was the least consensus that minorities and majoritiesmixed well (plusmn125) with slightly more consensus that diversity should bepursued as organizational (plusmn10) The most positive valence was theperception of equality of department support for minorities (393) and thatdiversity should be pursued as a goal (39)

34 Regression results on demographic shift and climateTo address our second research objective which was to examine whetherwork groups with higher demographic shifts had more positive groupclimates for each of the regve dependent variables separate hierarchicalregressions were run to examine the relation between group demographicchange and group consensus (see Table II) or valence (see Table III) In theregrst step of all regressions proportional change in size was entered as acontrol For the consensus regressions the current mean of each dependentvariable was also entered as a control In step 2 the group proportionalchange in women and the group proportional change in minorities wereentered (In early analyses for the second step of each regression interactionterms change in gender or race by tenure were included following Aikenand Westrsquos (1993) procedures No signiregcant interactions were found) Usingstandardized beta weights as seen in step 2 of Tables II and III ourhypothesis that work groups with higher demographic shifts may havehigher consensus and valence toward diversity change received partialsupport Group referents of diversity climate were unchanged bydemographic shifts remacrecting possible dilution of HR strategiesOrganizational referents of valence were effected but consensus wasreduced in terms of whether the organization should pursue diversity as anorganizational goal Departments with higher shifts in the proportion ofwomen had lower consensus on whether the organization should pursuediversity as an organizational goal Both regressions with organizationalreferents as dependent variables were signiregcant Groups with higherincreases in the proportion of women over time tended to have morefavorable valence regarding the desirability of promoting diversity as anorganizational goal but were less sure of managementrsquos commitment asthey had more experience with the reality of implementing theorganizational change

35 Regression results on positive senior women and minority effect on climateOur third research objective was to examine whether increasing the proportionof senior women and minorities at the department level positively relates togroup valence and consensus In our data analysis the proportional change inoverall tenured faculty was entered in step 2 and then the proportional change

HRM changestrategy

343

Org

aniz

atio

nD

epar

tmen

tD

iver

sity

asor

gan

izat

ional

goa

l

Com

mit

men

tof

man

agem

ent

todiv

ersi

ty

Per

ceiv

edw

ork-g

roup

mix

Equal

ity

ofdep

artm

ent

suppor

tof

wom

enE

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

Ste

p1

Dep

tsi

zeT

ime

2-T

ime

10

020

080

180

000

06M

ean

ofD

V0

35

022

017

002

029

D

R2

012

0

060

060

000

09T

otal

R2

012

0

060

060

000

09Ste

p2

Wom

enT

ime

2-T

ime

12

035

2

017

20

182

007

20

02W

hit

esT

ime

2-T

ime

12

002

20

052

003

004

021

Ten

ure

Tim

e2-

Tim

e1

009

003

022

010

000

DR

20

11

003

008

001

004

Tot

alR

20

23

009

014

001

013

Ste

p3

Whit

ete

nure

dm

ales

Tim

e2-

Tim

e1

20

092

012

005

20

302

031

DR

20

010

010

000

040

05T

otal

R2

024

0

100

140

050

18

Note

p

005

orgre

ater

p0

01or

gre

ater

Table IIRegressionpredictingconsensus onclimate for diversity

JOCM163

344

Org

aniz

atio

nD

epar

tmen

tD

iver

sity

asor

gan

izat

ional

goa

l

Com

mit

men

tof

man

agem

ent

todiv

ersi

ty

Per

ceiv

edw

ork-g

roup

mix

Equal

ity

ofdep

artm

ent

suppor

tof

wom

enE

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

Ste

p1

Dep

tsi

zeT

ime

2-T

ime

10

080

152

006

021

0

05D

R2

001

002

000

004

0

00T

otal

R2

001

002

000

004

0

00Ste

p2

Wom

enT

ime

2-T

ime

10

40

20

38

20

21

20

062

013

Whit

esT

ime

2-T

ime

12

007

019

008

20

172

001

Ten

ure

Tim

e2-

Tim

e1

20

072

013

012

016

019

DR

20

16

017

0

060

070

05T

otal

R2

017

0

19

006

011

0

05Ste

p3

Whit

ete

nure

dm

ales

Tim

e2-

Tim

e1

005

011

000

007

035

D

R2

000

001

000

000

004

T

otal

R2

017

0

20

006

011

009

Note

p

005

orgre

ater

p0

01or

gre

ater

Table IIIRegression

predicting valenceof climate for

diversity

HRM changestrategy

345

in tenured white males was entered in step 3 The results are that the thirdblock was not statistically signiregcant in the regressions showing that smallincreases in the group proportions of senior women and minorities did notincrease group consensus or valence regarding climate for change

4 DiscussionThis study explored links between actual change in organizational structureorganizational demography and work group interpretation of this change Ourresults showed that HR strategies that focus on structural change withoutworking to develop supportive group norms and positive climate areinadequate change strategies Current employer efforts to increase diversity asan isolated HR strategy may not necessarily lead to increased memberagreement that change is favorable Most change efforts like the diversitychange effort we studied are focused at the organizational or individualintervention levels and under-emphasize the work group level and theimportance of group tipping points Our study highlights a disconnect betweenorganizational level diversity practice which often assumes that hiring moreminorities and women will improve climate and what our research found at thegroup level a climate that was mixed and ambivalent at best and sometimesnegative

41 Group level measures of demographic change consensus and valence ondiversity climateThe regrst objective of this study was to develop new theory and measures ofgroup level change in diversity We provided new ways to measure and groupsocial construction of change These included consensus the degree to whichmembers held common perceptions of diversity climate and valence thedirection of those perceptions Our measures identify group and organizationallevel referents of HR strategies to increase demographic diversity Group levelreferents of diversity change were how well members perceived their workgroup mix and the perceived equality of resource allocation acrossdemographic groups Organizational referents were managementcommitment to change and whether diversity should be an organizationalstrategy assessed general support for change Our study highlights howimportant it is for research and practice to not only examine how change indemography relates to differences in individual attitudes but also to groupmember shared perceptions about the direction of the climate for changeUnderstanding linkages between HR changes (in this case group demographywhich raises unique psychological processes related to social identity) andpositive consensus is important because it is relevant to cooperation workexperiences interpersonal conmacrict stress and norms in groups striving tomanage change

JOCM163

346

42 Greater work group demographic shift does not necessarily result infavorable climateOur second objective was to assess whether work groups with higherdemographic shifts tend to have more positive climates that is higherconsensus and valence toward diversity change for both organizational andgroup level change referents Our results only partially supported this beliefthat greater structural demographic shift would positively relate to a morefavorable climate for diversity Although at the organizational level theuniversity had experienced signiregcant change in organizational demographyan increase in the overall representation of white women (36 percent) andminorities (41 percent) over time work group members in units with thegreatest change did not have higher consensus or more favorable valenceWe found more favorable results for organizational than group levelreferents of change Members may regnd it easier to agree with organizationalreferents and more difregcult to agree on group level referents due to theirgreater psychological proximity to group processes that affect how theypersonally experience conmacrict and wellbeing from change in their dailywork Work groups with more demographic change had higher valence butlower consensus on whether the organization should pursue diversity as astrategic goal When organizational change strategies to increase diversityare enacted at the work group level the group demographic shift may bediluted across the organization to only slightly increase the representation ofwomen and minorities within actual groups Due to this dilution effectgroup membersrsquo social construction of the HR strategies may lack consensusor positive valence regarding the organizational diversity change strategiesIf the change is not large enough to create a critical mass to tiprepresentation of minorities to at least 35 percent according to Kanter (1977)negative intergroup dynamics such as increased resource competition andtokenism may still occur

43 Diluting diversity gradually increasing the proportion of senior womenminorities does not necessarily improve group climateOur third objective was to investigate whether increasing the proportion ofsenior women and minorities at the department level positively relates togroup valence and consensus There was no relation between increasing theproportion of senior women and minorities and group valence andconsensus Due to diluted saliency of organizational HR change strategies toincrease diversity at the work group level slight demographic shifts toproportionally increase the representation of senior women and minoritiesdid not relate to higher group valence or consensus This increase did littleto improve group climate and may have even hurt the climate in theshort run

HRM changestrategy

347

5 Implications for future research and practice51 Incremental structural change not sufregcient to improve climateWe hope this study will encourage researchers and practitioners interested indiversity change strategies to augment individual and organizational levels ofintervention to focus on work group climates supporting change Our resultshighlight some disparities between the underlying assumptions of HRstrategies to manage organizational change (increasing structural diversitywill be positive for climate) and reality at the department level (incrementalincreases did not improve climate) Organizational strategists implementingdiversity recruitment initiatives typically assume that change strategies (in thiscase hiring diversity) would improve the group social climate due to a positivechange in member attitudes based on increased group diversity and positivesocial contact Yet our study does not support this view Although theorganization we studied had had signiregcant success in enacting the formal HRstrategy plusmn increasing the representation of women and minorities at theorganizational level plusmn our research shows that gradually shifting thedemography of work groups may not necessarily lead to increased consensusor valence about the social climate within work groups For example it did notlead to improved climate regarding the perceived fairness of resourceallocation or good mixing in social interaction In fact the incremental increasein the proportion of minority groups lowered consensus and agreement in theshort term as subgroups sorted out shifting and socially ambiguous powerrelationships within work units These results suggest the importance ofmeasuring change initiatives impacts at the group level and having group levelinterventions such as team-building and group-focused resource allocation tosupport organizational level diversity strategies

52 Focus change strategies on tipping points within targeted work unitsFuture research should consider how HR strategies to manage organizationaldemographic change are socially constructed and enacted at the work grouplevel It is especially important to identify what constitutes a group leveltipping point or critical mass across hierarchical levels and differentdemographic groups to enable positive consensus supporting changePractice might regnd it more fruitful to focus future HR strategies to increasediversity through hiring to alter tipping points within targeted work groupsThis will dramatically alter the saliency of strategic organizationaldemographic change and intervene to shape the climate of speciregc units Ourstudy suggests that organizations may need demographic earthquakes withinwork groups to effect group consensus and valence on the climate regardingthe HR change (ie managing diversity) Dilution of change at the group levelmay ameliorate intended positive effects unless certain group level tippingpoints such as at least 35 percent of the work group are effected in each unit by

JOCM163

348

the HR strategy After eight years of hiring new faculty in our sample womenhovered at about 15 percent in 89 percent of groups and minorities had amaximum representation of 25 percent The demographic shifts whileincreasing were not dramatic enough to create imbalances that markedlyincreased the salience and power of new demographic subgroups created bythe HR strategies Instead of being balanced many of our groups were stilllargely white male and remain tilted or skewed over time which Kanter (1977)notes perpetuates stereotypical negative dynamics

53 Avoiding backlash weaving tipping points across the hierarchyRecent work at MIT by Bailyn (2000) suggests that backlash and resistance toorganizational change may be occurring among senior white male faculty inresponse to HR diversity strategies Initially women and minorities now enterdepartments with relatively equal resources to white men as supported by theHR system design However by the time women and minorities reach theassociate and full professor levels subtle discrimination processes occur in howHR policies are applied such as lower access to resources promotions andpoor climate Despite the increase in upward mobility the representation ofwomen and minorities at the top is still relatively low Ely (1994) found that ifwomen are under-represented at higher organizational levels relations betweenwomen at lower levels might be of lower quality due to increased competitionand negative gender dynamics in the larger social system If HR changestrategies that alter the distribution of existing resources across organizationaland demographic groups are to be successful they must take into account hownot only tipping points in terms of overall departmental representation butmore importantly how tipping points are distributed in demographies acrossthe hierarchies of work groups or departments

HR diversity change strategies that bring in many new hires of differentbackgrounds may have negative ramiregcations for social functioning in groupsparticularly if new members are not supported by additional HR strategies thatallow them to enter work groups on an equal footing or if their work groups donot have leadership that mirror their demographic identities As Kanter (1983)found identity groups must be equal in their access to power resources ifcontact is to produce improvement in attitudes There must also be time for thenew members to be socialized and integrated into the culture Harrison et al(1998) found that the longer the length of time that members worked togetherthe greater the effects of deep level attitude similarity Organizations areincreasing surface (demographic) and deep level (attitudinal) diversitysimultaneously which is likely to have negative short-term ramiregcations forthe consensus and direction of climate We argue HR strategies must not onlyfocus on the surface level plusmn remacrecting structural and demographic attributes plusmnbut also simultaneously to deep level characteristics remacrecting values within

HRM changestrategy

349

the context of speciregc groups at all organizational levels In sum HR strategistsand change agents should focus on how HR change strategies relate to the jointalignment of group memberrsquos multiple characteristics (eg demographichierarchical status task) in work groups plusmn that is to examine group contextualinmacruences (Wharton 1992)

54 LimitationsOur study while making contributions to organizational research is notwithout limitations Though large in individual respondents by utilizing thegroup as the level of analysis our sample size includes only 74 groups Thiscould reduce the level of statistical power of our regndings Additionally becausethe sample was based on one organization it is important to replicate thisanalysis across multiple organizations to substantiate our results Althoughchanges in department heterogeneity across an eight-year period wereincluded this is not a longitudinal design Future research should employrepeated measures over a length of time across all variables measuringdemographic change in heterogeneity as well as consensus and valence Ourstudy points to some of the challenges in measuring ordfsuccess in increasingdiversityordm as we did not look at annual turnover or the effects of the glassceiling at the group level in this study which is a gap we hope futureresearchers will be motivated to study spurred by our research

Since little published work has been done on group consensus regarding theeffects of HR policies to increase diversity through demographic change weextend many social science literatures ranging from organizationaldevelopment to demography and HR Future research on diversity andorganizational change also should strive to be interdisciplinary as theliterature on HR strategy and change has been largely developed in the HRliterature separate from other social science disciplines reducing theoreticalintegration Without such integration and increased scholarly and practitionerattention to climate and consensus many organizations may fail to supportgroup membersrsquo explicit use of their cultural experiences to advanceorganizational learning from enacted diversity change strategies despite thegroup and intergroup social tensions that will inevitably occur (Thomas andEly 1996)

References

Aiken LA and West SG (1993) Multiple Regression Testing and Interpreting InteractionsSage Publications Newbury Park CA

Alderfer CP and Smith KK (1982) ordfStudying embedded intergroup relations inorganizationsordm Administrative Science Quarterly Vol 27 pp 35-65

Allport G (1954) The Nature of Prejudice Addison Wesley Cambridge MA

Bailyn L (2000) ordfGender and diversity in organizations preconference presentationordm paperpresented at the National Academy of Management meeting Toronto

JOCM163

350

Blalock HM (1967) Toward a Theory of Minority Group Relations Wiley New York NY

Bleise PD and Halverson RR (1998) ordfGroup consensus and psychological well-being a largeregeld studyordm Journal of Applied Social Psychology Vol 28 pp 563-80

Brown KG Kozlowski SWJ and Hattrup K (1996) Theory Issues and Recommendations inConceptualizing Agreement as a Construct in Organizational Research The Search forConsensus Regarding Consensus National Academy of Management Cincinnati OH

Cox T Jr (1993) Cultural Diversity in Organizations Berrett-Koehler San Francisco CA

Ely R (1994) ordfThe effects of organizational demographics and social identity on relationshipsamong professional womenordm Administrative Science Quarterly Vol 39 pp 203-38

Ely R (1995) ordfThe power in demography womenrsquos social constructions of gender identity atworkordm Academy of Management Journal Vol 38 pp 589-634

Festinger L (1954) ordfA theory of social comparison processesordm Human Relations Vol 7pp 117-40

Forsyth DR (1990) Group Dynamics BrooksCole Publishing Paciregc Grove CA

Harrison DA Price KH and Bell MP (1998) ordfBeyond relational demography time and theeffects of surface and deep-level diversity on work group cohesionordm Academy ofManagement Journal Vol 41 pp 96-107

James K Lovato C and Khoo G (1994) ordfSocial identity correlates of minority workersrsquo healthordmAcademy of Management Journal Vol 37 p 383+

Kanter R (1977) ordfSome effects of proportions on group life skewed sex ratios and reactions totoken womenordm American Journal of Sociology Vol 82 pp 965-90

Kanter R (1983) Men and Women of the Corporation Basic Books New York NY

Kossek EE and Lobel SA (1996) ordfIntroduction transforming human resource systems tomanage diversityordm in Managing Diversity Human Resource Strategies for Transforming

the Workplace Blackwell Oxford pp 1-19

Kossek EE and Zonia SC (1993) ordfAssessing diversity climate a regeld study of reactions toemployer efforts to promote diversityordm Journal of Organizational Behavior Vol 14pp 61-81

Larkey LK (1996) ordfToward a theory of communicative interactions in culturally diverseworkgroupsordm Academy of Management Review Vol 21 pp 463-91

Lau DC and Murnighan JK (1998) ordfDemographic diversity and faultlines the compositionaldynamics of organizational groupsordm Academy of Management Review Vol 29 pp 325-40

Lindell MK and Brandt CJ (2000) ordfClimate quality and climate consensus as mediators of therelationship between organizational antecedents and outcomesordm Journal of AppliedPsychology Vol 85 pp 331-48

McGrath JE Berdahl JL and Arrow H (1995) ordfTraits expectations culture and clout thedynamics of diversity in work groupsordm in Jackson SE and Ruderman MN (Eds)Diversity in Work Teams Research Paradigms for a Changing Workplace AmericanPsychological Association Washington DC

Payne RL and Pugh DS (1976) ordfOrganizational structure and climateordm in Dunnette MD(Ed) Handbook of Industrial and Organizational Psychology Rand McNally Chicago ILpp 1125-73

Reichers A and Schneider B (1990) ordfClimate and culture an evolution of constructsordm inSchneider B (Ed) Organizational Climate and Culture Jossey-Bass San Francisco CA

HRM changestrategy

351

prevails Here the argument is made that the diversity change is not largeenough to lead to greater consensus and valence The redistribution ofintergroup power through resource reallocation to reduce inequality (Alderferand Smith 1982) is likely to be experienced negatively by both majority andminority group members In groups where demographic shifts are dilutedminority members may still experience negative psychological processes anddiscrimination Majority members may see the slight demographic shifts aschanging the status quo to negatively impact them as they would see this aslosing power

H1 Work groups with higher demographic shifts may not necessarily havemore positive climates that is higher consensus and valence towarddiversity change for organizational and group level change referents

13 Is there a positive senior women and minority proportional effect on changeWe turn to our third objective that investigates ordfDoes increasing theproportion of senior women and minorities at the department level positivelyrelate to favorable group valence and high consensusordm Growing evidencesuggests that the interaction of demography with hierarchical level is critical tounderstanding organizational change processes The manner in which HRstrategies to promote demographic change shifts group demography acrossstatus levels within groups can lead to differing group climates Thus thecurrent race and gender composition of two groups can be identical in collectiverace and sex composition but have markedly different social reactions todynamics due to the saliency of membersrsquo multiple identities in relation to otherkey employee backgrounds such as hierarchical level

By examining the interaction of level (ie status) with race and sexdemography we are able to investigate the theory that it is not just the changein overall proportional demographic representation that effects climate but thecombination of demographic and non-demographic attributes that mightinmacruence the group social construction of diversity strategies in amultiplicative manner (Ely 1995) Demographic variables such as race andgender usually assume greater importance when associated with thedifferences in status (Tsui et al 1992) For example higher status membersmay be more likely to withdraw when groups are regrst integrated by lowerstatus members (Harrison et al 1998) Research on how status relates to groupproportional representation suggests that HR strategies to simply increase theoverall proportion of a minority group do not necessarily result in improvedorganizational effectiveness and increased inclusion of under-representedgroups (Ely 1994 1995) In order to have favorable climate the minority groupalso must be well represented in senior positions

Yet the dilution of organizational strategies across group hierarchiesagain may hamper the achievement of positive change in the status of

HRM changestrategy

333

groups When organizational diversity objectives to increase the upwardmobility of women and minorities are regrst enacted across work groupsstudies show that most organizations are likely to increase the number ofsenior women and racial minorities by only a few in each group (Tsui andGutek 1999) Again relying on Kanterrsquos (1977) theory of tipping points aslight gradual increase may not be strong enough to improve climatecompared to work groups that experienced no change or a decrease inhierarchical representation Recent work by Lau and Murnighan (1998)builds on Kanterrsquos (1977) work and suggests that groups have fault lines orways to align demographic characteristics across level gender and raceIncreasing gender and racial diversity through diversity hiring strategiescan alter the demographic strength and alignment or fault lines of internalsubgroups (eg senior and junior management tenured and junior faculty) Ifthese alignments are weak and work groups do not experienceordfearthquakesordm like the notion of a shift to a critical mass a slight shiftwork group demographic may be even more diluted at the senior levels Thedilution of change may result in senior groups being still largely seen aswhite male having little change impact on diversity climate

Weak demographic shifts create weak group fault lines and consequentlymay do little to change or coalesce the social construction of group climateUnder such conditions intergroup resource competition theories are againlikely to be apt explanations of group processes As a minority group gainsstature and starts to become larger and more of a realistic competitor for scarceresources the majority is more likely to discriminate (Blalock 1967) Gradualproportionate increases in the representation of women and minorities at seniorlevels might lead to greater hostility by majority members because of resourcecontrol issues This would lead to more strained intergroup relations andreduced social interaction among members resulting in greater variation of theclimate for diversity (ie reduced consensus) and more negative perceptions ofclimate

H2 Increasing the proportion of senior women and minorities at thedepartment level may not necessarily positively relate to group valenceand consensus

2 Method21 Organizational settingThe data collection period for this study was designed to mirror the timeperiod of implementation of an organizational change strategy to managediversity at a large public sector university in the USA Over eight yearsthe administration had been actively engaged in HR strategies to fosterorganizational change in diversity These included active recruitment of adiverse workforce and wide dissemination of a document afregrming the

JOCM163

334

universityrsquos commitment to diversity National consulting regrms had beencontracted to provide advice on managing diversity The administrationmade funds readily available (sometimes over the initial salary posting) tohire quality minorities and women even for units in which new tenurestream positions had not been approved Training to promote diversitysensitivity among administrators was initiated Small cash achievementawards were distributed to recognize department activities that promotedmulticulturalism

Using the human information system the researchers collected facultydemographic data by department from records in the HR information system attwo points over an eight-year period to measure change in demography bydepartment We used a stratireged sample of tenure stream faculty with fullselection of the population of women and minorities and random sampling ofwhite males At the end of the eight years a survey was developed to examinethe success of the change efforts The archival data on proportional change indemography by group were used to form the independent variables (changefrom year 1 to year 8) and the survey data collected at the end of the eight yearswere used to form the group climate dependent variables The survey to assessthe success of these change strategies was developed with expert and facultyinput

22 MeasuresGroup dependent variables climate for diversity (consensus and valence) Fivegroup dependent variables were developed to assess the climate for diversityExploratory factor analysis ensured that each of these measures developed forthe study tapped into unique constructs Each of these measures utilized thesame regve-point scale ranging from 1 (strongly disagree) to 5 (strongly agree)Below we regrst describe the regve dependent variables and then we discuss howwe aggregated them to create group level consensus and valence measures

Diversity as an organizational goal We used Kossek and Zoniarsquos (1993) six-item scale (alpha 089) to assess the perceptions of the extent to which membersbelieved that organizational excellence and effective functioning were related tothe recruitment and retention of faculty who are female or minority Sampleitem ordfIf organization X is to remain an excellent institution it must recruit andretain more minority facultyordm

Commitment of management to diversity We developed a regve-item scale(alpha 085) to assess the administrationrsquos commitment to diversity through itssupport of efforts to increase faculty diversity through hiring and relatedpolicies Sample item ordfThe Dean of my college is strongly committed torecruiting more minority facultyordm

Perceived work-group mix We developed a regve-item scale (alpha 072) toassess the degree to which the department had a mix where women and

HRM changestrategy

335

minorities had good representation Sample item ordfMy department has a goodmix of men and women faculty membersordm

Equality of department support of women We used Kossek and Zoniarsquos(1993) three-item scale (alpha 071) to assess the degree to which faculty womenhad the same opportunity to have graduate assistants teaching release andabove average merit increases compared to males

Equality of department support of racial minorities We used Kossek andZoniarsquos (1993) three-item scale (alpha 077) to assess the degree to whichminority faculty had the same opportunity to have graduate assistantsteaching release and above average merit increases compared to non-minorities

Group consensus about climate for diversity These scales measured groupagreement for each of the regve dependent variables assessing climate Weutilized procedures outlined by Bleise and Halverson (1998) to ensure the validaggregation of individual data Group consensus was computed by

(1) determining the mean variance for each scale item per department

(2) calculating the sum of each scalersquos mean item variance

(3) dividing the sum (from step 2) by the number of items to obtain theaverage mean item variance for each scale

(4) multiplying the result from step 3 by plusmn1 to capture the level agreement

The closer the score is to 0 the higher the consensusGroup valence about the climate for diversity This scale assessed the median

score (a robust measure of central tendency) to assess the favorability of thegroup climate for the regve dependent variables Groups whose scores were at orabove the median have a more favorable climate for diversity than thosewhose scores were below the median

Group independent variables All of the independent variables were drawnfrom the universityrsquos HR information systems These included measures ofsize race gender and tenure (level) distribution by department at time 1 andtime 2 eight years later Variables were created to assess proportionaldemographic change for each group comparing the two points in time Forexample a grouprsquos proportion of women at time 1 was subtracted from theproportion of women at time 2 to measure proportional change

Change in department size was entered as a control variable since thedegree to which a department is growing may inmacruence member climateperceptions Drawing on Allportrsquos (1954) theory of social contact the moreopportunity an individual has to interact with members of other social groupsthe more likely they are to disconregrm individually held stereotypes Howeverwith increasing department size members may migrate toward others withfamiliar characteristics

JOCM163

336

3 Results31 Organizational structural demographic changeThe data show considerable success in increasing organizational diversity overthe eight-year hiring period At the organizational level the universityexperienced a 36 percent increase in female faculty from 437 at time 1 (baseline)to 644 at time 2 Similarly there was a 41 percent increase in the number ofracial minority faculty (from 211 at time 1 to 298 at time 2) Overall there were2511 total faculty at time 1 and 2684 at time 2 The net change was from 19percent at baseline to 24 percent eight years later for women and from 8percent to 11 percent for minorities Upward mobility was also improving Atbaseline the data show 9 percent women and 5 percent minority tenuredfaculty which increased to 11 percent (women) and 6 percent (minority) at time2 We took these data and transferred them into the measures described above

32 Survey response rateA total of 1529 individuals received the survey distributed at time 2 and 775were returned plusmn a 51 percent response rate For those surveys in whichdemographic identiregcation was provided analysis showed that response rateswere 47 percent for white men 46 percent for racial minority women 51 percentfor white women and 43 percent for racial minority men Of the 81 departmentsin the original sample 74 had at least three departmental respondents completethe diversity survey which were included in the regnal analysis Analyses weredone to ensure the respondents from the seven omitted departments did notstatistically differ from the 74 included Only tenure stream faculty wereincluded in the regnal dataset since there were relatively few non-tenure streamfaculty

33 Group level descriptive measures of diversity changeMeans standard deviations and correlations of measures are shown in Table IIn support of our regrst research objective Table I shows the descriptivestatistics for our new group level measures of diversity change These included

work group demographic shift

group climate (ie consensus and valence) and

group and organizational change referents

As the change variables show the average work group shift to increase inwomen was 1 percent with a standard deviation of 7 percent The averagework group increase in minorities was 1 percent with a standard deviationof 3 percent The group climate variables for the group referents (eg mixwell equality of department support of women equality of departmentsupport of minorities) and organizational referents (diversity is anorganizational goal commitment of management to diversity) are shownin Table I There was the greatest consensus that departments had equal

HRM changestrategy

337

Var

iable

Mea

nSD

12

34

5

1D

epar

tmen

tsi

zeT

ime

125

52

160

4plusmn

2D

epar

tmen

tsi

zeT

ime

229

06

176

60

89

plusmn3

Pro

por

tion

ofw

hit

esT

ime

10

900

010

210

21plusmn

4P

ropor

tion

ofw

hit

esT

ime

20

810

180

070

120

47

plusmn5

Pro

por

tion

ofw

omen

Tim

e1

019

021

005

002

001

017

plusmn6

Pro

por

tion

ofw

omen

Tim

e2

043

027

006

006

017

035

0

58

7P

ropor

tion

wit

hte

nure

Tim

e1

081

015

010

010

20

152

005

20

098

Pro

por

tion

wit

hT

enure

Tim

e2

084

011

004

003

20

082

018

20

129

Pro

por

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

034

019

20

042

009

20

100

000

76

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

039

018

20

092

010

20

39

20

130

65

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

247

550

40

062

0

59

20

092

013

009

12C

han

ge

inte

nure

tim

e1-

Tim

e2

001

001

20

152

013

20

28

20

120

0613

C

han

ge

ingen

der

Tim

e1-

Tim

e2

001

007

004

20

020

090

100

76

14C

han

ge

inw

hit

esT

ime

1-T

ime

20

010

032

009

20

162

059

2

005

2

008

15D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s2

100

057

001

001

20

192

015

024

16

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s2

099

048

007

014

20

032

006

008

17P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s2

125

047

001

011

002

20

060

1218

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s2

024

016

007

007

015

009

001

19E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s2

025

021

006

011

20

050

142

006

20D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n3

900

382

003

20

052

023

2

007

039

21

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n3

530

402

010

001

20

222

007

029

22P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

314

048

20

212

021

20

112

017

030

23E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n2

450

292

009

20

032

017

20

31

002

24E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

393

050

20

010

020

072

010

20

1625

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

361

049

20

030

012

018

20

050

2026

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

324

057

20

092

003

20

24

001

014

27P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n2

500

372

009

20

112

003

000

027

28

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

200

014

20

060

022

013

20

242

008

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n1

920

152

018

20

020

070

062

010

(con

tinued

)

Table IMeans standardsdeviations andcorrelations ofvariables

JOCM163

338

Var

iable

67

89

1011

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

plusmn7

Pro

por

tion

wit

hte

nure

Tim

e1

20

20plusmn

8P

ropor

tion

wit

hte

nure

Tim

e2

20

190

31plusmn

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

054

2

058

2

029

plusmn

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

048

2

018

20

32

073

plusmn

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

004

009

012

006

003

plusmn12

C

han

ge

inte

nure

Tim

e1-

Tim

e2

20

062

016

022

014

006

008

13C

han

ge

ingen

der

Tim

e1-

Tim

e2

032

2

019

20

130

68

055

0

0514

C

han

ge

inw

hit

esT

ime

1-T

ime

22

018

023

014

001

034

0

1315

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s0

010

020

080

140

110

0816

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s2

013

001

000

009

003

007

17P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s2

008

20

170

070

170

092

011

18E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s2

005

20

162

009

007

20

080

1519

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s2

011

011

008

20

102

023

004

20D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n0

57

20

022

014

036

0

48

014

21C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n0

030

182

001

016

026

0

0122

P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

007

20

040

050

32

030

0

0423

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n2

013

001

012

008

011

005

24E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

20

33

20

052

007

20

072

012

20

0225

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

048

0

022

004

020

027

0

1626

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

20

110

140

010

030

172

003

27P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n0

050

030

190

210

150

0628

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

20

092

004

022

20

012

002

004

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n2

026

2

018

20

102

007

20

162

024

(con

tinued

)

Table I

HRM changestrategy

339

Var

iable

1213

1415

1617

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

7P

ropor

tion

wit

hte

nure

Tim

e1

8P

ropor

tion

wit

hte

nure

Tim

e2

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

12C

han

ge

inte

nure

Tim

e1-

Tim

e2

plusmn13

C

han

ge

ingen

der

Tim

e1-

Tim

e2

001

plusmn14

C

han

ge

inw

hit

esT

ime

1-T

ime

20

162

007

plusmn15

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s0

150

110

14plusmn

16C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s2

013

021

006

20

07plusmn

17P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s0

150

130

060

38

034

plusmn

18E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s0

130

182

001

20

110

45

008

19E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s0

000

040

042

011

042

0

0020

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n0

000

050

150

34

20

32

20

0721

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n2

020

022

022

022

023

002

22P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

010

011

018

017

010

016

23E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n0

022

009

015

20

090

062

007

24E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

001

007

011

20

062

019

007

25D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

005

20

060

100

080

012

010

26C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

20

27

009

021

024

20

030

0227

P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n0

070

040

120

232

003

013

28E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

010

20

190

102

003

20

142

014

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n0

022

002

009

000

001

020

(con

tinued

)

Table I

JOCM163

340

Var

iable

1819

2021

2223

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

7P

ropor

tion

wit

hte

nure

Tim

e1

8P

ropor

tion

wit

hte

nure

Tim

e2

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

12C

han

ge

inte

nure

Tim

e1-

Tim

e2

13C

han

ge

ingen

der

Tim

e1-

Tim

e2

14C

han

ge

inw

hit

esT

ime

1-T

ime

215

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s16

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s17

P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s18

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

splusmn

19E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s0

57

plusmn20

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n2

035

2

027

plusmn

21C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n0

180

180

08plusmn

22P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

010

20

050

150

27

plusmn23

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n0

010

002

008

008

014

plusmn24

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

030

0

34

20

26

026

0

192

021

25D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

20

222

013

083

0

040

082

007

26C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

20

070

020

150

79

018

000

27P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n2

002

20

040

140

120

84

012

28E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

20

052

009

20

020

022

004

086

29

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n0

140

31

20

210

100

042

008

(con

tinued

)

Table I

HRM changestrategy

341

Table I

Var

iable

2425

2627

2829

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

7P

ropor

tion

wit

hte

nure

Tim

e1

8P

ropor

tion

wit

hte

nure

Tim

e2

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

12C

han

ge

inte

nure

Tim

e1-

Tim

e2

13C

han

ge

ingen

der

Tim

e1-

Tim

e2

14C

han

ge

inw

hit

esT

ime

1-T

ime

215

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s16

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s17

P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s18

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s19

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s20

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n21

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n22

P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

23E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n24

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

plusmn25

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

20

20plusmn

26C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

025

002

plusmn27

P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n0

080

010

19plusmn

28E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

20

32

000

20

010

05plusmn

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n0

58

20

240

180

132

003

plusmn

Note

p

001

orgre

ater

JOCM163

342

support of resources for women and men and whites and non-whites asthese measures hovered at plusmn024 and 2 025 or close to 0 which would betotal consensus There was the least consensus that minorities and majoritiesmixed well (plusmn125) with slightly more consensus that diversity should bepursued as organizational (plusmn10) The most positive valence was theperception of equality of department support for minorities (393) and thatdiversity should be pursued as a goal (39)

34 Regression results on demographic shift and climateTo address our second research objective which was to examine whetherwork groups with higher demographic shifts had more positive groupclimates for each of the regve dependent variables separate hierarchicalregressions were run to examine the relation between group demographicchange and group consensus (see Table II) or valence (see Table III) In theregrst step of all regressions proportional change in size was entered as acontrol For the consensus regressions the current mean of each dependentvariable was also entered as a control In step 2 the group proportionalchange in women and the group proportional change in minorities wereentered (In early analyses for the second step of each regression interactionterms change in gender or race by tenure were included following Aikenand Westrsquos (1993) procedures No signiregcant interactions were found) Usingstandardized beta weights as seen in step 2 of Tables II and III ourhypothesis that work groups with higher demographic shifts may havehigher consensus and valence toward diversity change received partialsupport Group referents of diversity climate were unchanged bydemographic shifts remacrecting possible dilution of HR strategiesOrganizational referents of valence were effected but consensus wasreduced in terms of whether the organization should pursue diversity as anorganizational goal Departments with higher shifts in the proportion ofwomen had lower consensus on whether the organization should pursuediversity as an organizational goal Both regressions with organizationalreferents as dependent variables were signiregcant Groups with higherincreases in the proportion of women over time tended to have morefavorable valence regarding the desirability of promoting diversity as anorganizational goal but were less sure of managementrsquos commitment asthey had more experience with the reality of implementing theorganizational change

35 Regression results on positive senior women and minority effect on climateOur third research objective was to examine whether increasing the proportionof senior women and minorities at the department level positively relates togroup valence and consensus In our data analysis the proportional change inoverall tenured faculty was entered in step 2 and then the proportional change

HRM changestrategy

343

Org

aniz

atio

nD

epar

tmen

tD

iver

sity

asor

gan

izat

ional

goa

l

Com

mit

men

tof

man

agem

ent

todiv

ersi

ty

Per

ceiv

edw

ork-g

roup

mix

Equal

ity

ofdep

artm

ent

suppor

tof

wom

enE

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

Ste

p1

Dep

tsi

zeT

ime

2-T

ime

10

020

080

180

000

06M

ean

ofD

V0

35

022

017

002

029

D

R2

012

0

060

060

000

09T

otal

R2

012

0

060

060

000

09Ste

p2

Wom

enT

ime

2-T

ime

12

035

2

017

20

182

007

20

02W

hit

esT

ime

2-T

ime

12

002

20

052

003

004

021

Ten

ure

Tim

e2-

Tim

e1

009

003

022

010

000

DR

20

11

003

008

001

004

Tot

alR

20

23

009

014

001

013

Ste

p3

Whit

ete

nure

dm

ales

Tim

e2-

Tim

e1

20

092

012

005

20

302

031

DR

20

010

010

000

040

05T

otal

R2

024

0

100

140

050

18

Note

p

005

orgre

ater

p0

01or

gre

ater

Table IIRegressionpredictingconsensus onclimate for diversity

JOCM163

344

Org

aniz

atio

nD

epar

tmen

tD

iver

sity

asor

gan

izat

ional

goa

l

Com

mit

men

tof

man

agem

ent

todiv

ersi

ty

Per

ceiv

edw

ork-g

roup

mix

Equal

ity

ofdep

artm

ent

suppor

tof

wom

enE

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

Ste

p1

Dep

tsi

zeT

ime

2-T

ime

10

080

152

006

021

0

05D

R2

001

002

000

004

0

00T

otal

R2

001

002

000

004

0

00Ste

p2

Wom

enT

ime

2-T

ime

10

40

20

38

20

21

20

062

013

Whit

esT

ime

2-T

ime

12

007

019

008

20

172

001

Ten

ure

Tim

e2-

Tim

e1

20

072

013

012

016

019

DR

20

16

017

0

060

070

05T

otal

R2

017

0

19

006

011

0

05Ste

p3

Whit

ete

nure

dm

ales

Tim

e2-

Tim

e1

005

011

000

007

035

D

R2

000

001

000

000

004

T

otal

R2

017

0

20

006

011

009

Note

p

005

orgre

ater

p0

01or

gre

ater

Table IIIRegression

predicting valenceof climate for

diversity

HRM changestrategy

345

in tenured white males was entered in step 3 The results are that the thirdblock was not statistically signiregcant in the regressions showing that smallincreases in the group proportions of senior women and minorities did notincrease group consensus or valence regarding climate for change

4 DiscussionThis study explored links between actual change in organizational structureorganizational demography and work group interpretation of this change Ourresults showed that HR strategies that focus on structural change withoutworking to develop supportive group norms and positive climate areinadequate change strategies Current employer efforts to increase diversity asan isolated HR strategy may not necessarily lead to increased memberagreement that change is favorable Most change efforts like the diversitychange effort we studied are focused at the organizational or individualintervention levels and under-emphasize the work group level and theimportance of group tipping points Our study highlights a disconnect betweenorganizational level diversity practice which often assumes that hiring moreminorities and women will improve climate and what our research found at thegroup level a climate that was mixed and ambivalent at best and sometimesnegative

41 Group level measures of demographic change consensus and valence ondiversity climateThe regrst objective of this study was to develop new theory and measures ofgroup level change in diversity We provided new ways to measure and groupsocial construction of change These included consensus the degree to whichmembers held common perceptions of diversity climate and valence thedirection of those perceptions Our measures identify group and organizationallevel referents of HR strategies to increase demographic diversity Group levelreferents of diversity change were how well members perceived their workgroup mix and the perceived equality of resource allocation acrossdemographic groups Organizational referents were managementcommitment to change and whether diversity should be an organizationalstrategy assessed general support for change Our study highlights howimportant it is for research and practice to not only examine how change indemography relates to differences in individual attitudes but also to groupmember shared perceptions about the direction of the climate for changeUnderstanding linkages between HR changes (in this case group demographywhich raises unique psychological processes related to social identity) andpositive consensus is important because it is relevant to cooperation workexperiences interpersonal conmacrict stress and norms in groups striving tomanage change

JOCM163

346

42 Greater work group demographic shift does not necessarily result infavorable climateOur second objective was to assess whether work groups with higherdemographic shifts tend to have more positive climates that is higherconsensus and valence toward diversity change for both organizational andgroup level change referents Our results only partially supported this beliefthat greater structural demographic shift would positively relate to a morefavorable climate for diversity Although at the organizational level theuniversity had experienced signiregcant change in organizational demographyan increase in the overall representation of white women (36 percent) andminorities (41 percent) over time work group members in units with thegreatest change did not have higher consensus or more favorable valenceWe found more favorable results for organizational than group levelreferents of change Members may regnd it easier to agree with organizationalreferents and more difregcult to agree on group level referents due to theirgreater psychological proximity to group processes that affect how theypersonally experience conmacrict and wellbeing from change in their dailywork Work groups with more demographic change had higher valence butlower consensus on whether the organization should pursue diversity as astrategic goal When organizational change strategies to increase diversityare enacted at the work group level the group demographic shift may bediluted across the organization to only slightly increase the representation ofwomen and minorities within actual groups Due to this dilution effectgroup membersrsquo social construction of the HR strategies may lack consensusor positive valence regarding the organizational diversity change strategiesIf the change is not large enough to create a critical mass to tiprepresentation of minorities to at least 35 percent according to Kanter (1977)negative intergroup dynamics such as increased resource competition andtokenism may still occur

43 Diluting diversity gradually increasing the proportion of senior womenminorities does not necessarily improve group climateOur third objective was to investigate whether increasing the proportion ofsenior women and minorities at the department level positively relates togroup valence and consensus There was no relation between increasing theproportion of senior women and minorities and group valence andconsensus Due to diluted saliency of organizational HR change strategies toincrease diversity at the work group level slight demographic shifts toproportionally increase the representation of senior women and minoritiesdid not relate to higher group valence or consensus This increase did littleto improve group climate and may have even hurt the climate in theshort run

HRM changestrategy

347

5 Implications for future research and practice51 Incremental structural change not sufregcient to improve climateWe hope this study will encourage researchers and practitioners interested indiversity change strategies to augment individual and organizational levels ofintervention to focus on work group climates supporting change Our resultshighlight some disparities between the underlying assumptions of HRstrategies to manage organizational change (increasing structural diversitywill be positive for climate) and reality at the department level (incrementalincreases did not improve climate) Organizational strategists implementingdiversity recruitment initiatives typically assume that change strategies (in thiscase hiring diversity) would improve the group social climate due to a positivechange in member attitudes based on increased group diversity and positivesocial contact Yet our study does not support this view Although theorganization we studied had had signiregcant success in enacting the formal HRstrategy plusmn increasing the representation of women and minorities at theorganizational level plusmn our research shows that gradually shifting thedemography of work groups may not necessarily lead to increased consensusor valence about the social climate within work groups For example it did notlead to improved climate regarding the perceived fairness of resourceallocation or good mixing in social interaction In fact the incremental increasein the proportion of minority groups lowered consensus and agreement in theshort term as subgroups sorted out shifting and socially ambiguous powerrelationships within work units These results suggest the importance ofmeasuring change initiatives impacts at the group level and having group levelinterventions such as team-building and group-focused resource allocation tosupport organizational level diversity strategies

52 Focus change strategies on tipping points within targeted work unitsFuture research should consider how HR strategies to manage organizationaldemographic change are socially constructed and enacted at the work grouplevel It is especially important to identify what constitutes a group leveltipping point or critical mass across hierarchical levels and differentdemographic groups to enable positive consensus supporting changePractice might regnd it more fruitful to focus future HR strategies to increasediversity through hiring to alter tipping points within targeted work groupsThis will dramatically alter the saliency of strategic organizationaldemographic change and intervene to shape the climate of speciregc units Ourstudy suggests that organizations may need demographic earthquakes withinwork groups to effect group consensus and valence on the climate regardingthe HR change (ie managing diversity) Dilution of change at the group levelmay ameliorate intended positive effects unless certain group level tippingpoints such as at least 35 percent of the work group are effected in each unit by

JOCM163

348

the HR strategy After eight years of hiring new faculty in our sample womenhovered at about 15 percent in 89 percent of groups and minorities had amaximum representation of 25 percent The demographic shifts whileincreasing were not dramatic enough to create imbalances that markedlyincreased the salience and power of new demographic subgroups created bythe HR strategies Instead of being balanced many of our groups were stilllargely white male and remain tilted or skewed over time which Kanter (1977)notes perpetuates stereotypical negative dynamics

53 Avoiding backlash weaving tipping points across the hierarchyRecent work at MIT by Bailyn (2000) suggests that backlash and resistance toorganizational change may be occurring among senior white male faculty inresponse to HR diversity strategies Initially women and minorities now enterdepartments with relatively equal resources to white men as supported by theHR system design However by the time women and minorities reach theassociate and full professor levels subtle discrimination processes occur in howHR policies are applied such as lower access to resources promotions andpoor climate Despite the increase in upward mobility the representation ofwomen and minorities at the top is still relatively low Ely (1994) found that ifwomen are under-represented at higher organizational levels relations betweenwomen at lower levels might be of lower quality due to increased competitionand negative gender dynamics in the larger social system If HR changestrategies that alter the distribution of existing resources across organizationaland demographic groups are to be successful they must take into account hownot only tipping points in terms of overall departmental representation butmore importantly how tipping points are distributed in demographies acrossthe hierarchies of work groups or departments

HR diversity change strategies that bring in many new hires of differentbackgrounds may have negative ramiregcations for social functioning in groupsparticularly if new members are not supported by additional HR strategies thatallow them to enter work groups on an equal footing or if their work groups donot have leadership that mirror their demographic identities As Kanter (1983)found identity groups must be equal in their access to power resources ifcontact is to produce improvement in attitudes There must also be time for thenew members to be socialized and integrated into the culture Harrison et al(1998) found that the longer the length of time that members worked togetherthe greater the effects of deep level attitude similarity Organizations areincreasing surface (demographic) and deep level (attitudinal) diversitysimultaneously which is likely to have negative short-term ramiregcations forthe consensus and direction of climate We argue HR strategies must not onlyfocus on the surface level plusmn remacrecting structural and demographic attributes plusmnbut also simultaneously to deep level characteristics remacrecting values within

HRM changestrategy

349

the context of speciregc groups at all organizational levels In sum HR strategistsand change agents should focus on how HR change strategies relate to the jointalignment of group memberrsquos multiple characteristics (eg demographichierarchical status task) in work groups plusmn that is to examine group contextualinmacruences (Wharton 1992)

54 LimitationsOur study while making contributions to organizational research is notwithout limitations Though large in individual respondents by utilizing thegroup as the level of analysis our sample size includes only 74 groups Thiscould reduce the level of statistical power of our regndings Additionally becausethe sample was based on one organization it is important to replicate thisanalysis across multiple organizations to substantiate our results Althoughchanges in department heterogeneity across an eight-year period wereincluded this is not a longitudinal design Future research should employrepeated measures over a length of time across all variables measuringdemographic change in heterogeneity as well as consensus and valence Ourstudy points to some of the challenges in measuring ordfsuccess in increasingdiversityordm as we did not look at annual turnover or the effects of the glassceiling at the group level in this study which is a gap we hope futureresearchers will be motivated to study spurred by our research

Since little published work has been done on group consensus regarding theeffects of HR policies to increase diversity through demographic change weextend many social science literatures ranging from organizationaldevelopment to demography and HR Future research on diversity andorganizational change also should strive to be interdisciplinary as theliterature on HR strategy and change has been largely developed in the HRliterature separate from other social science disciplines reducing theoreticalintegration Without such integration and increased scholarly and practitionerattention to climate and consensus many organizations may fail to supportgroup membersrsquo explicit use of their cultural experiences to advanceorganizational learning from enacted diversity change strategies despite thegroup and intergroup social tensions that will inevitably occur (Thomas andEly 1996)

References

Aiken LA and West SG (1993) Multiple Regression Testing and Interpreting InteractionsSage Publications Newbury Park CA

Alderfer CP and Smith KK (1982) ordfStudying embedded intergroup relations inorganizationsordm Administrative Science Quarterly Vol 27 pp 35-65

Allport G (1954) The Nature of Prejudice Addison Wesley Cambridge MA

Bailyn L (2000) ordfGender and diversity in organizations preconference presentationordm paperpresented at the National Academy of Management meeting Toronto

JOCM163

350

Blalock HM (1967) Toward a Theory of Minority Group Relations Wiley New York NY

Bleise PD and Halverson RR (1998) ordfGroup consensus and psychological well-being a largeregeld studyordm Journal of Applied Social Psychology Vol 28 pp 563-80

Brown KG Kozlowski SWJ and Hattrup K (1996) Theory Issues and Recommendations inConceptualizing Agreement as a Construct in Organizational Research The Search forConsensus Regarding Consensus National Academy of Management Cincinnati OH

Cox T Jr (1993) Cultural Diversity in Organizations Berrett-Koehler San Francisco CA

Ely R (1994) ordfThe effects of organizational demographics and social identity on relationshipsamong professional womenordm Administrative Science Quarterly Vol 39 pp 203-38

Ely R (1995) ordfThe power in demography womenrsquos social constructions of gender identity atworkordm Academy of Management Journal Vol 38 pp 589-634

Festinger L (1954) ordfA theory of social comparison processesordm Human Relations Vol 7pp 117-40

Forsyth DR (1990) Group Dynamics BrooksCole Publishing Paciregc Grove CA

Harrison DA Price KH and Bell MP (1998) ordfBeyond relational demography time and theeffects of surface and deep-level diversity on work group cohesionordm Academy ofManagement Journal Vol 41 pp 96-107

James K Lovato C and Khoo G (1994) ordfSocial identity correlates of minority workersrsquo healthordmAcademy of Management Journal Vol 37 p 383+

Kanter R (1977) ordfSome effects of proportions on group life skewed sex ratios and reactions totoken womenordm American Journal of Sociology Vol 82 pp 965-90

Kanter R (1983) Men and Women of the Corporation Basic Books New York NY

Kossek EE and Lobel SA (1996) ordfIntroduction transforming human resource systems tomanage diversityordm in Managing Diversity Human Resource Strategies for Transforming

the Workplace Blackwell Oxford pp 1-19

Kossek EE and Zonia SC (1993) ordfAssessing diversity climate a regeld study of reactions toemployer efforts to promote diversityordm Journal of Organizational Behavior Vol 14pp 61-81

Larkey LK (1996) ordfToward a theory of communicative interactions in culturally diverseworkgroupsordm Academy of Management Review Vol 21 pp 463-91

Lau DC and Murnighan JK (1998) ordfDemographic diversity and faultlines the compositionaldynamics of organizational groupsordm Academy of Management Review Vol 29 pp 325-40

Lindell MK and Brandt CJ (2000) ordfClimate quality and climate consensus as mediators of therelationship between organizational antecedents and outcomesordm Journal of AppliedPsychology Vol 85 pp 331-48

McGrath JE Berdahl JL and Arrow H (1995) ordfTraits expectations culture and clout thedynamics of diversity in work groupsordm in Jackson SE and Ruderman MN (Eds)Diversity in Work Teams Research Paradigms for a Changing Workplace AmericanPsychological Association Washington DC

Payne RL and Pugh DS (1976) ordfOrganizational structure and climateordm in Dunnette MD(Ed) Handbook of Industrial and Organizational Psychology Rand McNally Chicago ILpp 1125-73

Reichers A and Schneider B (1990) ordfClimate and culture an evolution of constructsordm inSchneider B (Ed) Organizational Climate and Culture Jossey-Bass San Francisco CA

HRM changestrategy

351

groups When organizational diversity objectives to increase the upwardmobility of women and minorities are regrst enacted across work groupsstudies show that most organizations are likely to increase the number ofsenior women and racial minorities by only a few in each group (Tsui andGutek 1999) Again relying on Kanterrsquos (1977) theory of tipping points aslight gradual increase may not be strong enough to improve climatecompared to work groups that experienced no change or a decrease inhierarchical representation Recent work by Lau and Murnighan (1998)builds on Kanterrsquos (1977) work and suggests that groups have fault lines orways to align demographic characteristics across level gender and raceIncreasing gender and racial diversity through diversity hiring strategiescan alter the demographic strength and alignment or fault lines of internalsubgroups (eg senior and junior management tenured and junior faculty) Ifthese alignments are weak and work groups do not experienceordfearthquakesordm like the notion of a shift to a critical mass a slight shiftwork group demographic may be even more diluted at the senior levels Thedilution of change may result in senior groups being still largely seen aswhite male having little change impact on diversity climate

Weak demographic shifts create weak group fault lines and consequentlymay do little to change or coalesce the social construction of group climateUnder such conditions intergroup resource competition theories are againlikely to be apt explanations of group processes As a minority group gainsstature and starts to become larger and more of a realistic competitor for scarceresources the majority is more likely to discriminate (Blalock 1967) Gradualproportionate increases in the representation of women and minorities at seniorlevels might lead to greater hostility by majority members because of resourcecontrol issues This would lead to more strained intergroup relations andreduced social interaction among members resulting in greater variation of theclimate for diversity (ie reduced consensus) and more negative perceptions ofclimate

H2 Increasing the proportion of senior women and minorities at thedepartment level may not necessarily positively relate to group valenceand consensus

2 Method21 Organizational settingThe data collection period for this study was designed to mirror the timeperiod of implementation of an organizational change strategy to managediversity at a large public sector university in the USA Over eight yearsthe administration had been actively engaged in HR strategies to fosterorganizational change in diversity These included active recruitment of adiverse workforce and wide dissemination of a document afregrming the

JOCM163

334

universityrsquos commitment to diversity National consulting regrms had beencontracted to provide advice on managing diversity The administrationmade funds readily available (sometimes over the initial salary posting) tohire quality minorities and women even for units in which new tenurestream positions had not been approved Training to promote diversitysensitivity among administrators was initiated Small cash achievementawards were distributed to recognize department activities that promotedmulticulturalism

Using the human information system the researchers collected facultydemographic data by department from records in the HR information system attwo points over an eight-year period to measure change in demography bydepartment We used a stratireged sample of tenure stream faculty with fullselection of the population of women and minorities and random sampling ofwhite males At the end of the eight years a survey was developed to examinethe success of the change efforts The archival data on proportional change indemography by group were used to form the independent variables (changefrom year 1 to year 8) and the survey data collected at the end of the eight yearswere used to form the group climate dependent variables The survey to assessthe success of these change strategies was developed with expert and facultyinput

22 MeasuresGroup dependent variables climate for diversity (consensus and valence) Fivegroup dependent variables were developed to assess the climate for diversityExploratory factor analysis ensured that each of these measures developed forthe study tapped into unique constructs Each of these measures utilized thesame regve-point scale ranging from 1 (strongly disagree) to 5 (strongly agree)Below we regrst describe the regve dependent variables and then we discuss howwe aggregated them to create group level consensus and valence measures

Diversity as an organizational goal We used Kossek and Zoniarsquos (1993) six-item scale (alpha 089) to assess the perceptions of the extent to which membersbelieved that organizational excellence and effective functioning were related tothe recruitment and retention of faculty who are female or minority Sampleitem ordfIf organization X is to remain an excellent institution it must recruit andretain more minority facultyordm

Commitment of management to diversity We developed a regve-item scale(alpha 085) to assess the administrationrsquos commitment to diversity through itssupport of efforts to increase faculty diversity through hiring and relatedpolicies Sample item ordfThe Dean of my college is strongly committed torecruiting more minority facultyordm

Perceived work-group mix We developed a regve-item scale (alpha 072) toassess the degree to which the department had a mix where women and

HRM changestrategy

335

minorities had good representation Sample item ordfMy department has a goodmix of men and women faculty membersordm

Equality of department support of women We used Kossek and Zoniarsquos(1993) three-item scale (alpha 071) to assess the degree to which faculty womenhad the same opportunity to have graduate assistants teaching release andabove average merit increases compared to males

Equality of department support of racial minorities We used Kossek andZoniarsquos (1993) three-item scale (alpha 077) to assess the degree to whichminority faculty had the same opportunity to have graduate assistantsteaching release and above average merit increases compared to non-minorities

Group consensus about climate for diversity These scales measured groupagreement for each of the regve dependent variables assessing climate Weutilized procedures outlined by Bleise and Halverson (1998) to ensure the validaggregation of individual data Group consensus was computed by

(1) determining the mean variance for each scale item per department

(2) calculating the sum of each scalersquos mean item variance

(3) dividing the sum (from step 2) by the number of items to obtain theaverage mean item variance for each scale

(4) multiplying the result from step 3 by plusmn1 to capture the level agreement

The closer the score is to 0 the higher the consensusGroup valence about the climate for diversity This scale assessed the median

score (a robust measure of central tendency) to assess the favorability of thegroup climate for the regve dependent variables Groups whose scores were at orabove the median have a more favorable climate for diversity than thosewhose scores were below the median

Group independent variables All of the independent variables were drawnfrom the universityrsquos HR information systems These included measures ofsize race gender and tenure (level) distribution by department at time 1 andtime 2 eight years later Variables were created to assess proportionaldemographic change for each group comparing the two points in time Forexample a grouprsquos proportion of women at time 1 was subtracted from theproportion of women at time 2 to measure proportional change

Change in department size was entered as a control variable since thedegree to which a department is growing may inmacruence member climateperceptions Drawing on Allportrsquos (1954) theory of social contact the moreopportunity an individual has to interact with members of other social groupsthe more likely they are to disconregrm individually held stereotypes Howeverwith increasing department size members may migrate toward others withfamiliar characteristics

JOCM163

336

3 Results31 Organizational structural demographic changeThe data show considerable success in increasing organizational diversity overthe eight-year hiring period At the organizational level the universityexperienced a 36 percent increase in female faculty from 437 at time 1 (baseline)to 644 at time 2 Similarly there was a 41 percent increase in the number ofracial minority faculty (from 211 at time 1 to 298 at time 2) Overall there were2511 total faculty at time 1 and 2684 at time 2 The net change was from 19percent at baseline to 24 percent eight years later for women and from 8percent to 11 percent for minorities Upward mobility was also improving Atbaseline the data show 9 percent women and 5 percent minority tenuredfaculty which increased to 11 percent (women) and 6 percent (minority) at time2 We took these data and transferred them into the measures described above

32 Survey response rateA total of 1529 individuals received the survey distributed at time 2 and 775were returned plusmn a 51 percent response rate For those surveys in whichdemographic identiregcation was provided analysis showed that response rateswere 47 percent for white men 46 percent for racial minority women 51 percentfor white women and 43 percent for racial minority men Of the 81 departmentsin the original sample 74 had at least three departmental respondents completethe diversity survey which were included in the regnal analysis Analyses weredone to ensure the respondents from the seven omitted departments did notstatistically differ from the 74 included Only tenure stream faculty wereincluded in the regnal dataset since there were relatively few non-tenure streamfaculty

33 Group level descriptive measures of diversity changeMeans standard deviations and correlations of measures are shown in Table IIn support of our regrst research objective Table I shows the descriptivestatistics for our new group level measures of diversity change These included

work group demographic shift

group climate (ie consensus and valence) and

group and organizational change referents

As the change variables show the average work group shift to increase inwomen was 1 percent with a standard deviation of 7 percent The averagework group increase in minorities was 1 percent with a standard deviationof 3 percent The group climate variables for the group referents (eg mixwell equality of department support of women equality of departmentsupport of minorities) and organizational referents (diversity is anorganizational goal commitment of management to diversity) are shownin Table I There was the greatest consensus that departments had equal

HRM changestrategy

337

Var

iable

Mea

nSD

12

34

5

1D

epar

tmen

tsi

zeT

ime

125

52

160

4plusmn

2D

epar

tmen

tsi

zeT

ime

229

06

176

60

89

plusmn3

Pro

por

tion

ofw

hit

esT

ime

10

900

010

210

21plusmn

4P

ropor

tion

ofw

hit

esT

ime

20

810

180

070

120

47

plusmn5

Pro

por

tion

ofw

omen

Tim

e1

019

021

005

002

001

017

plusmn6

Pro

por

tion

ofw

omen

Tim

e2

043

027

006

006

017

035

0

58

7P

ropor

tion

wit

hte

nure

Tim

e1

081

015

010

010

20

152

005

20

098

Pro

por

tion

wit

hT

enure

Tim

e2

084

011

004

003

20

082

018

20

129

Pro

por

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

034

019

20

042

009

20

100

000

76

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

039

018

20

092

010

20

39

20

130

65

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

247

550

40

062

0

59

20

092

013

009

12C

han

ge

inte

nure

tim

e1-

Tim

e2

001

001

20

152

013

20

28

20

120

0613

C

han

ge

ingen

der

Tim

e1-

Tim

e2

001

007

004

20

020

090

100

76

14C

han

ge

inw

hit

esT

ime

1-T

ime

20

010

032

009

20

162

059

2

005

2

008

15D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s2

100

057

001

001

20

192

015

024

16

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s2

099

048

007

014

20

032

006

008

17P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s2

125

047

001

011

002

20

060

1218

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s2

024

016

007

007

015

009

001

19E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s2

025

021

006

011

20

050

142

006

20D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n3

900

382

003

20

052

023

2

007

039

21

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n3

530

402

010

001

20

222

007

029

22P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

314

048

20

212

021

20

112

017

030

23E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n2

450

292

009

20

032

017

20

31

002

24E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

393

050

20

010

020

072

010

20

1625

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

361

049

20

030

012

018

20

050

2026

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

324

057

20

092

003

20

24

001

014

27P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n2

500

372

009

20

112

003

000

027

28

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

200

014

20

060

022

013

20

242

008

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n1

920

152

018

20

020

070

062

010

(con

tinued

)

Table IMeans standardsdeviations andcorrelations ofvariables

JOCM163

338

Var

iable

67

89

1011

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

plusmn7

Pro

por

tion

wit

hte

nure

Tim

e1

20

20plusmn

8P

ropor

tion

wit

hte

nure

Tim

e2

20

190

31plusmn

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

054

2

058

2

029

plusmn

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

048

2

018

20

32

073

plusmn

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

004

009

012

006

003

plusmn12

C

han

ge

inte

nure

Tim

e1-

Tim

e2

20

062

016

022

014

006

008

13C

han

ge

ingen

der

Tim

e1-

Tim

e2

032

2

019

20

130

68

055

0

0514

C

han

ge

inw

hit

esT

ime

1-T

ime

22

018

023

014

001

034

0

1315

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s0

010

020

080

140

110

0816

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s2

013

001

000

009

003

007

17P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s2

008

20

170

070

170

092

011

18E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s2

005

20

162

009

007

20

080

1519

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s2

011

011

008

20

102

023

004

20D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n0

57

20

022

014

036

0

48

014

21C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n0

030

182

001

016

026

0

0122

P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

007

20

040

050

32

030

0

0423

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n2

013

001

012

008

011

005

24E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

20

33

20

052

007

20

072

012

20

0225

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

048

0

022

004

020

027

0

1626

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

20

110

140

010

030

172

003

27P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n0

050

030

190

210

150

0628

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

20

092

004

022

20

012

002

004

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n2

026

2

018

20

102

007

20

162

024

(con

tinued

)

Table I

HRM changestrategy

339

Var

iable

1213

1415

1617

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

7P

ropor

tion

wit

hte

nure

Tim

e1

8P

ropor

tion

wit

hte

nure

Tim

e2

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

12C

han

ge

inte

nure

Tim

e1-

Tim

e2

plusmn13

C

han

ge

ingen

der

Tim

e1-

Tim

e2

001

plusmn14

C

han

ge

inw

hit

esT

ime

1-T

ime

20

162

007

plusmn15

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s0

150

110

14plusmn

16C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s2

013

021

006

20

07plusmn

17P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s0

150

130

060

38

034

plusmn

18E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s0

130

182

001

20

110

45

008

19E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s0

000

040

042

011

042

0

0020

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n0

000

050

150

34

20

32

20

0721

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n2

020

022

022

022

023

002

22P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

010

011

018

017

010

016

23E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n0

022

009

015

20

090

062

007

24E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

001

007

011

20

062

019

007

25D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

005

20

060

100

080

012

010

26C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

20

27

009

021

024

20

030

0227

P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n0

070

040

120

232

003

013

28E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

010

20

190

102

003

20

142

014

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n0

022

002

009

000

001

020

(con

tinued

)

Table I

JOCM163

340

Var

iable

1819

2021

2223

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

7P

ropor

tion

wit

hte

nure

Tim

e1

8P

ropor

tion

wit

hte

nure

Tim

e2

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

12C

han

ge

inte

nure

Tim

e1-

Tim

e2

13C

han

ge

ingen

der

Tim

e1-

Tim

e2

14C

han

ge

inw

hit

esT

ime

1-T

ime

215

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s16

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s17

P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s18

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

splusmn

19E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s0

57

plusmn20

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n2

035

2

027

plusmn

21C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n0

180

180

08plusmn

22P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

010

20

050

150

27

plusmn23

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n0

010

002

008

008

014

plusmn24

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

030

0

34

20

26

026

0

192

021

25D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

20

222

013

083

0

040

082

007

26C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

20

070

020

150

79

018

000

27P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n2

002

20

040

140

120

84

012

28E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

20

052

009

20

020

022

004

086

29

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n0

140

31

20

210

100

042

008

(con

tinued

)

Table I

HRM changestrategy

341

Table I

Var

iable

2425

2627

2829

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

7P

ropor

tion

wit

hte

nure

Tim

e1

8P

ropor

tion

wit

hte

nure

Tim

e2

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

12C

han

ge

inte

nure

Tim

e1-

Tim

e2

13C

han

ge

ingen

der

Tim

e1-

Tim

e2

14C

han

ge

inw

hit

esT

ime

1-T

ime

215

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s16

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s17

P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s18

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s19

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s20

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n21

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n22

P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

23E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n24

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

plusmn25

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

20

20plusmn

26C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

025

002

plusmn27

P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n0

080

010

19plusmn

28E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

20

32

000

20

010

05plusmn

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n0

58

20

240

180

132

003

plusmn

Note

p

001

orgre

ater

JOCM163

342

support of resources for women and men and whites and non-whites asthese measures hovered at plusmn024 and 2 025 or close to 0 which would betotal consensus There was the least consensus that minorities and majoritiesmixed well (plusmn125) with slightly more consensus that diversity should bepursued as organizational (plusmn10) The most positive valence was theperception of equality of department support for minorities (393) and thatdiversity should be pursued as a goal (39)

34 Regression results on demographic shift and climateTo address our second research objective which was to examine whetherwork groups with higher demographic shifts had more positive groupclimates for each of the regve dependent variables separate hierarchicalregressions were run to examine the relation between group demographicchange and group consensus (see Table II) or valence (see Table III) In theregrst step of all regressions proportional change in size was entered as acontrol For the consensus regressions the current mean of each dependentvariable was also entered as a control In step 2 the group proportionalchange in women and the group proportional change in minorities wereentered (In early analyses for the second step of each regression interactionterms change in gender or race by tenure were included following Aikenand Westrsquos (1993) procedures No signiregcant interactions were found) Usingstandardized beta weights as seen in step 2 of Tables II and III ourhypothesis that work groups with higher demographic shifts may havehigher consensus and valence toward diversity change received partialsupport Group referents of diversity climate were unchanged bydemographic shifts remacrecting possible dilution of HR strategiesOrganizational referents of valence were effected but consensus wasreduced in terms of whether the organization should pursue diversity as anorganizational goal Departments with higher shifts in the proportion ofwomen had lower consensus on whether the organization should pursuediversity as an organizational goal Both regressions with organizationalreferents as dependent variables were signiregcant Groups with higherincreases in the proportion of women over time tended to have morefavorable valence regarding the desirability of promoting diversity as anorganizational goal but were less sure of managementrsquos commitment asthey had more experience with the reality of implementing theorganizational change

35 Regression results on positive senior women and minority effect on climateOur third research objective was to examine whether increasing the proportionof senior women and minorities at the department level positively relates togroup valence and consensus In our data analysis the proportional change inoverall tenured faculty was entered in step 2 and then the proportional change

HRM changestrategy

343

Org

aniz

atio

nD

epar

tmen

tD

iver

sity

asor

gan

izat

ional

goa

l

Com

mit

men

tof

man

agem

ent

todiv

ersi

ty

Per

ceiv

edw

ork-g

roup

mix

Equal

ity

ofdep

artm

ent

suppor

tof

wom

enE

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

Ste

p1

Dep

tsi

zeT

ime

2-T

ime

10

020

080

180

000

06M

ean

ofD

V0

35

022

017

002

029

D

R2

012

0

060

060

000

09T

otal

R2

012

0

060

060

000

09Ste

p2

Wom

enT

ime

2-T

ime

12

035

2

017

20

182

007

20

02W

hit

esT

ime

2-T

ime

12

002

20

052

003

004

021

Ten

ure

Tim

e2-

Tim

e1

009

003

022

010

000

DR

20

11

003

008

001

004

Tot

alR

20

23

009

014

001

013

Ste

p3

Whit

ete

nure

dm

ales

Tim

e2-

Tim

e1

20

092

012

005

20

302

031

DR

20

010

010

000

040

05T

otal

R2

024

0

100

140

050

18

Note

p

005

orgre

ater

p0

01or

gre

ater

Table IIRegressionpredictingconsensus onclimate for diversity

JOCM163

344

Org

aniz

atio

nD

epar

tmen

tD

iver

sity

asor

gan

izat

ional

goa

l

Com

mit

men

tof

man

agem

ent

todiv

ersi

ty

Per

ceiv

edw

ork-g

roup

mix

Equal

ity

ofdep

artm

ent

suppor

tof

wom

enE

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

Ste

p1

Dep

tsi

zeT

ime

2-T

ime

10

080

152

006

021

0

05D

R2

001

002

000

004

0

00T

otal

R2

001

002

000

004

0

00Ste

p2

Wom

enT

ime

2-T

ime

10

40

20

38

20

21

20

062

013

Whit

esT

ime

2-T

ime

12

007

019

008

20

172

001

Ten

ure

Tim

e2-

Tim

e1

20

072

013

012

016

019

DR

20

16

017

0

060

070

05T

otal

R2

017

0

19

006

011

0

05Ste

p3

Whit

ete

nure

dm

ales

Tim

e2-

Tim

e1

005

011

000

007

035

D

R2

000

001

000

000

004

T

otal

R2

017

0

20

006

011

009

Note

p

005

orgre

ater

p0

01or

gre

ater

Table IIIRegression

predicting valenceof climate for

diversity

HRM changestrategy

345

in tenured white males was entered in step 3 The results are that the thirdblock was not statistically signiregcant in the regressions showing that smallincreases in the group proportions of senior women and minorities did notincrease group consensus or valence regarding climate for change

4 DiscussionThis study explored links between actual change in organizational structureorganizational demography and work group interpretation of this change Ourresults showed that HR strategies that focus on structural change withoutworking to develop supportive group norms and positive climate areinadequate change strategies Current employer efforts to increase diversity asan isolated HR strategy may not necessarily lead to increased memberagreement that change is favorable Most change efforts like the diversitychange effort we studied are focused at the organizational or individualintervention levels and under-emphasize the work group level and theimportance of group tipping points Our study highlights a disconnect betweenorganizational level diversity practice which often assumes that hiring moreminorities and women will improve climate and what our research found at thegroup level a climate that was mixed and ambivalent at best and sometimesnegative

41 Group level measures of demographic change consensus and valence ondiversity climateThe regrst objective of this study was to develop new theory and measures ofgroup level change in diversity We provided new ways to measure and groupsocial construction of change These included consensus the degree to whichmembers held common perceptions of diversity climate and valence thedirection of those perceptions Our measures identify group and organizationallevel referents of HR strategies to increase demographic diversity Group levelreferents of diversity change were how well members perceived their workgroup mix and the perceived equality of resource allocation acrossdemographic groups Organizational referents were managementcommitment to change and whether diversity should be an organizationalstrategy assessed general support for change Our study highlights howimportant it is for research and practice to not only examine how change indemography relates to differences in individual attitudes but also to groupmember shared perceptions about the direction of the climate for changeUnderstanding linkages between HR changes (in this case group demographywhich raises unique psychological processes related to social identity) andpositive consensus is important because it is relevant to cooperation workexperiences interpersonal conmacrict stress and norms in groups striving tomanage change

JOCM163

346

42 Greater work group demographic shift does not necessarily result infavorable climateOur second objective was to assess whether work groups with higherdemographic shifts tend to have more positive climates that is higherconsensus and valence toward diversity change for both organizational andgroup level change referents Our results only partially supported this beliefthat greater structural demographic shift would positively relate to a morefavorable climate for diversity Although at the organizational level theuniversity had experienced signiregcant change in organizational demographyan increase in the overall representation of white women (36 percent) andminorities (41 percent) over time work group members in units with thegreatest change did not have higher consensus or more favorable valenceWe found more favorable results for organizational than group levelreferents of change Members may regnd it easier to agree with organizationalreferents and more difregcult to agree on group level referents due to theirgreater psychological proximity to group processes that affect how theypersonally experience conmacrict and wellbeing from change in their dailywork Work groups with more demographic change had higher valence butlower consensus on whether the organization should pursue diversity as astrategic goal When organizational change strategies to increase diversityare enacted at the work group level the group demographic shift may bediluted across the organization to only slightly increase the representation ofwomen and minorities within actual groups Due to this dilution effectgroup membersrsquo social construction of the HR strategies may lack consensusor positive valence regarding the organizational diversity change strategiesIf the change is not large enough to create a critical mass to tiprepresentation of minorities to at least 35 percent according to Kanter (1977)negative intergroup dynamics such as increased resource competition andtokenism may still occur

43 Diluting diversity gradually increasing the proportion of senior womenminorities does not necessarily improve group climateOur third objective was to investigate whether increasing the proportion ofsenior women and minorities at the department level positively relates togroup valence and consensus There was no relation between increasing theproportion of senior women and minorities and group valence andconsensus Due to diluted saliency of organizational HR change strategies toincrease diversity at the work group level slight demographic shifts toproportionally increase the representation of senior women and minoritiesdid not relate to higher group valence or consensus This increase did littleto improve group climate and may have even hurt the climate in theshort run

HRM changestrategy

347

5 Implications for future research and practice51 Incremental structural change not sufregcient to improve climateWe hope this study will encourage researchers and practitioners interested indiversity change strategies to augment individual and organizational levels ofintervention to focus on work group climates supporting change Our resultshighlight some disparities between the underlying assumptions of HRstrategies to manage organizational change (increasing structural diversitywill be positive for climate) and reality at the department level (incrementalincreases did not improve climate) Organizational strategists implementingdiversity recruitment initiatives typically assume that change strategies (in thiscase hiring diversity) would improve the group social climate due to a positivechange in member attitudes based on increased group diversity and positivesocial contact Yet our study does not support this view Although theorganization we studied had had signiregcant success in enacting the formal HRstrategy plusmn increasing the representation of women and minorities at theorganizational level plusmn our research shows that gradually shifting thedemography of work groups may not necessarily lead to increased consensusor valence about the social climate within work groups For example it did notlead to improved climate regarding the perceived fairness of resourceallocation or good mixing in social interaction In fact the incremental increasein the proportion of minority groups lowered consensus and agreement in theshort term as subgroups sorted out shifting and socially ambiguous powerrelationships within work units These results suggest the importance ofmeasuring change initiatives impacts at the group level and having group levelinterventions such as team-building and group-focused resource allocation tosupport organizational level diversity strategies

52 Focus change strategies on tipping points within targeted work unitsFuture research should consider how HR strategies to manage organizationaldemographic change are socially constructed and enacted at the work grouplevel It is especially important to identify what constitutes a group leveltipping point or critical mass across hierarchical levels and differentdemographic groups to enable positive consensus supporting changePractice might regnd it more fruitful to focus future HR strategies to increasediversity through hiring to alter tipping points within targeted work groupsThis will dramatically alter the saliency of strategic organizationaldemographic change and intervene to shape the climate of speciregc units Ourstudy suggests that organizations may need demographic earthquakes withinwork groups to effect group consensus and valence on the climate regardingthe HR change (ie managing diversity) Dilution of change at the group levelmay ameliorate intended positive effects unless certain group level tippingpoints such as at least 35 percent of the work group are effected in each unit by

JOCM163

348

the HR strategy After eight years of hiring new faculty in our sample womenhovered at about 15 percent in 89 percent of groups and minorities had amaximum representation of 25 percent The demographic shifts whileincreasing were not dramatic enough to create imbalances that markedlyincreased the salience and power of new demographic subgroups created bythe HR strategies Instead of being balanced many of our groups were stilllargely white male and remain tilted or skewed over time which Kanter (1977)notes perpetuates stereotypical negative dynamics

53 Avoiding backlash weaving tipping points across the hierarchyRecent work at MIT by Bailyn (2000) suggests that backlash and resistance toorganizational change may be occurring among senior white male faculty inresponse to HR diversity strategies Initially women and minorities now enterdepartments with relatively equal resources to white men as supported by theHR system design However by the time women and minorities reach theassociate and full professor levels subtle discrimination processes occur in howHR policies are applied such as lower access to resources promotions andpoor climate Despite the increase in upward mobility the representation ofwomen and minorities at the top is still relatively low Ely (1994) found that ifwomen are under-represented at higher organizational levels relations betweenwomen at lower levels might be of lower quality due to increased competitionand negative gender dynamics in the larger social system If HR changestrategies that alter the distribution of existing resources across organizationaland demographic groups are to be successful they must take into account hownot only tipping points in terms of overall departmental representation butmore importantly how tipping points are distributed in demographies acrossthe hierarchies of work groups or departments

HR diversity change strategies that bring in many new hires of differentbackgrounds may have negative ramiregcations for social functioning in groupsparticularly if new members are not supported by additional HR strategies thatallow them to enter work groups on an equal footing or if their work groups donot have leadership that mirror their demographic identities As Kanter (1983)found identity groups must be equal in their access to power resources ifcontact is to produce improvement in attitudes There must also be time for thenew members to be socialized and integrated into the culture Harrison et al(1998) found that the longer the length of time that members worked togetherthe greater the effects of deep level attitude similarity Organizations areincreasing surface (demographic) and deep level (attitudinal) diversitysimultaneously which is likely to have negative short-term ramiregcations forthe consensus and direction of climate We argue HR strategies must not onlyfocus on the surface level plusmn remacrecting structural and demographic attributes plusmnbut also simultaneously to deep level characteristics remacrecting values within

HRM changestrategy

349

the context of speciregc groups at all organizational levels In sum HR strategistsand change agents should focus on how HR change strategies relate to the jointalignment of group memberrsquos multiple characteristics (eg demographichierarchical status task) in work groups plusmn that is to examine group contextualinmacruences (Wharton 1992)

54 LimitationsOur study while making contributions to organizational research is notwithout limitations Though large in individual respondents by utilizing thegroup as the level of analysis our sample size includes only 74 groups Thiscould reduce the level of statistical power of our regndings Additionally becausethe sample was based on one organization it is important to replicate thisanalysis across multiple organizations to substantiate our results Althoughchanges in department heterogeneity across an eight-year period wereincluded this is not a longitudinal design Future research should employrepeated measures over a length of time across all variables measuringdemographic change in heterogeneity as well as consensus and valence Ourstudy points to some of the challenges in measuring ordfsuccess in increasingdiversityordm as we did not look at annual turnover or the effects of the glassceiling at the group level in this study which is a gap we hope futureresearchers will be motivated to study spurred by our research

Since little published work has been done on group consensus regarding theeffects of HR policies to increase diversity through demographic change weextend many social science literatures ranging from organizationaldevelopment to demography and HR Future research on diversity andorganizational change also should strive to be interdisciplinary as theliterature on HR strategy and change has been largely developed in the HRliterature separate from other social science disciplines reducing theoreticalintegration Without such integration and increased scholarly and practitionerattention to climate and consensus many organizations may fail to supportgroup membersrsquo explicit use of their cultural experiences to advanceorganizational learning from enacted diversity change strategies despite thegroup and intergroup social tensions that will inevitably occur (Thomas andEly 1996)

References

Aiken LA and West SG (1993) Multiple Regression Testing and Interpreting InteractionsSage Publications Newbury Park CA

Alderfer CP and Smith KK (1982) ordfStudying embedded intergroup relations inorganizationsordm Administrative Science Quarterly Vol 27 pp 35-65

Allport G (1954) The Nature of Prejudice Addison Wesley Cambridge MA

Bailyn L (2000) ordfGender and diversity in organizations preconference presentationordm paperpresented at the National Academy of Management meeting Toronto

JOCM163

350

Blalock HM (1967) Toward a Theory of Minority Group Relations Wiley New York NY

Bleise PD and Halverson RR (1998) ordfGroup consensus and psychological well-being a largeregeld studyordm Journal of Applied Social Psychology Vol 28 pp 563-80

Brown KG Kozlowski SWJ and Hattrup K (1996) Theory Issues and Recommendations inConceptualizing Agreement as a Construct in Organizational Research The Search forConsensus Regarding Consensus National Academy of Management Cincinnati OH

Cox T Jr (1993) Cultural Diversity in Organizations Berrett-Koehler San Francisco CA

Ely R (1994) ordfThe effects of organizational demographics and social identity on relationshipsamong professional womenordm Administrative Science Quarterly Vol 39 pp 203-38

Ely R (1995) ordfThe power in demography womenrsquos social constructions of gender identity atworkordm Academy of Management Journal Vol 38 pp 589-634

Festinger L (1954) ordfA theory of social comparison processesordm Human Relations Vol 7pp 117-40

Forsyth DR (1990) Group Dynamics BrooksCole Publishing Paciregc Grove CA

Harrison DA Price KH and Bell MP (1998) ordfBeyond relational demography time and theeffects of surface and deep-level diversity on work group cohesionordm Academy ofManagement Journal Vol 41 pp 96-107

James K Lovato C and Khoo G (1994) ordfSocial identity correlates of minority workersrsquo healthordmAcademy of Management Journal Vol 37 p 383+

Kanter R (1977) ordfSome effects of proportions on group life skewed sex ratios and reactions totoken womenordm American Journal of Sociology Vol 82 pp 965-90

Kanter R (1983) Men and Women of the Corporation Basic Books New York NY

Kossek EE and Lobel SA (1996) ordfIntroduction transforming human resource systems tomanage diversityordm in Managing Diversity Human Resource Strategies for Transforming

the Workplace Blackwell Oxford pp 1-19

Kossek EE and Zonia SC (1993) ordfAssessing diversity climate a regeld study of reactions toemployer efforts to promote diversityordm Journal of Organizational Behavior Vol 14pp 61-81

Larkey LK (1996) ordfToward a theory of communicative interactions in culturally diverseworkgroupsordm Academy of Management Review Vol 21 pp 463-91

Lau DC and Murnighan JK (1998) ordfDemographic diversity and faultlines the compositionaldynamics of organizational groupsordm Academy of Management Review Vol 29 pp 325-40

Lindell MK and Brandt CJ (2000) ordfClimate quality and climate consensus as mediators of therelationship between organizational antecedents and outcomesordm Journal of AppliedPsychology Vol 85 pp 331-48

McGrath JE Berdahl JL and Arrow H (1995) ordfTraits expectations culture and clout thedynamics of diversity in work groupsordm in Jackson SE and Ruderman MN (Eds)Diversity in Work Teams Research Paradigms for a Changing Workplace AmericanPsychological Association Washington DC

Payne RL and Pugh DS (1976) ordfOrganizational structure and climateordm in Dunnette MD(Ed) Handbook of Industrial and Organizational Psychology Rand McNally Chicago ILpp 1125-73

Reichers A and Schneider B (1990) ordfClimate and culture an evolution of constructsordm inSchneider B (Ed) Organizational Climate and Culture Jossey-Bass San Francisco CA

HRM changestrategy

351

universityrsquos commitment to diversity National consulting regrms had beencontracted to provide advice on managing diversity The administrationmade funds readily available (sometimes over the initial salary posting) tohire quality minorities and women even for units in which new tenurestream positions had not been approved Training to promote diversitysensitivity among administrators was initiated Small cash achievementawards were distributed to recognize department activities that promotedmulticulturalism

Using the human information system the researchers collected facultydemographic data by department from records in the HR information system attwo points over an eight-year period to measure change in demography bydepartment We used a stratireged sample of tenure stream faculty with fullselection of the population of women and minorities and random sampling ofwhite males At the end of the eight years a survey was developed to examinethe success of the change efforts The archival data on proportional change indemography by group were used to form the independent variables (changefrom year 1 to year 8) and the survey data collected at the end of the eight yearswere used to form the group climate dependent variables The survey to assessthe success of these change strategies was developed with expert and facultyinput

22 MeasuresGroup dependent variables climate for diversity (consensus and valence) Fivegroup dependent variables were developed to assess the climate for diversityExploratory factor analysis ensured that each of these measures developed forthe study tapped into unique constructs Each of these measures utilized thesame regve-point scale ranging from 1 (strongly disagree) to 5 (strongly agree)Below we regrst describe the regve dependent variables and then we discuss howwe aggregated them to create group level consensus and valence measures

Diversity as an organizational goal We used Kossek and Zoniarsquos (1993) six-item scale (alpha 089) to assess the perceptions of the extent to which membersbelieved that organizational excellence and effective functioning were related tothe recruitment and retention of faculty who are female or minority Sampleitem ordfIf organization X is to remain an excellent institution it must recruit andretain more minority facultyordm

Commitment of management to diversity We developed a regve-item scale(alpha 085) to assess the administrationrsquos commitment to diversity through itssupport of efforts to increase faculty diversity through hiring and relatedpolicies Sample item ordfThe Dean of my college is strongly committed torecruiting more minority facultyordm

Perceived work-group mix We developed a regve-item scale (alpha 072) toassess the degree to which the department had a mix where women and

HRM changestrategy

335

minorities had good representation Sample item ordfMy department has a goodmix of men and women faculty membersordm

Equality of department support of women We used Kossek and Zoniarsquos(1993) three-item scale (alpha 071) to assess the degree to which faculty womenhad the same opportunity to have graduate assistants teaching release andabove average merit increases compared to males

Equality of department support of racial minorities We used Kossek andZoniarsquos (1993) three-item scale (alpha 077) to assess the degree to whichminority faculty had the same opportunity to have graduate assistantsteaching release and above average merit increases compared to non-minorities

Group consensus about climate for diversity These scales measured groupagreement for each of the regve dependent variables assessing climate Weutilized procedures outlined by Bleise and Halverson (1998) to ensure the validaggregation of individual data Group consensus was computed by

(1) determining the mean variance for each scale item per department

(2) calculating the sum of each scalersquos mean item variance

(3) dividing the sum (from step 2) by the number of items to obtain theaverage mean item variance for each scale

(4) multiplying the result from step 3 by plusmn1 to capture the level agreement

The closer the score is to 0 the higher the consensusGroup valence about the climate for diversity This scale assessed the median

score (a robust measure of central tendency) to assess the favorability of thegroup climate for the regve dependent variables Groups whose scores were at orabove the median have a more favorable climate for diversity than thosewhose scores were below the median

Group independent variables All of the independent variables were drawnfrom the universityrsquos HR information systems These included measures ofsize race gender and tenure (level) distribution by department at time 1 andtime 2 eight years later Variables were created to assess proportionaldemographic change for each group comparing the two points in time Forexample a grouprsquos proportion of women at time 1 was subtracted from theproportion of women at time 2 to measure proportional change

Change in department size was entered as a control variable since thedegree to which a department is growing may inmacruence member climateperceptions Drawing on Allportrsquos (1954) theory of social contact the moreopportunity an individual has to interact with members of other social groupsthe more likely they are to disconregrm individually held stereotypes Howeverwith increasing department size members may migrate toward others withfamiliar characteristics

JOCM163

336

3 Results31 Organizational structural demographic changeThe data show considerable success in increasing organizational diversity overthe eight-year hiring period At the organizational level the universityexperienced a 36 percent increase in female faculty from 437 at time 1 (baseline)to 644 at time 2 Similarly there was a 41 percent increase in the number ofracial minority faculty (from 211 at time 1 to 298 at time 2) Overall there were2511 total faculty at time 1 and 2684 at time 2 The net change was from 19percent at baseline to 24 percent eight years later for women and from 8percent to 11 percent for minorities Upward mobility was also improving Atbaseline the data show 9 percent women and 5 percent minority tenuredfaculty which increased to 11 percent (women) and 6 percent (minority) at time2 We took these data and transferred them into the measures described above

32 Survey response rateA total of 1529 individuals received the survey distributed at time 2 and 775were returned plusmn a 51 percent response rate For those surveys in whichdemographic identiregcation was provided analysis showed that response rateswere 47 percent for white men 46 percent for racial minority women 51 percentfor white women and 43 percent for racial minority men Of the 81 departmentsin the original sample 74 had at least three departmental respondents completethe diversity survey which were included in the regnal analysis Analyses weredone to ensure the respondents from the seven omitted departments did notstatistically differ from the 74 included Only tenure stream faculty wereincluded in the regnal dataset since there were relatively few non-tenure streamfaculty

33 Group level descriptive measures of diversity changeMeans standard deviations and correlations of measures are shown in Table IIn support of our regrst research objective Table I shows the descriptivestatistics for our new group level measures of diversity change These included

work group demographic shift

group climate (ie consensus and valence) and

group and organizational change referents

As the change variables show the average work group shift to increase inwomen was 1 percent with a standard deviation of 7 percent The averagework group increase in minorities was 1 percent with a standard deviationof 3 percent The group climate variables for the group referents (eg mixwell equality of department support of women equality of departmentsupport of minorities) and organizational referents (diversity is anorganizational goal commitment of management to diversity) are shownin Table I There was the greatest consensus that departments had equal

HRM changestrategy

337

Var

iable

Mea

nSD

12

34

5

1D

epar

tmen

tsi

zeT

ime

125

52

160

4plusmn

2D

epar

tmen

tsi

zeT

ime

229

06

176

60

89

plusmn3

Pro

por

tion

ofw

hit

esT

ime

10

900

010

210

21plusmn

4P

ropor

tion

ofw

hit

esT

ime

20

810

180

070

120

47

plusmn5

Pro

por

tion

ofw

omen

Tim

e1

019

021

005

002

001

017

plusmn6

Pro

por

tion

ofw

omen

Tim

e2

043

027

006

006

017

035

0

58

7P

ropor

tion

wit

hte

nure

Tim

e1

081

015

010

010

20

152

005

20

098

Pro

por

tion

wit

hT

enure

Tim

e2

084

011

004

003

20

082

018

20

129

Pro

por

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

034

019

20

042

009

20

100

000

76

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

039

018

20

092

010

20

39

20

130

65

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

247

550

40

062

0

59

20

092

013

009

12C

han

ge

inte

nure

tim

e1-

Tim

e2

001

001

20

152

013

20

28

20

120

0613

C

han

ge

ingen

der

Tim

e1-

Tim

e2

001

007

004

20

020

090

100

76

14C

han

ge

inw

hit

esT

ime

1-T

ime

20

010

032

009

20

162

059

2

005

2

008

15D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s2

100

057

001

001

20

192

015

024

16

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s2

099

048

007

014

20

032

006

008

17P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s2

125

047

001

011

002

20

060

1218

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s2

024

016

007

007

015

009

001

19E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s2

025

021

006

011

20

050

142

006

20D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n3

900

382

003

20

052

023

2

007

039

21

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n3

530

402

010

001

20

222

007

029

22P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

314

048

20

212

021

20

112

017

030

23E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n2

450

292

009

20

032

017

20

31

002

24E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

393

050

20

010

020

072

010

20

1625

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

361

049

20

030

012

018

20

050

2026

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

324

057

20

092

003

20

24

001

014

27P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n2

500

372

009

20

112

003

000

027

28

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

200

014

20

060

022

013

20

242

008

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n1

920

152

018

20

020

070

062

010

(con

tinued

)

Table IMeans standardsdeviations andcorrelations ofvariables

JOCM163

338

Var

iable

67

89

1011

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

plusmn7

Pro

por

tion

wit

hte

nure

Tim

e1

20

20plusmn

8P

ropor

tion

wit

hte

nure

Tim

e2

20

190

31plusmn

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

054

2

058

2

029

plusmn

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

048

2

018

20

32

073

plusmn

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

004

009

012

006

003

plusmn12

C

han

ge

inte

nure

Tim

e1-

Tim

e2

20

062

016

022

014

006

008

13C

han

ge

ingen

der

Tim

e1-

Tim

e2

032

2

019

20

130

68

055

0

0514

C

han

ge

inw

hit

esT

ime

1-T

ime

22

018

023

014

001

034

0

1315

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s0

010

020

080

140

110

0816

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s2

013

001

000

009

003

007

17P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s2

008

20

170

070

170

092

011

18E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s2

005

20

162

009

007

20

080

1519

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s2

011

011

008

20

102

023

004

20D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n0

57

20

022

014

036

0

48

014

21C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n0

030

182

001

016

026

0

0122

P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

007

20

040

050

32

030

0

0423

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n2

013

001

012

008

011

005

24E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

20

33

20

052

007

20

072

012

20

0225

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

048

0

022

004

020

027

0

1626

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

20

110

140

010

030

172

003

27P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n0

050

030

190

210

150

0628

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

20

092

004

022

20

012

002

004

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n2

026

2

018

20

102

007

20

162

024

(con

tinued

)

Table I

HRM changestrategy

339

Var

iable

1213

1415

1617

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

7P

ropor

tion

wit

hte

nure

Tim

e1

8P

ropor

tion

wit

hte

nure

Tim

e2

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

12C

han

ge

inte

nure

Tim

e1-

Tim

e2

plusmn13

C

han

ge

ingen

der

Tim

e1-

Tim

e2

001

plusmn14

C

han

ge

inw

hit

esT

ime

1-T

ime

20

162

007

plusmn15

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s0

150

110

14plusmn

16C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s2

013

021

006

20

07plusmn

17P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s0

150

130

060

38

034

plusmn

18E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s0

130

182

001

20

110

45

008

19E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s0

000

040

042

011

042

0

0020

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n0

000

050

150

34

20

32

20

0721

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n2

020

022

022

022

023

002

22P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

010

011

018

017

010

016

23E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n0

022

009

015

20

090

062

007

24E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

001

007

011

20

062

019

007

25D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

005

20

060

100

080

012

010

26C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

20

27

009

021

024

20

030

0227

P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n0

070

040

120

232

003

013

28E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

010

20

190

102

003

20

142

014

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n0

022

002

009

000

001

020

(con

tinued

)

Table I

JOCM163

340

Var

iable

1819

2021

2223

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

7P

ropor

tion

wit

hte

nure

Tim

e1

8P

ropor

tion

wit

hte

nure

Tim

e2

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

12C

han

ge

inte

nure

Tim

e1-

Tim

e2

13C

han

ge

ingen

der

Tim

e1-

Tim

e2

14C

han

ge

inw

hit

esT

ime

1-T

ime

215

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s16

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s17

P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s18

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

splusmn

19E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s0

57

plusmn20

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n2

035

2

027

plusmn

21C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n0

180

180

08plusmn

22P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

010

20

050

150

27

plusmn23

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n0

010

002

008

008

014

plusmn24

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

030

0

34

20

26

026

0

192

021

25D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

20

222

013

083

0

040

082

007

26C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

20

070

020

150

79

018

000

27P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n2

002

20

040

140

120

84

012

28E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

20

052

009

20

020

022

004

086

29

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n0

140

31

20

210

100

042

008

(con

tinued

)

Table I

HRM changestrategy

341

Table I

Var

iable

2425

2627

2829

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

7P

ropor

tion

wit

hte

nure

Tim

e1

8P

ropor

tion

wit

hte

nure

Tim

e2

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

12C

han

ge

inte

nure

Tim

e1-

Tim

e2

13C

han

ge

ingen

der

Tim

e1-

Tim

e2

14C

han

ge

inw

hit

esT

ime

1-T

ime

215

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s16

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s17

P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s18

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s19

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s20

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n21

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n22

P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

23E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n24

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

plusmn25

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

20

20plusmn

26C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

025

002

plusmn27

P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n0

080

010

19plusmn

28E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

20

32

000

20

010

05plusmn

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n0

58

20

240

180

132

003

plusmn

Note

p

001

orgre

ater

JOCM163

342

support of resources for women and men and whites and non-whites asthese measures hovered at plusmn024 and 2 025 or close to 0 which would betotal consensus There was the least consensus that minorities and majoritiesmixed well (plusmn125) with slightly more consensus that diversity should bepursued as organizational (plusmn10) The most positive valence was theperception of equality of department support for minorities (393) and thatdiversity should be pursued as a goal (39)

34 Regression results on demographic shift and climateTo address our second research objective which was to examine whetherwork groups with higher demographic shifts had more positive groupclimates for each of the regve dependent variables separate hierarchicalregressions were run to examine the relation between group demographicchange and group consensus (see Table II) or valence (see Table III) In theregrst step of all regressions proportional change in size was entered as acontrol For the consensus regressions the current mean of each dependentvariable was also entered as a control In step 2 the group proportionalchange in women and the group proportional change in minorities wereentered (In early analyses for the second step of each regression interactionterms change in gender or race by tenure were included following Aikenand Westrsquos (1993) procedures No signiregcant interactions were found) Usingstandardized beta weights as seen in step 2 of Tables II and III ourhypothesis that work groups with higher demographic shifts may havehigher consensus and valence toward diversity change received partialsupport Group referents of diversity climate were unchanged bydemographic shifts remacrecting possible dilution of HR strategiesOrganizational referents of valence were effected but consensus wasreduced in terms of whether the organization should pursue diversity as anorganizational goal Departments with higher shifts in the proportion ofwomen had lower consensus on whether the organization should pursuediversity as an organizational goal Both regressions with organizationalreferents as dependent variables were signiregcant Groups with higherincreases in the proportion of women over time tended to have morefavorable valence regarding the desirability of promoting diversity as anorganizational goal but were less sure of managementrsquos commitment asthey had more experience with the reality of implementing theorganizational change

35 Regression results on positive senior women and minority effect on climateOur third research objective was to examine whether increasing the proportionof senior women and minorities at the department level positively relates togroup valence and consensus In our data analysis the proportional change inoverall tenured faculty was entered in step 2 and then the proportional change

HRM changestrategy

343

Org

aniz

atio

nD

epar

tmen

tD

iver

sity

asor

gan

izat

ional

goa

l

Com

mit

men

tof

man

agem

ent

todiv

ersi

ty

Per

ceiv

edw

ork-g

roup

mix

Equal

ity

ofdep

artm

ent

suppor

tof

wom

enE

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

Ste

p1

Dep

tsi

zeT

ime

2-T

ime

10

020

080

180

000

06M

ean

ofD

V0

35

022

017

002

029

D

R2

012

0

060

060

000

09T

otal

R2

012

0

060

060

000

09Ste

p2

Wom

enT

ime

2-T

ime

12

035

2

017

20

182

007

20

02W

hit

esT

ime

2-T

ime

12

002

20

052

003

004

021

Ten

ure

Tim

e2-

Tim

e1

009

003

022

010

000

DR

20

11

003

008

001

004

Tot

alR

20

23

009

014

001

013

Ste

p3

Whit

ete

nure

dm

ales

Tim

e2-

Tim

e1

20

092

012

005

20

302

031

DR

20

010

010

000

040

05T

otal

R2

024

0

100

140

050

18

Note

p

005

orgre

ater

p0

01or

gre

ater

Table IIRegressionpredictingconsensus onclimate for diversity

JOCM163

344

Org

aniz

atio

nD

epar

tmen

tD

iver

sity

asor

gan

izat

ional

goa

l

Com

mit

men

tof

man

agem

ent

todiv

ersi

ty

Per

ceiv

edw

ork-g

roup

mix

Equal

ity

ofdep

artm

ent

suppor

tof

wom

enE

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

Ste

p1

Dep

tsi

zeT

ime

2-T

ime

10

080

152

006

021

0

05D

R2

001

002

000

004

0

00T

otal

R2

001

002

000

004

0

00Ste

p2

Wom

enT

ime

2-T

ime

10

40

20

38

20

21

20

062

013

Whit

esT

ime

2-T

ime

12

007

019

008

20

172

001

Ten

ure

Tim

e2-

Tim

e1

20

072

013

012

016

019

DR

20

16

017

0

060

070

05T

otal

R2

017

0

19

006

011

0

05Ste

p3

Whit

ete

nure

dm

ales

Tim

e2-

Tim

e1

005

011

000

007

035

D

R2

000

001

000

000

004

T

otal

R2

017

0

20

006

011

009

Note

p

005

orgre

ater

p0

01or

gre

ater

Table IIIRegression

predicting valenceof climate for

diversity

HRM changestrategy

345

in tenured white males was entered in step 3 The results are that the thirdblock was not statistically signiregcant in the regressions showing that smallincreases in the group proportions of senior women and minorities did notincrease group consensus or valence regarding climate for change

4 DiscussionThis study explored links between actual change in organizational structureorganizational demography and work group interpretation of this change Ourresults showed that HR strategies that focus on structural change withoutworking to develop supportive group norms and positive climate areinadequate change strategies Current employer efforts to increase diversity asan isolated HR strategy may not necessarily lead to increased memberagreement that change is favorable Most change efforts like the diversitychange effort we studied are focused at the organizational or individualintervention levels and under-emphasize the work group level and theimportance of group tipping points Our study highlights a disconnect betweenorganizational level diversity practice which often assumes that hiring moreminorities and women will improve climate and what our research found at thegroup level a climate that was mixed and ambivalent at best and sometimesnegative

41 Group level measures of demographic change consensus and valence ondiversity climateThe regrst objective of this study was to develop new theory and measures ofgroup level change in diversity We provided new ways to measure and groupsocial construction of change These included consensus the degree to whichmembers held common perceptions of diversity climate and valence thedirection of those perceptions Our measures identify group and organizationallevel referents of HR strategies to increase demographic diversity Group levelreferents of diversity change were how well members perceived their workgroup mix and the perceived equality of resource allocation acrossdemographic groups Organizational referents were managementcommitment to change and whether diversity should be an organizationalstrategy assessed general support for change Our study highlights howimportant it is for research and practice to not only examine how change indemography relates to differences in individual attitudes but also to groupmember shared perceptions about the direction of the climate for changeUnderstanding linkages between HR changes (in this case group demographywhich raises unique psychological processes related to social identity) andpositive consensus is important because it is relevant to cooperation workexperiences interpersonal conmacrict stress and norms in groups striving tomanage change

JOCM163

346

42 Greater work group demographic shift does not necessarily result infavorable climateOur second objective was to assess whether work groups with higherdemographic shifts tend to have more positive climates that is higherconsensus and valence toward diversity change for both organizational andgroup level change referents Our results only partially supported this beliefthat greater structural demographic shift would positively relate to a morefavorable climate for diversity Although at the organizational level theuniversity had experienced signiregcant change in organizational demographyan increase in the overall representation of white women (36 percent) andminorities (41 percent) over time work group members in units with thegreatest change did not have higher consensus or more favorable valenceWe found more favorable results for organizational than group levelreferents of change Members may regnd it easier to agree with organizationalreferents and more difregcult to agree on group level referents due to theirgreater psychological proximity to group processes that affect how theypersonally experience conmacrict and wellbeing from change in their dailywork Work groups with more demographic change had higher valence butlower consensus on whether the organization should pursue diversity as astrategic goal When organizational change strategies to increase diversityare enacted at the work group level the group demographic shift may bediluted across the organization to only slightly increase the representation ofwomen and minorities within actual groups Due to this dilution effectgroup membersrsquo social construction of the HR strategies may lack consensusor positive valence regarding the organizational diversity change strategiesIf the change is not large enough to create a critical mass to tiprepresentation of minorities to at least 35 percent according to Kanter (1977)negative intergroup dynamics such as increased resource competition andtokenism may still occur

43 Diluting diversity gradually increasing the proportion of senior womenminorities does not necessarily improve group climateOur third objective was to investigate whether increasing the proportion ofsenior women and minorities at the department level positively relates togroup valence and consensus There was no relation between increasing theproportion of senior women and minorities and group valence andconsensus Due to diluted saliency of organizational HR change strategies toincrease diversity at the work group level slight demographic shifts toproportionally increase the representation of senior women and minoritiesdid not relate to higher group valence or consensus This increase did littleto improve group climate and may have even hurt the climate in theshort run

HRM changestrategy

347

5 Implications for future research and practice51 Incremental structural change not sufregcient to improve climateWe hope this study will encourage researchers and practitioners interested indiversity change strategies to augment individual and organizational levels ofintervention to focus on work group climates supporting change Our resultshighlight some disparities between the underlying assumptions of HRstrategies to manage organizational change (increasing structural diversitywill be positive for climate) and reality at the department level (incrementalincreases did not improve climate) Organizational strategists implementingdiversity recruitment initiatives typically assume that change strategies (in thiscase hiring diversity) would improve the group social climate due to a positivechange in member attitudes based on increased group diversity and positivesocial contact Yet our study does not support this view Although theorganization we studied had had signiregcant success in enacting the formal HRstrategy plusmn increasing the representation of women and minorities at theorganizational level plusmn our research shows that gradually shifting thedemography of work groups may not necessarily lead to increased consensusor valence about the social climate within work groups For example it did notlead to improved climate regarding the perceived fairness of resourceallocation or good mixing in social interaction In fact the incremental increasein the proportion of minority groups lowered consensus and agreement in theshort term as subgroups sorted out shifting and socially ambiguous powerrelationships within work units These results suggest the importance ofmeasuring change initiatives impacts at the group level and having group levelinterventions such as team-building and group-focused resource allocation tosupport organizational level diversity strategies

52 Focus change strategies on tipping points within targeted work unitsFuture research should consider how HR strategies to manage organizationaldemographic change are socially constructed and enacted at the work grouplevel It is especially important to identify what constitutes a group leveltipping point or critical mass across hierarchical levels and differentdemographic groups to enable positive consensus supporting changePractice might regnd it more fruitful to focus future HR strategies to increasediversity through hiring to alter tipping points within targeted work groupsThis will dramatically alter the saliency of strategic organizationaldemographic change and intervene to shape the climate of speciregc units Ourstudy suggests that organizations may need demographic earthquakes withinwork groups to effect group consensus and valence on the climate regardingthe HR change (ie managing diversity) Dilution of change at the group levelmay ameliorate intended positive effects unless certain group level tippingpoints such as at least 35 percent of the work group are effected in each unit by

JOCM163

348

the HR strategy After eight years of hiring new faculty in our sample womenhovered at about 15 percent in 89 percent of groups and minorities had amaximum representation of 25 percent The demographic shifts whileincreasing were not dramatic enough to create imbalances that markedlyincreased the salience and power of new demographic subgroups created bythe HR strategies Instead of being balanced many of our groups were stilllargely white male and remain tilted or skewed over time which Kanter (1977)notes perpetuates stereotypical negative dynamics

53 Avoiding backlash weaving tipping points across the hierarchyRecent work at MIT by Bailyn (2000) suggests that backlash and resistance toorganizational change may be occurring among senior white male faculty inresponse to HR diversity strategies Initially women and minorities now enterdepartments with relatively equal resources to white men as supported by theHR system design However by the time women and minorities reach theassociate and full professor levels subtle discrimination processes occur in howHR policies are applied such as lower access to resources promotions andpoor climate Despite the increase in upward mobility the representation ofwomen and minorities at the top is still relatively low Ely (1994) found that ifwomen are under-represented at higher organizational levels relations betweenwomen at lower levels might be of lower quality due to increased competitionand negative gender dynamics in the larger social system If HR changestrategies that alter the distribution of existing resources across organizationaland demographic groups are to be successful they must take into account hownot only tipping points in terms of overall departmental representation butmore importantly how tipping points are distributed in demographies acrossthe hierarchies of work groups or departments

HR diversity change strategies that bring in many new hires of differentbackgrounds may have negative ramiregcations for social functioning in groupsparticularly if new members are not supported by additional HR strategies thatallow them to enter work groups on an equal footing or if their work groups donot have leadership that mirror their demographic identities As Kanter (1983)found identity groups must be equal in their access to power resources ifcontact is to produce improvement in attitudes There must also be time for thenew members to be socialized and integrated into the culture Harrison et al(1998) found that the longer the length of time that members worked togetherthe greater the effects of deep level attitude similarity Organizations areincreasing surface (demographic) and deep level (attitudinal) diversitysimultaneously which is likely to have negative short-term ramiregcations forthe consensus and direction of climate We argue HR strategies must not onlyfocus on the surface level plusmn remacrecting structural and demographic attributes plusmnbut also simultaneously to deep level characteristics remacrecting values within

HRM changestrategy

349

the context of speciregc groups at all organizational levels In sum HR strategistsand change agents should focus on how HR change strategies relate to the jointalignment of group memberrsquos multiple characteristics (eg demographichierarchical status task) in work groups plusmn that is to examine group contextualinmacruences (Wharton 1992)

54 LimitationsOur study while making contributions to organizational research is notwithout limitations Though large in individual respondents by utilizing thegroup as the level of analysis our sample size includes only 74 groups Thiscould reduce the level of statistical power of our regndings Additionally becausethe sample was based on one organization it is important to replicate thisanalysis across multiple organizations to substantiate our results Althoughchanges in department heterogeneity across an eight-year period wereincluded this is not a longitudinal design Future research should employrepeated measures over a length of time across all variables measuringdemographic change in heterogeneity as well as consensus and valence Ourstudy points to some of the challenges in measuring ordfsuccess in increasingdiversityordm as we did not look at annual turnover or the effects of the glassceiling at the group level in this study which is a gap we hope futureresearchers will be motivated to study spurred by our research

Since little published work has been done on group consensus regarding theeffects of HR policies to increase diversity through demographic change weextend many social science literatures ranging from organizationaldevelopment to demography and HR Future research on diversity andorganizational change also should strive to be interdisciplinary as theliterature on HR strategy and change has been largely developed in the HRliterature separate from other social science disciplines reducing theoreticalintegration Without such integration and increased scholarly and practitionerattention to climate and consensus many organizations may fail to supportgroup membersrsquo explicit use of their cultural experiences to advanceorganizational learning from enacted diversity change strategies despite thegroup and intergroup social tensions that will inevitably occur (Thomas andEly 1996)

References

Aiken LA and West SG (1993) Multiple Regression Testing and Interpreting InteractionsSage Publications Newbury Park CA

Alderfer CP and Smith KK (1982) ordfStudying embedded intergroup relations inorganizationsordm Administrative Science Quarterly Vol 27 pp 35-65

Allport G (1954) The Nature of Prejudice Addison Wesley Cambridge MA

Bailyn L (2000) ordfGender and diversity in organizations preconference presentationordm paperpresented at the National Academy of Management meeting Toronto

JOCM163

350

Blalock HM (1967) Toward a Theory of Minority Group Relations Wiley New York NY

Bleise PD and Halverson RR (1998) ordfGroup consensus and psychological well-being a largeregeld studyordm Journal of Applied Social Psychology Vol 28 pp 563-80

Brown KG Kozlowski SWJ and Hattrup K (1996) Theory Issues and Recommendations inConceptualizing Agreement as a Construct in Organizational Research The Search forConsensus Regarding Consensus National Academy of Management Cincinnati OH

Cox T Jr (1993) Cultural Diversity in Organizations Berrett-Koehler San Francisco CA

Ely R (1994) ordfThe effects of organizational demographics and social identity on relationshipsamong professional womenordm Administrative Science Quarterly Vol 39 pp 203-38

Ely R (1995) ordfThe power in demography womenrsquos social constructions of gender identity atworkordm Academy of Management Journal Vol 38 pp 589-634

Festinger L (1954) ordfA theory of social comparison processesordm Human Relations Vol 7pp 117-40

Forsyth DR (1990) Group Dynamics BrooksCole Publishing Paciregc Grove CA

Harrison DA Price KH and Bell MP (1998) ordfBeyond relational demography time and theeffects of surface and deep-level diversity on work group cohesionordm Academy ofManagement Journal Vol 41 pp 96-107

James K Lovato C and Khoo G (1994) ordfSocial identity correlates of minority workersrsquo healthordmAcademy of Management Journal Vol 37 p 383+

Kanter R (1977) ordfSome effects of proportions on group life skewed sex ratios and reactions totoken womenordm American Journal of Sociology Vol 82 pp 965-90

Kanter R (1983) Men and Women of the Corporation Basic Books New York NY

Kossek EE and Lobel SA (1996) ordfIntroduction transforming human resource systems tomanage diversityordm in Managing Diversity Human Resource Strategies for Transforming

the Workplace Blackwell Oxford pp 1-19

Kossek EE and Zonia SC (1993) ordfAssessing diversity climate a regeld study of reactions toemployer efforts to promote diversityordm Journal of Organizational Behavior Vol 14pp 61-81

Larkey LK (1996) ordfToward a theory of communicative interactions in culturally diverseworkgroupsordm Academy of Management Review Vol 21 pp 463-91

Lau DC and Murnighan JK (1998) ordfDemographic diversity and faultlines the compositionaldynamics of organizational groupsordm Academy of Management Review Vol 29 pp 325-40

Lindell MK and Brandt CJ (2000) ordfClimate quality and climate consensus as mediators of therelationship between organizational antecedents and outcomesordm Journal of AppliedPsychology Vol 85 pp 331-48

McGrath JE Berdahl JL and Arrow H (1995) ordfTraits expectations culture and clout thedynamics of diversity in work groupsordm in Jackson SE and Ruderman MN (Eds)Diversity in Work Teams Research Paradigms for a Changing Workplace AmericanPsychological Association Washington DC

Payne RL and Pugh DS (1976) ordfOrganizational structure and climateordm in Dunnette MD(Ed) Handbook of Industrial and Organizational Psychology Rand McNally Chicago ILpp 1125-73

Reichers A and Schneider B (1990) ordfClimate and culture an evolution of constructsordm inSchneider B (Ed) Organizational Climate and Culture Jossey-Bass San Francisco CA

HRM changestrategy

351

minorities had good representation Sample item ordfMy department has a goodmix of men and women faculty membersordm

Equality of department support of women We used Kossek and Zoniarsquos(1993) three-item scale (alpha 071) to assess the degree to which faculty womenhad the same opportunity to have graduate assistants teaching release andabove average merit increases compared to males

Equality of department support of racial minorities We used Kossek andZoniarsquos (1993) three-item scale (alpha 077) to assess the degree to whichminority faculty had the same opportunity to have graduate assistantsteaching release and above average merit increases compared to non-minorities

Group consensus about climate for diversity These scales measured groupagreement for each of the regve dependent variables assessing climate Weutilized procedures outlined by Bleise and Halverson (1998) to ensure the validaggregation of individual data Group consensus was computed by

(1) determining the mean variance for each scale item per department

(2) calculating the sum of each scalersquos mean item variance

(3) dividing the sum (from step 2) by the number of items to obtain theaverage mean item variance for each scale

(4) multiplying the result from step 3 by plusmn1 to capture the level agreement

The closer the score is to 0 the higher the consensusGroup valence about the climate for diversity This scale assessed the median

score (a robust measure of central tendency) to assess the favorability of thegroup climate for the regve dependent variables Groups whose scores were at orabove the median have a more favorable climate for diversity than thosewhose scores were below the median

Group independent variables All of the independent variables were drawnfrom the universityrsquos HR information systems These included measures ofsize race gender and tenure (level) distribution by department at time 1 andtime 2 eight years later Variables were created to assess proportionaldemographic change for each group comparing the two points in time Forexample a grouprsquos proportion of women at time 1 was subtracted from theproportion of women at time 2 to measure proportional change

Change in department size was entered as a control variable since thedegree to which a department is growing may inmacruence member climateperceptions Drawing on Allportrsquos (1954) theory of social contact the moreopportunity an individual has to interact with members of other social groupsthe more likely they are to disconregrm individually held stereotypes Howeverwith increasing department size members may migrate toward others withfamiliar characteristics

JOCM163

336

3 Results31 Organizational structural demographic changeThe data show considerable success in increasing organizational diversity overthe eight-year hiring period At the organizational level the universityexperienced a 36 percent increase in female faculty from 437 at time 1 (baseline)to 644 at time 2 Similarly there was a 41 percent increase in the number ofracial minority faculty (from 211 at time 1 to 298 at time 2) Overall there were2511 total faculty at time 1 and 2684 at time 2 The net change was from 19percent at baseline to 24 percent eight years later for women and from 8percent to 11 percent for minorities Upward mobility was also improving Atbaseline the data show 9 percent women and 5 percent minority tenuredfaculty which increased to 11 percent (women) and 6 percent (minority) at time2 We took these data and transferred them into the measures described above

32 Survey response rateA total of 1529 individuals received the survey distributed at time 2 and 775were returned plusmn a 51 percent response rate For those surveys in whichdemographic identiregcation was provided analysis showed that response rateswere 47 percent for white men 46 percent for racial minority women 51 percentfor white women and 43 percent for racial minority men Of the 81 departmentsin the original sample 74 had at least three departmental respondents completethe diversity survey which were included in the regnal analysis Analyses weredone to ensure the respondents from the seven omitted departments did notstatistically differ from the 74 included Only tenure stream faculty wereincluded in the regnal dataset since there were relatively few non-tenure streamfaculty

33 Group level descriptive measures of diversity changeMeans standard deviations and correlations of measures are shown in Table IIn support of our regrst research objective Table I shows the descriptivestatistics for our new group level measures of diversity change These included

work group demographic shift

group climate (ie consensus and valence) and

group and organizational change referents

As the change variables show the average work group shift to increase inwomen was 1 percent with a standard deviation of 7 percent The averagework group increase in minorities was 1 percent with a standard deviationof 3 percent The group climate variables for the group referents (eg mixwell equality of department support of women equality of departmentsupport of minorities) and organizational referents (diversity is anorganizational goal commitment of management to diversity) are shownin Table I There was the greatest consensus that departments had equal

HRM changestrategy

337

Var

iable

Mea

nSD

12

34

5

1D

epar

tmen

tsi

zeT

ime

125

52

160

4plusmn

2D

epar

tmen

tsi

zeT

ime

229

06

176

60

89

plusmn3

Pro

por

tion

ofw

hit

esT

ime

10

900

010

210

21plusmn

4P

ropor

tion

ofw

hit

esT

ime

20

810

180

070

120

47

plusmn5

Pro

por

tion

ofw

omen

Tim

e1

019

021

005

002

001

017

plusmn6

Pro

por

tion

ofw

omen

Tim

e2

043

027

006

006

017

035

0

58

7P

ropor

tion

wit

hte

nure

Tim

e1

081

015

010

010

20

152

005

20

098

Pro

por

tion

wit

hT

enure

Tim

e2

084

011

004

003

20

082

018

20

129

Pro

por

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

034

019

20

042

009

20

100

000

76

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

039

018

20

092

010

20

39

20

130

65

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

247

550

40

062

0

59

20

092

013

009

12C

han

ge

inte

nure

tim

e1-

Tim

e2

001

001

20

152

013

20

28

20

120

0613

C

han

ge

ingen

der

Tim

e1-

Tim

e2

001

007

004

20

020

090

100

76

14C

han

ge

inw

hit

esT

ime

1-T

ime

20

010

032

009

20

162

059

2

005

2

008

15D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s2

100

057

001

001

20

192

015

024

16

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s2

099

048

007

014

20

032

006

008

17P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s2

125

047

001

011

002

20

060

1218

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s2

024

016

007

007

015

009

001

19E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s2

025

021

006

011

20

050

142

006

20D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n3

900

382

003

20

052

023

2

007

039

21

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n3

530

402

010

001

20

222

007

029

22P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

314

048

20

212

021

20

112

017

030

23E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n2

450

292

009

20

032

017

20

31

002

24E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

393

050

20

010

020

072

010

20

1625

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

361

049

20

030

012

018

20

050

2026

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

324

057

20

092

003

20

24

001

014

27P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n2

500

372

009

20

112

003

000

027

28

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

200

014

20

060

022

013

20

242

008

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n1

920

152

018

20

020

070

062

010

(con

tinued

)

Table IMeans standardsdeviations andcorrelations ofvariables

JOCM163

338

Var

iable

67

89

1011

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

plusmn7

Pro

por

tion

wit

hte

nure

Tim

e1

20

20plusmn

8P

ropor

tion

wit

hte

nure

Tim

e2

20

190

31plusmn

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

054

2

058

2

029

plusmn

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

048

2

018

20

32

073

plusmn

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

004

009

012

006

003

plusmn12

C

han

ge

inte

nure

Tim

e1-

Tim

e2

20

062

016

022

014

006

008

13C

han

ge

ingen

der

Tim

e1-

Tim

e2

032

2

019

20

130

68

055

0

0514

C

han

ge

inw

hit

esT

ime

1-T

ime

22

018

023

014

001

034

0

1315

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s0

010

020

080

140

110

0816

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s2

013

001

000

009

003

007

17P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s2

008

20

170

070

170

092

011

18E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s2

005

20

162

009

007

20

080

1519

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s2

011

011

008

20

102

023

004

20D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n0

57

20

022

014

036

0

48

014

21C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n0

030

182

001

016

026

0

0122

P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

007

20

040

050

32

030

0

0423

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n2

013

001

012

008

011

005

24E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

20

33

20

052

007

20

072

012

20

0225

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

048

0

022

004

020

027

0

1626

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

20

110

140

010

030

172

003

27P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n0

050

030

190

210

150

0628

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

20

092

004

022

20

012

002

004

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n2

026

2

018

20

102

007

20

162

024

(con

tinued

)

Table I

HRM changestrategy

339

Var

iable

1213

1415

1617

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

7P

ropor

tion

wit

hte

nure

Tim

e1

8P

ropor

tion

wit

hte

nure

Tim

e2

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

12C

han

ge

inte

nure

Tim

e1-

Tim

e2

plusmn13

C

han

ge

ingen

der

Tim

e1-

Tim

e2

001

plusmn14

C

han

ge

inw

hit

esT

ime

1-T

ime

20

162

007

plusmn15

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s0

150

110

14plusmn

16C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s2

013

021

006

20

07plusmn

17P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s0

150

130

060

38

034

plusmn

18E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s0

130

182

001

20

110

45

008

19E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s0

000

040

042

011

042

0

0020

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n0

000

050

150

34

20

32

20

0721

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n2

020

022

022

022

023

002

22P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

010

011

018

017

010

016

23E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n0

022

009

015

20

090

062

007

24E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

001

007

011

20

062

019

007

25D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

005

20

060

100

080

012

010

26C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

20

27

009

021

024

20

030

0227

P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n0

070

040

120

232

003

013

28E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

010

20

190

102

003

20

142

014

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n0

022

002

009

000

001

020

(con

tinued

)

Table I

JOCM163

340

Var

iable

1819

2021

2223

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

7P

ropor

tion

wit

hte

nure

Tim

e1

8P

ropor

tion

wit

hte

nure

Tim

e2

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

12C

han

ge

inte

nure

Tim

e1-

Tim

e2

13C

han

ge

ingen

der

Tim

e1-

Tim

e2

14C

han

ge

inw

hit

esT

ime

1-T

ime

215

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s16

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s17

P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s18

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

splusmn

19E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s0

57

plusmn20

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n2

035

2

027

plusmn

21C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n0

180

180

08plusmn

22P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

010

20

050

150

27

plusmn23

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n0

010

002

008

008

014

plusmn24

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

030

0

34

20

26

026

0

192

021

25D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

20

222

013

083

0

040

082

007

26C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

20

070

020

150

79

018

000

27P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n2

002

20

040

140

120

84

012

28E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

20

052

009

20

020

022

004

086

29

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n0

140

31

20

210

100

042

008

(con

tinued

)

Table I

HRM changestrategy

341

Table I

Var

iable

2425

2627

2829

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

7P

ropor

tion

wit

hte

nure

Tim

e1

8P

ropor

tion

wit

hte

nure

Tim

e2

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

12C

han

ge

inte

nure

Tim

e1-

Tim

e2

13C

han

ge

ingen

der

Tim

e1-

Tim

e2

14C

han

ge

inw

hit

esT

ime

1-T

ime

215

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s16

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s17

P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s18

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s19

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s20

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n21

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n22

P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

23E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n24

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

plusmn25

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

20

20plusmn

26C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

025

002

plusmn27

P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n0

080

010

19plusmn

28E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

20

32

000

20

010

05plusmn

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n0

58

20

240

180

132

003

plusmn

Note

p

001

orgre

ater

JOCM163

342

support of resources for women and men and whites and non-whites asthese measures hovered at plusmn024 and 2 025 or close to 0 which would betotal consensus There was the least consensus that minorities and majoritiesmixed well (plusmn125) with slightly more consensus that diversity should bepursued as organizational (plusmn10) The most positive valence was theperception of equality of department support for minorities (393) and thatdiversity should be pursued as a goal (39)

34 Regression results on demographic shift and climateTo address our second research objective which was to examine whetherwork groups with higher demographic shifts had more positive groupclimates for each of the regve dependent variables separate hierarchicalregressions were run to examine the relation between group demographicchange and group consensus (see Table II) or valence (see Table III) In theregrst step of all regressions proportional change in size was entered as acontrol For the consensus regressions the current mean of each dependentvariable was also entered as a control In step 2 the group proportionalchange in women and the group proportional change in minorities wereentered (In early analyses for the second step of each regression interactionterms change in gender or race by tenure were included following Aikenand Westrsquos (1993) procedures No signiregcant interactions were found) Usingstandardized beta weights as seen in step 2 of Tables II and III ourhypothesis that work groups with higher demographic shifts may havehigher consensus and valence toward diversity change received partialsupport Group referents of diversity climate were unchanged bydemographic shifts remacrecting possible dilution of HR strategiesOrganizational referents of valence were effected but consensus wasreduced in terms of whether the organization should pursue diversity as anorganizational goal Departments with higher shifts in the proportion ofwomen had lower consensus on whether the organization should pursuediversity as an organizational goal Both regressions with organizationalreferents as dependent variables were signiregcant Groups with higherincreases in the proportion of women over time tended to have morefavorable valence regarding the desirability of promoting diversity as anorganizational goal but were less sure of managementrsquos commitment asthey had more experience with the reality of implementing theorganizational change

35 Regression results on positive senior women and minority effect on climateOur third research objective was to examine whether increasing the proportionof senior women and minorities at the department level positively relates togroup valence and consensus In our data analysis the proportional change inoverall tenured faculty was entered in step 2 and then the proportional change

HRM changestrategy

343

Org

aniz

atio

nD

epar

tmen

tD

iver

sity

asor

gan

izat

ional

goa

l

Com

mit

men

tof

man

agem

ent

todiv

ersi

ty

Per

ceiv

edw

ork-g

roup

mix

Equal

ity

ofdep

artm

ent

suppor

tof

wom

enE

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

Ste

p1

Dep

tsi

zeT

ime

2-T

ime

10

020

080

180

000

06M

ean

ofD

V0

35

022

017

002

029

D

R2

012

0

060

060

000

09T

otal

R2

012

0

060

060

000

09Ste

p2

Wom

enT

ime

2-T

ime

12

035

2

017

20

182

007

20

02W

hit

esT

ime

2-T

ime

12

002

20

052

003

004

021

Ten

ure

Tim

e2-

Tim

e1

009

003

022

010

000

DR

20

11

003

008

001

004

Tot

alR

20

23

009

014

001

013

Ste

p3

Whit

ete

nure

dm

ales

Tim

e2-

Tim

e1

20

092

012

005

20

302

031

DR

20

010

010

000

040

05T

otal

R2

024

0

100

140

050

18

Note

p

005

orgre

ater

p0

01or

gre

ater

Table IIRegressionpredictingconsensus onclimate for diversity

JOCM163

344

Org

aniz

atio

nD

epar

tmen

tD

iver

sity

asor

gan

izat

ional

goa

l

Com

mit

men

tof

man

agem

ent

todiv

ersi

ty

Per

ceiv

edw

ork-g

roup

mix

Equal

ity

ofdep

artm

ent

suppor

tof

wom

enE

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

Ste

p1

Dep

tsi

zeT

ime

2-T

ime

10

080

152

006

021

0

05D

R2

001

002

000

004

0

00T

otal

R2

001

002

000

004

0

00Ste

p2

Wom

enT

ime

2-T

ime

10

40

20

38

20

21

20

062

013

Whit

esT

ime

2-T

ime

12

007

019

008

20

172

001

Ten

ure

Tim

e2-

Tim

e1

20

072

013

012

016

019

DR

20

16

017

0

060

070

05T

otal

R2

017

0

19

006

011

0

05Ste

p3

Whit

ete

nure

dm

ales

Tim

e2-

Tim

e1

005

011

000

007

035

D

R2

000

001

000

000

004

T

otal

R2

017

0

20

006

011

009

Note

p

005

orgre

ater

p0

01or

gre

ater

Table IIIRegression

predicting valenceof climate for

diversity

HRM changestrategy

345

in tenured white males was entered in step 3 The results are that the thirdblock was not statistically signiregcant in the regressions showing that smallincreases in the group proportions of senior women and minorities did notincrease group consensus or valence regarding climate for change

4 DiscussionThis study explored links between actual change in organizational structureorganizational demography and work group interpretation of this change Ourresults showed that HR strategies that focus on structural change withoutworking to develop supportive group norms and positive climate areinadequate change strategies Current employer efforts to increase diversity asan isolated HR strategy may not necessarily lead to increased memberagreement that change is favorable Most change efforts like the diversitychange effort we studied are focused at the organizational or individualintervention levels and under-emphasize the work group level and theimportance of group tipping points Our study highlights a disconnect betweenorganizational level diversity practice which often assumes that hiring moreminorities and women will improve climate and what our research found at thegroup level a climate that was mixed and ambivalent at best and sometimesnegative

41 Group level measures of demographic change consensus and valence ondiversity climateThe regrst objective of this study was to develop new theory and measures ofgroup level change in diversity We provided new ways to measure and groupsocial construction of change These included consensus the degree to whichmembers held common perceptions of diversity climate and valence thedirection of those perceptions Our measures identify group and organizationallevel referents of HR strategies to increase demographic diversity Group levelreferents of diversity change were how well members perceived their workgroup mix and the perceived equality of resource allocation acrossdemographic groups Organizational referents were managementcommitment to change and whether diversity should be an organizationalstrategy assessed general support for change Our study highlights howimportant it is for research and practice to not only examine how change indemography relates to differences in individual attitudes but also to groupmember shared perceptions about the direction of the climate for changeUnderstanding linkages between HR changes (in this case group demographywhich raises unique psychological processes related to social identity) andpositive consensus is important because it is relevant to cooperation workexperiences interpersonal conmacrict stress and norms in groups striving tomanage change

JOCM163

346

42 Greater work group demographic shift does not necessarily result infavorable climateOur second objective was to assess whether work groups with higherdemographic shifts tend to have more positive climates that is higherconsensus and valence toward diversity change for both organizational andgroup level change referents Our results only partially supported this beliefthat greater structural demographic shift would positively relate to a morefavorable climate for diversity Although at the organizational level theuniversity had experienced signiregcant change in organizational demographyan increase in the overall representation of white women (36 percent) andminorities (41 percent) over time work group members in units with thegreatest change did not have higher consensus or more favorable valenceWe found more favorable results for organizational than group levelreferents of change Members may regnd it easier to agree with organizationalreferents and more difregcult to agree on group level referents due to theirgreater psychological proximity to group processes that affect how theypersonally experience conmacrict and wellbeing from change in their dailywork Work groups with more demographic change had higher valence butlower consensus on whether the organization should pursue diversity as astrategic goal When organizational change strategies to increase diversityare enacted at the work group level the group demographic shift may bediluted across the organization to only slightly increase the representation ofwomen and minorities within actual groups Due to this dilution effectgroup membersrsquo social construction of the HR strategies may lack consensusor positive valence regarding the organizational diversity change strategiesIf the change is not large enough to create a critical mass to tiprepresentation of minorities to at least 35 percent according to Kanter (1977)negative intergroup dynamics such as increased resource competition andtokenism may still occur

43 Diluting diversity gradually increasing the proportion of senior womenminorities does not necessarily improve group climateOur third objective was to investigate whether increasing the proportion ofsenior women and minorities at the department level positively relates togroup valence and consensus There was no relation between increasing theproportion of senior women and minorities and group valence andconsensus Due to diluted saliency of organizational HR change strategies toincrease diversity at the work group level slight demographic shifts toproportionally increase the representation of senior women and minoritiesdid not relate to higher group valence or consensus This increase did littleto improve group climate and may have even hurt the climate in theshort run

HRM changestrategy

347

5 Implications for future research and practice51 Incremental structural change not sufregcient to improve climateWe hope this study will encourage researchers and practitioners interested indiversity change strategies to augment individual and organizational levels ofintervention to focus on work group climates supporting change Our resultshighlight some disparities between the underlying assumptions of HRstrategies to manage organizational change (increasing structural diversitywill be positive for climate) and reality at the department level (incrementalincreases did not improve climate) Organizational strategists implementingdiversity recruitment initiatives typically assume that change strategies (in thiscase hiring diversity) would improve the group social climate due to a positivechange in member attitudes based on increased group diversity and positivesocial contact Yet our study does not support this view Although theorganization we studied had had signiregcant success in enacting the formal HRstrategy plusmn increasing the representation of women and minorities at theorganizational level plusmn our research shows that gradually shifting thedemography of work groups may not necessarily lead to increased consensusor valence about the social climate within work groups For example it did notlead to improved climate regarding the perceived fairness of resourceallocation or good mixing in social interaction In fact the incremental increasein the proportion of minority groups lowered consensus and agreement in theshort term as subgroups sorted out shifting and socially ambiguous powerrelationships within work units These results suggest the importance ofmeasuring change initiatives impacts at the group level and having group levelinterventions such as team-building and group-focused resource allocation tosupport organizational level diversity strategies

52 Focus change strategies on tipping points within targeted work unitsFuture research should consider how HR strategies to manage organizationaldemographic change are socially constructed and enacted at the work grouplevel It is especially important to identify what constitutes a group leveltipping point or critical mass across hierarchical levels and differentdemographic groups to enable positive consensus supporting changePractice might regnd it more fruitful to focus future HR strategies to increasediversity through hiring to alter tipping points within targeted work groupsThis will dramatically alter the saliency of strategic organizationaldemographic change and intervene to shape the climate of speciregc units Ourstudy suggests that organizations may need demographic earthquakes withinwork groups to effect group consensus and valence on the climate regardingthe HR change (ie managing diversity) Dilution of change at the group levelmay ameliorate intended positive effects unless certain group level tippingpoints such as at least 35 percent of the work group are effected in each unit by

JOCM163

348

the HR strategy After eight years of hiring new faculty in our sample womenhovered at about 15 percent in 89 percent of groups and minorities had amaximum representation of 25 percent The demographic shifts whileincreasing were not dramatic enough to create imbalances that markedlyincreased the salience and power of new demographic subgroups created bythe HR strategies Instead of being balanced many of our groups were stilllargely white male and remain tilted or skewed over time which Kanter (1977)notes perpetuates stereotypical negative dynamics

53 Avoiding backlash weaving tipping points across the hierarchyRecent work at MIT by Bailyn (2000) suggests that backlash and resistance toorganizational change may be occurring among senior white male faculty inresponse to HR diversity strategies Initially women and minorities now enterdepartments with relatively equal resources to white men as supported by theHR system design However by the time women and minorities reach theassociate and full professor levels subtle discrimination processes occur in howHR policies are applied such as lower access to resources promotions andpoor climate Despite the increase in upward mobility the representation ofwomen and minorities at the top is still relatively low Ely (1994) found that ifwomen are under-represented at higher organizational levels relations betweenwomen at lower levels might be of lower quality due to increased competitionand negative gender dynamics in the larger social system If HR changestrategies that alter the distribution of existing resources across organizationaland demographic groups are to be successful they must take into account hownot only tipping points in terms of overall departmental representation butmore importantly how tipping points are distributed in demographies acrossthe hierarchies of work groups or departments

HR diversity change strategies that bring in many new hires of differentbackgrounds may have negative ramiregcations for social functioning in groupsparticularly if new members are not supported by additional HR strategies thatallow them to enter work groups on an equal footing or if their work groups donot have leadership that mirror their demographic identities As Kanter (1983)found identity groups must be equal in their access to power resources ifcontact is to produce improvement in attitudes There must also be time for thenew members to be socialized and integrated into the culture Harrison et al(1998) found that the longer the length of time that members worked togetherthe greater the effects of deep level attitude similarity Organizations areincreasing surface (demographic) and deep level (attitudinal) diversitysimultaneously which is likely to have negative short-term ramiregcations forthe consensus and direction of climate We argue HR strategies must not onlyfocus on the surface level plusmn remacrecting structural and demographic attributes plusmnbut also simultaneously to deep level characteristics remacrecting values within

HRM changestrategy

349

the context of speciregc groups at all organizational levels In sum HR strategistsand change agents should focus on how HR change strategies relate to the jointalignment of group memberrsquos multiple characteristics (eg demographichierarchical status task) in work groups plusmn that is to examine group contextualinmacruences (Wharton 1992)

54 LimitationsOur study while making contributions to organizational research is notwithout limitations Though large in individual respondents by utilizing thegroup as the level of analysis our sample size includes only 74 groups Thiscould reduce the level of statistical power of our regndings Additionally becausethe sample was based on one organization it is important to replicate thisanalysis across multiple organizations to substantiate our results Althoughchanges in department heterogeneity across an eight-year period wereincluded this is not a longitudinal design Future research should employrepeated measures over a length of time across all variables measuringdemographic change in heterogeneity as well as consensus and valence Ourstudy points to some of the challenges in measuring ordfsuccess in increasingdiversityordm as we did not look at annual turnover or the effects of the glassceiling at the group level in this study which is a gap we hope futureresearchers will be motivated to study spurred by our research

Since little published work has been done on group consensus regarding theeffects of HR policies to increase diversity through demographic change weextend many social science literatures ranging from organizationaldevelopment to demography and HR Future research on diversity andorganizational change also should strive to be interdisciplinary as theliterature on HR strategy and change has been largely developed in the HRliterature separate from other social science disciplines reducing theoreticalintegration Without such integration and increased scholarly and practitionerattention to climate and consensus many organizations may fail to supportgroup membersrsquo explicit use of their cultural experiences to advanceorganizational learning from enacted diversity change strategies despite thegroup and intergroup social tensions that will inevitably occur (Thomas andEly 1996)

References

Aiken LA and West SG (1993) Multiple Regression Testing and Interpreting InteractionsSage Publications Newbury Park CA

Alderfer CP and Smith KK (1982) ordfStudying embedded intergroup relations inorganizationsordm Administrative Science Quarterly Vol 27 pp 35-65

Allport G (1954) The Nature of Prejudice Addison Wesley Cambridge MA

Bailyn L (2000) ordfGender and diversity in organizations preconference presentationordm paperpresented at the National Academy of Management meeting Toronto

JOCM163

350

Blalock HM (1967) Toward a Theory of Minority Group Relations Wiley New York NY

Bleise PD and Halverson RR (1998) ordfGroup consensus and psychological well-being a largeregeld studyordm Journal of Applied Social Psychology Vol 28 pp 563-80

Brown KG Kozlowski SWJ and Hattrup K (1996) Theory Issues and Recommendations inConceptualizing Agreement as a Construct in Organizational Research The Search forConsensus Regarding Consensus National Academy of Management Cincinnati OH

Cox T Jr (1993) Cultural Diversity in Organizations Berrett-Koehler San Francisco CA

Ely R (1994) ordfThe effects of organizational demographics and social identity on relationshipsamong professional womenordm Administrative Science Quarterly Vol 39 pp 203-38

Ely R (1995) ordfThe power in demography womenrsquos social constructions of gender identity atworkordm Academy of Management Journal Vol 38 pp 589-634

Festinger L (1954) ordfA theory of social comparison processesordm Human Relations Vol 7pp 117-40

Forsyth DR (1990) Group Dynamics BrooksCole Publishing Paciregc Grove CA

Harrison DA Price KH and Bell MP (1998) ordfBeyond relational demography time and theeffects of surface and deep-level diversity on work group cohesionordm Academy ofManagement Journal Vol 41 pp 96-107

James K Lovato C and Khoo G (1994) ordfSocial identity correlates of minority workersrsquo healthordmAcademy of Management Journal Vol 37 p 383+

Kanter R (1977) ordfSome effects of proportions on group life skewed sex ratios and reactions totoken womenordm American Journal of Sociology Vol 82 pp 965-90

Kanter R (1983) Men and Women of the Corporation Basic Books New York NY

Kossek EE and Lobel SA (1996) ordfIntroduction transforming human resource systems tomanage diversityordm in Managing Diversity Human Resource Strategies for Transforming

the Workplace Blackwell Oxford pp 1-19

Kossek EE and Zonia SC (1993) ordfAssessing diversity climate a regeld study of reactions toemployer efforts to promote diversityordm Journal of Organizational Behavior Vol 14pp 61-81

Larkey LK (1996) ordfToward a theory of communicative interactions in culturally diverseworkgroupsordm Academy of Management Review Vol 21 pp 463-91

Lau DC and Murnighan JK (1998) ordfDemographic diversity and faultlines the compositionaldynamics of organizational groupsordm Academy of Management Review Vol 29 pp 325-40

Lindell MK and Brandt CJ (2000) ordfClimate quality and climate consensus as mediators of therelationship between organizational antecedents and outcomesordm Journal of AppliedPsychology Vol 85 pp 331-48

McGrath JE Berdahl JL and Arrow H (1995) ordfTraits expectations culture and clout thedynamics of diversity in work groupsordm in Jackson SE and Ruderman MN (Eds)Diversity in Work Teams Research Paradigms for a Changing Workplace AmericanPsychological Association Washington DC

Payne RL and Pugh DS (1976) ordfOrganizational structure and climateordm in Dunnette MD(Ed) Handbook of Industrial and Organizational Psychology Rand McNally Chicago ILpp 1125-73

Reichers A and Schneider B (1990) ordfClimate and culture an evolution of constructsordm inSchneider B (Ed) Organizational Climate and Culture Jossey-Bass San Francisco CA

HRM changestrategy

351

3 Results31 Organizational structural demographic changeThe data show considerable success in increasing organizational diversity overthe eight-year hiring period At the organizational level the universityexperienced a 36 percent increase in female faculty from 437 at time 1 (baseline)to 644 at time 2 Similarly there was a 41 percent increase in the number ofracial minority faculty (from 211 at time 1 to 298 at time 2) Overall there were2511 total faculty at time 1 and 2684 at time 2 The net change was from 19percent at baseline to 24 percent eight years later for women and from 8percent to 11 percent for minorities Upward mobility was also improving Atbaseline the data show 9 percent women and 5 percent minority tenuredfaculty which increased to 11 percent (women) and 6 percent (minority) at time2 We took these data and transferred them into the measures described above

32 Survey response rateA total of 1529 individuals received the survey distributed at time 2 and 775were returned plusmn a 51 percent response rate For those surveys in whichdemographic identiregcation was provided analysis showed that response rateswere 47 percent for white men 46 percent for racial minority women 51 percentfor white women and 43 percent for racial minority men Of the 81 departmentsin the original sample 74 had at least three departmental respondents completethe diversity survey which were included in the regnal analysis Analyses weredone to ensure the respondents from the seven omitted departments did notstatistically differ from the 74 included Only tenure stream faculty wereincluded in the regnal dataset since there were relatively few non-tenure streamfaculty

33 Group level descriptive measures of diversity changeMeans standard deviations and correlations of measures are shown in Table IIn support of our regrst research objective Table I shows the descriptivestatistics for our new group level measures of diversity change These included

work group demographic shift

group climate (ie consensus and valence) and

group and organizational change referents

As the change variables show the average work group shift to increase inwomen was 1 percent with a standard deviation of 7 percent The averagework group increase in minorities was 1 percent with a standard deviationof 3 percent The group climate variables for the group referents (eg mixwell equality of department support of women equality of departmentsupport of minorities) and organizational referents (diversity is anorganizational goal commitment of management to diversity) are shownin Table I There was the greatest consensus that departments had equal

HRM changestrategy

337

Var

iable

Mea

nSD

12

34

5

1D

epar

tmen

tsi

zeT

ime

125

52

160

4plusmn

2D

epar

tmen

tsi

zeT

ime

229

06

176

60

89

plusmn3

Pro

por

tion

ofw

hit

esT

ime

10

900

010

210

21plusmn

4P

ropor

tion

ofw

hit

esT

ime

20

810

180

070

120

47

plusmn5

Pro

por

tion

ofw

omen

Tim

e1

019

021

005

002

001

017

plusmn6

Pro

por

tion

ofw

omen

Tim

e2

043

027

006

006

017

035

0

58

7P

ropor

tion

wit

hte

nure

Tim

e1

081

015

010

010

20

152

005

20

098

Pro

por

tion

wit

hT

enure

Tim

e2

084

011

004

003

20

082

018

20

129

Pro

por

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

034

019

20

042

009

20

100

000

76

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

039

018

20

092

010

20

39

20

130

65

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

247

550

40

062

0

59

20

092

013

009

12C

han

ge

inte

nure

tim

e1-

Tim

e2

001

001

20

152

013

20

28

20

120

0613

C

han

ge

ingen

der

Tim

e1-

Tim

e2

001

007

004

20

020

090

100

76

14C

han

ge

inw

hit

esT

ime

1-T

ime

20

010

032

009

20

162

059

2

005

2

008

15D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s2

100

057

001

001

20

192

015

024

16

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s2

099

048

007

014

20

032

006

008

17P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s2

125

047

001

011

002

20

060

1218

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s2

024

016

007

007

015

009

001

19E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s2

025

021

006

011

20

050

142

006

20D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n3

900

382

003

20

052

023

2

007

039

21

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n3

530

402

010

001

20

222

007

029

22P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

314

048

20

212

021

20

112

017

030

23E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n2

450

292

009

20

032

017

20

31

002

24E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

393

050

20

010

020

072

010

20

1625

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

361

049

20

030

012

018

20

050

2026

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

324

057

20

092

003

20

24

001

014

27P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n2

500

372

009

20

112

003

000

027

28

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

200

014

20

060

022

013

20

242

008

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n1

920

152

018

20

020

070

062

010

(con

tinued

)

Table IMeans standardsdeviations andcorrelations ofvariables

JOCM163

338

Var

iable

67

89

1011

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

plusmn7

Pro

por

tion

wit

hte

nure

Tim

e1

20

20plusmn

8P

ropor

tion

wit

hte

nure

Tim

e2

20

190

31plusmn

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

054

2

058

2

029

plusmn

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

048

2

018

20

32

073

plusmn

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

004

009

012

006

003

plusmn12

C

han

ge

inte

nure

Tim

e1-

Tim

e2

20

062

016

022

014

006

008

13C

han

ge

ingen

der

Tim

e1-

Tim

e2

032

2

019

20

130

68

055

0

0514

C

han

ge

inw

hit

esT

ime

1-T

ime

22

018

023

014

001

034

0

1315

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s0

010

020

080

140

110

0816

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s2

013

001

000

009

003

007

17P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s2

008

20

170

070

170

092

011

18E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s2

005

20

162

009

007

20

080

1519

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s2

011

011

008

20

102

023

004

20D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n0

57

20

022

014

036

0

48

014

21C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n0

030

182

001

016

026

0

0122

P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

007

20

040

050

32

030

0

0423

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n2

013

001

012

008

011

005

24E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

20

33

20

052

007

20

072

012

20

0225

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

048

0

022

004

020

027

0

1626

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

20

110

140

010

030

172

003

27P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n0

050

030

190

210

150

0628

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

20

092

004

022

20

012

002

004

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n2

026

2

018

20

102

007

20

162

024

(con

tinued

)

Table I

HRM changestrategy

339

Var

iable

1213

1415

1617

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

7P

ropor

tion

wit

hte

nure

Tim

e1

8P

ropor

tion

wit

hte

nure

Tim

e2

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

12C

han

ge

inte

nure

Tim

e1-

Tim

e2

plusmn13

C

han

ge

ingen

der

Tim

e1-

Tim

e2

001

plusmn14

C

han

ge

inw

hit

esT

ime

1-T

ime

20

162

007

plusmn15

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s0

150

110

14plusmn

16C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s2

013

021

006

20

07plusmn

17P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s0

150

130

060

38

034

plusmn

18E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s0

130

182

001

20

110

45

008

19E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s0

000

040

042

011

042

0

0020

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n0

000

050

150

34

20

32

20

0721

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n2

020

022

022

022

023

002

22P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

010

011

018

017

010

016

23E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n0

022

009

015

20

090

062

007

24E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

001

007

011

20

062

019

007

25D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

005

20

060

100

080

012

010

26C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

20

27

009

021

024

20

030

0227

P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n0

070

040

120

232

003

013

28E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

010

20

190

102

003

20

142

014

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n0

022

002

009

000

001

020

(con

tinued

)

Table I

JOCM163

340

Var

iable

1819

2021

2223

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

7P

ropor

tion

wit

hte

nure

Tim

e1

8P

ropor

tion

wit

hte

nure

Tim

e2

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

12C

han

ge

inte

nure

Tim

e1-

Tim

e2

13C

han

ge

ingen

der

Tim

e1-

Tim

e2

14C

han

ge

inw

hit

esT

ime

1-T

ime

215

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s16

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s17

P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s18

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

splusmn

19E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s0

57

plusmn20

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n2

035

2

027

plusmn

21C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n0

180

180

08plusmn

22P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

010

20

050

150

27

plusmn23

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n0

010

002

008

008

014

plusmn24

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

030

0

34

20

26

026

0

192

021

25D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

20

222

013

083

0

040

082

007

26C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

20

070

020

150

79

018

000

27P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n2

002

20

040

140

120

84

012

28E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

20

052

009

20

020

022

004

086

29

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n0

140

31

20

210

100

042

008

(con

tinued

)

Table I

HRM changestrategy

341

Table I

Var

iable

2425

2627

2829

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

7P

ropor

tion

wit

hte

nure

Tim

e1

8P

ropor

tion

wit

hte

nure

Tim

e2

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

12C

han

ge

inte

nure

Tim

e1-

Tim

e2

13C

han

ge

ingen

der

Tim

e1-

Tim

e2

14C

han

ge

inw

hit

esT

ime

1-T

ime

215

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s16

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s17

P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s18

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s19

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s20

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n21

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n22

P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

23E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n24

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

plusmn25

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

20

20plusmn

26C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

025

002

plusmn27

P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n0

080

010

19plusmn

28E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

20

32

000

20

010

05plusmn

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n0

58

20

240

180

132

003

plusmn

Note

p

001

orgre

ater

JOCM163

342

support of resources for women and men and whites and non-whites asthese measures hovered at plusmn024 and 2 025 or close to 0 which would betotal consensus There was the least consensus that minorities and majoritiesmixed well (plusmn125) with slightly more consensus that diversity should bepursued as organizational (plusmn10) The most positive valence was theperception of equality of department support for minorities (393) and thatdiversity should be pursued as a goal (39)

34 Regression results on demographic shift and climateTo address our second research objective which was to examine whetherwork groups with higher demographic shifts had more positive groupclimates for each of the regve dependent variables separate hierarchicalregressions were run to examine the relation between group demographicchange and group consensus (see Table II) or valence (see Table III) In theregrst step of all regressions proportional change in size was entered as acontrol For the consensus regressions the current mean of each dependentvariable was also entered as a control In step 2 the group proportionalchange in women and the group proportional change in minorities wereentered (In early analyses for the second step of each regression interactionterms change in gender or race by tenure were included following Aikenand Westrsquos (1993) procedures No signiregcant interactions were found) Usingstandardized beta weights as seen in step 2 of Tables II and III ourhypothesis that work groups with higher demographic shifts may havehigher consensus and valence toward diversity change received partialsupport Group referents of diversity climate were unchanged bydemographic shifts remacrecting possible dilution of HR strategiesOrganizational referents of valence were effected but consensus wasreduced in terms of whether the organization should pursue diversity as anorganizational goal Departments with higher shifts in the proportion ofwomen had lower consensus on whether the organization should pursuediversity as an organizational goal Both regressions with organizationalreferents as dependent variables were signiregcant Groups with higherincreases in the proportion of women over time tended to have morefavorable valence regarding the desirability of promoting diversity as anorganizational goal but were less sure of managementrsquos commitment asthey had more experience with the reality of implementing theorganizational change

35 Regression results on positive senior women and minority effect on climateOur third research objective was to examine whether increasing the proportionof senior women and minorities at the department level positively relates togroup valence and consensus In our data analysis the proportional change inoverall tenured faculty was entered in step 2 and then the proportional change

HRM changestrategy

343

Org

aniz

atio

nD

epar

tmen

tD

iver

sity

asor

gan

izat

ional

goa

l

Com

mit

men

tof

man

agem

ent

todiv

ersi

ty

Per

ceiv

edw

ork-g

roup

mix

Equal

ity

ofdep

artm

ent

suppor

tof

wom

enE

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

Ste

p1

Dep

tsi

zeT

ime

2-T

ime

10

020

080

180

000

06M

ean

ofD

V0

35

022

017

002

029

D

R2

012

0

060

060

000

09T

otal

R2

012

0

060

060

000

09Ste

p2

Wom

enT

ime

2-T

ime

12

035

2

017

20

182

007

20

02W

hit

esT

ime

2-T

ime

12

002

20

052

003

004

021

Ten

ure

Tim

e2-

Tim

e1

009

003

022

010

000

DR

20

11

003

008

001

004

Tot

alR

20

23

009

014

001

013

Ste

p3

Whit

ete

nure

dm

ales

Tim

e2-

Tim

e1

20

092

012

005

20

302

031

DR

20

010

010

000

040

05T

otal

R2

024

0

100

140

050

18

Note

p

005

orgre

ater

p0

01or

gre

ater

Table IIRegressionpredictingconsensus onclimate for diversity

JOCM163

344

Org

aniz

atio

nD

epar

tmen

tD

iver

sity

asor

gan

izat

ional

goa

l

Com

mit

men

tof

man

agem

ent

todiv

ersi

ty

Per

ceiv

edw

ork-g

roup

mix

Equal

ity

ofdep

artm

ent

suppor

tof

wom

enE

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

Ste

p1

Dep

tsi

zeT

ime

2-T

ime

10

080

152

006

021

0

05D

R2

001

002

000

004

0

00T

otal

R2

001

002

000

004

0

00Ste

p2

Wom

enT

ime

2-T

ime

10

40

20

38

20

21

20

062

013

Whit

esT

ime

2-T

ime

12

007

019

008

20

172

001

Ten

ure

Tim

e2-

Tim

e1

20

072

013

012

016

019

DR

20

16

017

0

060

070

05T

otal

R2

017

0

19

006

011

0

05Ste

p3

Whit

ete

nure

dm

ales

Tim

e2-

Tim

e1

005

011

000

007

035

D

R2

000

001

000

000

004

T

otal

R2

017

0

20

006

011

009

Note

p

005

orgre

ater

p0

01or

gre

ater

Table IIIRegression

predicting valenceof climate for

diversity

HRM changestrategy

345

in tenured white males was entered in step 3 The results are that the thirdblock was not statistically signiregcant in the regressions showing that smallincreases in the group proportions of senior women and minorities did notincrease group consensus or valence regarding climate for change

4 DiscussionThis study explored links between actual change in organizational structureorganizational demography and work group interpretation of this change Ourresults showed that HR strategies that focus on structural change withoutworking to develop supportive group norms and positive climate areinadequate change strategies Current employer efforts to increase diversity asan isolated HR strategy may not necessarily lead to increased memberagreement that change is favorable Most change efforts like the diversitychange effort we studied are focused at the organizational or individualintervention levels and under-emphasize the work group level and theimportance of group tipping points Our study highlights a disconnect betweenorganizational level diversity practice which often assumes that hiring moreminorities and women will improve climate and what our research found at thegroup level a climate that was mixed and ambivalent at best and sometimesnegative

41 Group level measures of demographic change consensus and valence ondiversity climateThe regrst objective of this study was to develop new theory and measures ofgroup level change in diversity We provided new ways to measure and groupsocial construction of change These included consensus the degree to whichmembers held common perceptions of diversity climate and valence thedirection of those perceptions Our measures identify group and organizationallevel referents of HR strategies to increase demographic diversity Group levelreferents of diversity change were how well members perceived their workgroup mix and the perceived equality of resource allocation acrossdemographic groups Organizational referents were managementcommitment to change and whether diversity should be an organizationalstrategy assessed general support for change Our study highlights howimportant it is for research and practice to not only examine how change indemography relates to differences in individual attitudes but also to groupmember shared perceptions about the direction of the climate for changeUnderstanding linkages between HR changes (in this case group demographywhich raises unique psychological processes related to social identity) andpositive consensus is important because it is relevant to cooperation workexperiences interpersonal conmacrict stress and norms in groups striving tomanage change

JOCM163

346

42 Greater work group demographic shift does not necessarily result infavorable climateOur second objective was to assess whether work groups with higherdemographic shifts tend to have more positive climates that is higherconsensus and valence toward diversity change for both organizational andgroup level change referents Our results only partially supported this beliefthat greater structural demographic shift would positively relate to a morefavorable climate for diversity Although at the organizational level theuniversity had experienced signiregcant change in organizational demographyan increase in the overall representation of white women (36 percent) andminorities (41 percent) over time work group members in units with thegreatest change did not have higher consensus or more favorable valenceWe found more favorable results for organizational than group levelreferents of change Members may regnd it easier to agree with organizationalreferents and more difregcult to agree on group level referents due to theirgreater psychological proximity to group processes that affect how theypersonally experience conmacrict and wellbeing from change in their dailywork Work groups with more demographic change had higher valence butlower consensus on whether the organization should pursue diversity as astrategic goal When organizational change strategies to increase diversityare enacted at the work group level the group demographic shift may bediluted across the organization to only slightly increase the representation ofwomen and minorities within actual groups Due to this dilution effectgroup membersrsquo social construction of the HR strategies may lack consensusor positive valence regarding the organizational diversity change strategiesIf the change is not large enough to create a critical mass to tiprepresentation of minorities to at least 35 percent according to Kanter (1977)negative intergroup dynamics such as increased resource competition andtokenism may still occur

43 Diluting diversity gradually increasing the proportion of senior womenminorities does not necessarily improve group climateOur third objective was to investigate whether increasing the proportion ofsenior women and minorities at the department level positively relates togroup valence and consensus There was no relation between increasing theproportion of senior women and minorities and group valence andconsensus Due to diluted saliency of organizational HR change strategies toincrease diversity at the work group level slight demographic shifts toproportionally increase the representation of senior women and minoritiesdid not relate to higher group valence or consensus This increase did littleto improve group climate and may have even hurt the climate in theshort run

HRM changestrategy

347

5 Implications for future research and practice51 Incremental structural change not sufregcient to improve climateWe hope this study will encourage researchers and practitioners interested indiversity change strategies to augment individual and organizational levels ofintervention to focus on work group climates supporting change Our resultshighlight some disparities between the underlying assumptions of HRstrategies to manage organizational change (increasing structural diversitywill be positive for climate) and reality at the department level (incrementalincreases did not improve climate) Organizational strategists implementingdiversity recruitment initiatives typically assume that change strategies (in thiscase hiring diversity) would improve the group social climate due to a positivechange in member attitudes based on increased group diversity and positivesocial contact Yet our study does not support this view Although theorganization we studied had had signiregcant success in enacting the formal HRstrategy plusmn increasing the representation of women and minorities at theorganizational level plusmn our research shows that gradually shifting thedemography of work groups may not necessarily lead to increased consensusor valence about the social climate within work groups For example it did notlead to improved climate regarding the perceived fairness of resourceallocation or good mixing in social interaction In fact the incremental increasein the proportion of minority groups lowered consensus and agreement in theshort term as subgroups sorted out shifting and socially ambiguous powerrelationships within work units These results suggest the importance ofmeasuring change initiatives impacts at the group level and having group levelinterventions such as team-building and group-focused resource allocation tosupport organizational level diversity strategies

52 Focus change strategies on tipping points within targeted work unitsFuture research should consider how HR strategies to manage organizationaldemographic change are socially constructed and enacted at the work grouplevel It is especially important to identify what constitutes a group leveltipping point or critical mass across hierarchical levels and differentdemographic groups to enable positive consensus supporting changePractice might regnd it more fruitful to focus future HR strategies to increasediversity through hiring to alter tipping points within targeted work groupsThis will dramatically alter the saliency of strategic organizationaldemographic change and intervene to shape the climate of speciregc units Ourstudy suggests that organizations may need demographic earthquakes withinwork groups to effect group consensus and valence on the climate regardingthe HR change (ie managing diversity) Dilution of change at the group levelmay ameliorate intended positive effects unless certain group level tippingpoints such as at least 35 percent of the work group are effected in each unit by

JOCM163

348

the HR strategy After eight years of hiring new faculty in our sample womenhovered at about 15 percent in 89 percent of groups and minorities had amaximum representation of 25 percent The demographic shifts whileincreasing were not dramatic enough to create imbalances that markedlyincreased the salience and power of new demographic subgroups created bythe HR strategies Instead of being balanced many of our groups were stilllargely white male and remain tilted or skewed over time which Kanter (1977)notes perpetuates stereotypical negative dynamics

53 Avoiding backlash weaving tipping points across the hierarchyRecent work at MIT by Bailyn (2000) suggests that backlash and resistance toorganizational change may be occurring among senior white male faculty inresponse to HR diversity strategies Initially women and minorities now enterdepartments with relatively equal resources to white men as supported by theHR system design However by the time women and minorities reach theassociate and full professor levels subtle discrimination processes occur in howHR policies are applied such as lower access to resources promotions andpoor climate Despite the increase in upward mobility the representation ofwomen and minorities at the top is still relatively low Ely (1994) found that ifwomen are under-represented at higher organizational levels relations betweenwomen at lower levels might be of lower quality due to increased competitionand negative gender dynamics in the larger social system If HR changestrategies that alter the distribution of existing resources across organizationaland demographic groups are to be successful they must take into account hownot only tipping points in terms of overall departmental representation butmore importantly how tipping points are distributed in demographies acrossthe hierarchies of work groups or departments

HR diversity change strategies that bring in many new hires of differentbackgrounds may have negative ramiregcations for social functioning in groupsparticularly if new members are not supported by additional HR strategies thatallow them to enter work groups on an equal footing or if their work groups donot have leadership that mirror their demographic identities As Kanter (1983)found identity groups must be equal in their access to power resources ifcontact is to produce improvement in attitudes There must also be time for thenew members to be socialized and integrated into the culture Harrison et al(1998) found that the longer the length of time that members worked togetherthe greater the effects of deep level attitude similarity Organizations areincreasing surface (demographic) and deep level (attitudinal) diversitysimultaneously which is likely to have negative short-term ramiregcations forthe consensus and direction of climate We argue HR strategies must not onlyfocus on the surface level plusmn remacrecting structural and demographic attributes plusmnbut also simultaneously to deep level characteristics remacrecting values within

HRM changestrategy

349

the context of speciregc groups at all organizational levels In sum HR strategistsand change agents should focus on how HR change strategies relate to the jointalignment of group memberrsquos multiple characteristics (eg demographichierarchical status task) in work groups plusmn that is to examine group contextualinmacruences (Wharton 1992)

54 LimitationsOur study while making contributions to organizational research is notwithout limitations Though large in individual respondents by utilizing thegroup as the level of analysis our sample size includes only 74 groups Thiscould reduce the level of statistical power of our regndings Additionally becausethe sample was based on one organization it is important to replicate thisanalysis across multiple organizations to substantiate our results Althoughchanges in department heterogeneity across an eight-year period wereincluded this is not a longitudinal design Future research should employrepeated measures over a length of time across all variables measuringdemographic change in heterogeneity as well as consensus and valence Ourstudy points to some of the challenges in measuring ordfsuccess in increasingdiversityordm as we did not look at annual turnover or the effects of the glassceiling at the group level in this study which is a gap we hope futureresearchers will be motivated to study spurred by our research

Since little published work has been done on group consensus regarding theeffects of HR policies to increase diversity through demographic change weextend many social science literatures ranging from organizationaldevelopment to demography and HR Future research on diversity andorganizational change also should strive to be interdisciplinary as theliterature on HR strategy and change has been largely developed in the HRliterature separate from other social science disciplines reducing theoreticalintegration Without such integration and increased scholarly and practitionerattention to climate and consensus many organizations may fail to supportgroup membersrsquo explicit use of their cultural experiences to advanceorganizational learning from enacted diversity change strategies despite thegroup and intergroup social tensions that will inevitably occur (Thomas andEly 1996)

References

Aiken LA and West SG (1993) Multiple Regression Testing and Interpreting InteractionsSage Publications Newbury Park CA

Alderfer CP and Smith KK (1982) ordfStudying embedded intergroup relations inorganizationsordm Administrative Science Quarterly Vol 27 pp 35-65

Allport G (1954) The Nature of Prejudice Addison Wesley Cambridge MA

Bailyn L (2000) ordfGender and diversity in organizations preconference presentationordm paperpresented at the National Academy of Management meeting Toronto

JOCM163

350

Blalock HM (1967) Toward a Theory of Minority Group Relations Wiley New York NY

Bleise PD and Halverson RR (1998) ordfGroup consensus and psychological well-being a largeregeld studyordm Journal of Applied Social Psychology Vol 28 pp 563-80

Brown KG Kozlowski SWJ and Hattrup K (1996) Theory Issues and Recommendations inConceptualizing Agreement as a Construct in Organizational Research The Search forConsensus Regarding Consensus National Academy of Management Cincinnati OH

Cox T Jr (1993) Cultural Diversity in Organizations Berrett-Koehler San Francisco CA

Ely R (1994) ordfThe effects of organizational demographics and social identity on relationshipsamong professional womenordm Administrative Science Quarterly Vol 39 pp 203-38

Ely R (1995) ordfThe power in demography womenrsquos social constructions of gender identity atworkordm Academy of Management Journal Vol 38 pp 589-634

Festinger L (1954) ordfA theory of social comparison processesordm Human Relations Vol 7pp 117-40

Forsyth DR (1990) Group Dynamics BrooksCole Publishing Paciregc Grove CA

Harrison DA Price KH and Bell MP (1998) ordfBeyond relational demography time and theeffects of surface and deep-level diversity on work group cohesionordm Academy ofManagement Journal Vol 41 pp 96-107

James K Lovato C and Khoo G (1994) ordfSocial identity correlates of minority workersrsquo healthordmAcademy of Management Journal Vol 37 p 383+

Kanter R (1977) ordfSome effects of proportions on group life skewed sex ratios and reactions totoken womenordm American Journal of Sociology Vol 82 pp 965-90

Kanter R (1983) Men and Women of the Corporation Basic Books New York NY

Kossek EE and Lobel SA (1996) ordfIntroduction transforming human resource systems tomanage diversityordm in Managing Diversity Human Resource Strategies for Transforming

the Workplace Blackwell Oxford pp 1-19

Kossek EE and Zonia SC (1993) ordfAssessing diversity climate a regeld study of reactions toemployer efforts to promote diversityordm Journal of Organizational Behavior Vol 14pp 61-81

Larkey LK (1996) ordfToward a theory of communicative interactions in culturally diverseworkgroupsordm Academy of Management Review Vol 21 pp 463-91

Lau DC and Murnighan JK (1998) ordfDemographic diversity and faultlines the compositionaldynamics of organizational groupsordm Academy of Management Review Vol 29 pp 325-40

Lindell MK and Brandt CJ (2000) ordfClimate quality and climate consensus as mediators of therelationship between organizational antecedents and outcomesordm Journal of AppliedPsychology Vol 85 pp 331-48

McGrath JE Berdahl JL and Arrow H (1995) ordfTraits expectations culture and clout thedynamics of diversity in work groupsordm in Jackson SE and Ruderman MN (Eds)Diversity in Work Teams Research Paradigms for a Changing Workplace AmericanPsychological Association Washington DC

Payne RL and Pugh DS (1976) ordfOrganizational structure and climateordm in Dunnette MD(Ed) Handbook of Industrial and Organizational Psychology Rand McNally Chicago ILpp 1125-73

Reichers A and Schneider B (1990) ordfClimate and culture an evolution of constructsordm inSchneider B (Ed) Organizational Climate and Culture Jossey-Bass San Francisco CA

HRM changestrategy

351

Var

iable

Mea

nSD

12

34

5

1D

epar

tmen

tsi

zeT

ime

125

52

160

4plusmn

2D

epar

tmen

tsi

zeT

ime

229

06

176

60

89

plusmn3

Pro

por

tion

ofw

hit

esT

ime

10

900

010

210

21plusmn

4P

ropor

tion

ofw

hit

esT

ime

20

810

180

070

120

47

plusmn5

Pro

por

tion

ofw

omen

Tim

e1

019

021

005

002

001

017

plusmn6

Pro

por

tion

ofw

omen

Tim

e2

043

027

006

006

017

035

0

58

7P

ropor

tion

wit

hte

nure

Tim

e1

081

015

010

010

20

152

005

20

098

Pro

por

tion

wit

hT

enure

Tim

e2

084

011

004

003

20

082

018

20

129

Pro

por

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

034

019

20

042

009

20

100

000

76

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

039

018

20

092

010

20

39

20

130

65

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

247

550

40

062

0

59

20

092

013

009

12C

han

ge

inte

nure

tim

e1-

Tim

e2

001

001

20

152

013

20

28

20

120

0613

C

han

ge

ingen

der

Tim

e1-

Tim

e2

001

007

004

20

020

090

100

76

14C

han

ge

inw

hit

esT

ime

1-T

ime

20

010

032

009

20

162

059

2

005

2

008

15D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s2

100

057

001

001

20

192

015

024

16

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s2

099

048

007

014

20

032

006

008

17P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s2

125

047

001

011

002

20

060

1218

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s2

024

016

007

007

015

009

001

19E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s2

025

021

006

011

20

050

142

006

20D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n3

900

382

003

20

052

023

2

007

039

21

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n3

530

402

010

001

20

222

007

029

22P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

314

048

20

212

021

20

112

017

030

23E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n2

450

292

009

20

032

017

20

31

002

24E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

393

050

20

010

020

072

010

20

1625

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

361

049

20

030

012

018

20

050

2026

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

324

057

20

092

003

20

24

001

014

27P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n2

500

372

009

20

112

003

000

027

28

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

200

014

20

060

022

013

20

242

008

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n1

920

152

018

20

020

070

062

010

(con

tinued

)

Table IMeans standardsdeviations andcorrelations ofvariables

JOCM163

338

Var

iable

67

89

1011

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

plusmn7

Pro

por

tion

wit

hte

nure

Tim

e1

20

20plusmn

8P

ropor

tion

wit

hte

nure

Tim

e2

20

190

31plusmn

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

054

2

058

2

029

plusmn

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

048

2

018

20

32

073

plusmn

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

004

009

012

006

003

plusmn12

C

han

ge

inte

nure

Tim

e1-

Tim

e2

20

062

016

022

014

006

008

13C

han

ge

ingen

der

Tim

e1-

Tim

e2

032

2

019

20

130

68

055

0

0514

C

han

ge

inw

hit

esT

ime

1-T

ime

22

018

023

014

001

034

0

1315

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s0

010

020

080

140

110

0816

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s2

013

001

000

009

003

007

17P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s2

008

20

170

070

170

092

011

18E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s2

005

20

162

009

007

20

080

1519

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s2

011

011

008

20

102

023

004

20D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n0

57

20

022

014

036

0

48

014

21C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n0

030

182

001

016

026

0

0122

P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

007

20

040

050

32

030

0

0423

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n2

013

001

012

008

011

005

24E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

20

33

20

052

007

20

072

012

20

0225

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

048

0

022

004

020

027

0

1626

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

20

110

140

010

030

172

003

27P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n0

050

030

190

210

150

0628

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

20

092

004

022

20

012

002

004

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n2

026

2

018

20

102

007

20

162

024

(con

tinued

)

Table I

HRM changestrategy

339

Var

iable

1213

1415

1617

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

7P

ropor

tion

wit

hte

nure

Tim

e1

8P

ropor

tion

wit

hte

nure

Tim

e2

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

12C

han

ge

inte

nure

Tim

e1-

Tim

e2

plusmn13

C

han

ge

ingen

der

Tim

e1-

Tim

e2

001

plusmn14

C

han

ge

inw

hit

esT

ime

1-T

ime

20

162

007

plusmn15

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s0

150

110

14plusmn

16C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s2

013

021

006

20

07plusmn

17P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s0

150

130

060

38

034

plusmn

18E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s0

130

182

001

20

110

45

008

19E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s0

000

040

042

011

042

0

0020

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n0

000

050

150

34

20

32

20

0721

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n2

020

022

022

022

023

002

22P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

010

011

018

017

010

016

23E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n0

022

009

015

20

090

062

007

24E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

001

007

011

20

062

019

007

25D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

005

20

060

100

080

012

010

26C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

20

27

009

021

024

20

030

0227

P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n0

070

040

120

232

003

013

28E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

010

20

190

102

003

20

142

014

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n0

022

002

009

000

001

020

(con

tinued

)

Table I

JOCM163

340

Var

iable

1819

2021

2223

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

7P

ropor

tion

wit

hte

nure

Tim

e1

8P

ropor

tion

wit

hte

nure

Tim

e2

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

12C

han

ge

inte

nure

Tim

e1-

Tim

e2

13C

han

ge

ingen

der

Tim

e1-

Tim

e2

14C

han

ge

inw

hit

esT

ime

1-T

ime

215

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s16

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s17

P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s18

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

splusmn

19E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s0

57

plusmn20

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n2

035

2

027

plusmn

21C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n0

180

180

08plusmn

22P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

010

20

050

150

27

plusmn23

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n0

010

002

008

008

014

plusmn24

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

030

0

34

20

26

026

0

192

021

25D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

20

222

013

083

0

040

082

007

26C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

20

070

020

150

79

018

000

27P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n2

002

20

040

140

120

84

012

28E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

20

052

009

20

020

022

004

086

29

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n0

140

31

20

210

100

042

008

(con

tinued

)

Table I

HRM changestrategy

341

Table I

Var

iable

2425

2627

2829

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

7P

ropor

tion

wit

hte

nure

Tim

e1

8P

ropor

tion

wit

hte

nure

Tim

e2

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

12C

han

ge

inte

nure

Tim

e1-

Tim

e2

13C

han

ge

ingen

der

Tim

e1-

Tim

e2

14C

han

ge

inw

hit

esT

ime

1-T

ime

215

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s16

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s17

P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s18

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s19

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s20

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n21

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n22

P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

23E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n24

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

plusmn25

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

20

20plusmn

26C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

025

002

plusmn27

P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n0

080

010

19plusmn

28E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

20

32

000

20

010

05plusmn

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n0

58

20

240

180

132

003

plusmn

Note

p

001

orgre

ater

JOCM163

342

support of resources for women and men and whites and non-whites asthese measures hovered at plusmn024 and 2 025 or close to 0 which would betotal consensus There was the least consensus that minorities and majoritiesmixed well (plusmn125) with slightly more consensus that diversity should bepursued as organizational (plusmn10) The most positive valence was theperception of equality of department support for minorities (393) and thatdiversity should be pursued as a goal (39)

34 Regression results on demographic shift and climateTo address our second research objective which was to examine whetherwork groups with higher demographic shifts had more positive groupclimates for each of the regve dependent variables separate hierarchicalregressions were run to examine the relation between group demographicchange and group consensus (see Table II) or valence (see Table III) In theregrst step of all regressions proportional change in size was entered as acontrol For the consensus regressions the current mean of each dependentvariable was also entered as a control In step 2 the group proportionalchange in women and the group proportional change in minorities wereentered (In early analyses for the second step of each regression interactionterms change in gender or race by tenure were included following Aikenand Westrsquos (1993) procedures No signiregcant interactions were found) Usingstandardized beta weights as seen in step 2 of Tables II and III ourhypothesis that work groups with higher demographic shifts may havehigher consensus and valence toward diversity change received partialsupport Group referents of diversity climate were unchanged bydemographic shifts remacrecting possible dilution of HR strategiesOrganizational referents of valence were effected but consensus wasreduced in terms of whether the organization should pursue diversity as anorganizational goal Departments with higher shifts in the proportion ofwomen had lower consensus on whether the organization should pursuediversity as an organizational goal Both regressions with organizationalreferents as dependent variables were signiregcant Groups with higherincreases in the proportion of women over time tended to have morefavorable valence regarding the desirability of promoting diversity as anorganizational goal but were less sure of managementrsquos commitment asthey had more experience with the reality of implementing theorganizational change

35 Regression results on positive senior women and minority effect on climateOur third research objective was to examine whether increasing the proportionof senior women and minorities at the department level positively relates togroup valence and consensus In our data analysis the proportional change inoverall tenured faculty was entered in step 2 and then the proportional change

HRM changestrategy

343

Org

aniz

atio

nD

epar

tmen

tD

iver

sity

asor

gan

izat

ional

goa

l

Com

mit

men

tof

man

agem

ent

todiv

ersi

ty

Per

ceiv

edw

ork-g

roup

mix

Equal

ity

ofdep

artm

ent

suppor

tof

wom

enE

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

Ste

p1

Dep

tsi

zeT

ime

2-T

ime

10

020

080

180

000

06M

ean

ofD

V0

35

022

017

002

029

D

R2

012

0

060

060

000

09T

otal

R2

012

0

060

060

000

09Ste

p2

Wom

enT

ime

2-T

ime

12

035

2

017

20

182

007

20

02W

hit

esT

ime

2-T

ime

12

002

20

052

003

004

021

Ten

ure

Tim

e2-

Tim

e1

009

003

022

010

000

DR

20

11

003

008

001

004

Tot

alR

20

23

009

014

001

013

Ste

p3

Whit

ete

nure

dm

ales

Tim

e2-

Tim

e1

20

092

012

005

20

302

031

DR

20

010

010

000

040

05T

otal

R2

024

0

100

140

050

18

Note

p

005

orgre

ater

p0

01or

gre

ater

Table IIRegressionpredictingconsensus onclimate for diversity

JOCM163

344

Org

aniz

atio

nD

epar

tmen

tD

iver

sity

asor

gan

izat

ional

goa

l

Com

mit

men

tof

man

agem

ent

todiv

ersi

ty

Per

ceiv

edw

ork-g

roup

mix

Equal

ity

ofdep

artm

ent

suppor

tof

wom

enE

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

Ste

p1

Dep

tsi

zeT

ime

2-T

ime

10

080

152

006

021

0

05D

R2

001

002

000

004

0

00T

otal

R2

001

002

000

004

0

00Ste

p2

Wom

enT

ime

2-T

ime

10

40

20

38

20

21

20

062

013

Whit

esT

ime

2-T

ime

12

007

019

008

20

172

001

Ten

ure

Tim

e2-

Tim

e1

20

072

013

012

016

019

DR

20

16

017

0

060

070

05T

otal

R2

017

0

19

006

011

0

05Ste

p3

Whit

ete

nure

dm

ales

Tim

e2-

Tim

e1

005

011

000

007

035

D

R2

000

001

000

000

004

T

otal

R2

017

0

20

006

011

009

Note

p

005

orgre

ater

p0

01or

gre

ater

Table IIIRegression

predicting valenceof climate for

diversity

HRM changestrategy

345

in tenured white males was entered in step 3 The results are that the thirdblock was not statistically signiregcant in the regressions showing that smallincreases in the group proportions of senior women and minorities did notincrease group consensus or valence regarding climate for change

4 DiscussionThis study explored links between actual change in organizational structureorganizational demography and work group interpretation of this change Ourresults showed that HR strategies that focus on structural change withoutworking to develop supportive group norms and positive climate areinadequate change strategies Current employer efforts to increase diversity asan isolated HR strategy may not necessarily lead to increased memberagreement that change is favorable Most change efforts like the diversitychange effort we studied are focused at the organizational or individualintervention levels and under-emphasize the work group level and theimportance of group tipping points Our study highlights a disconnect betweenorganizational level diversity practice which often assumes that hiring moreminorities and women will improve climate and what our research found at thegroup level a climate that was mixed and ambivalent at best and sometimesnegative

41 Group level measures of demographic change consensus and valence ondiversity climateThe regrst objective of this study was to develop new theory and measures ofgroup level change in diversity We provided new ways to measure and groupsocial construction of change These included consensus the degree to whichmembers held common perceptions of diversity climate and valence thedirection of those perceptions Our measures identify group and organizationallevel referents of HR strategies to increase demographic diversity Group levelreferents of diversity change were how well members perceived their workgroup mix and the perceived equality of resource allocation acrossdemographic groups Organizational referents were managementcommitment to change and whether diversity should be an organizationalstrategy assessed general support for change Our study highlights howimportant it is for research and practice to not only examine how change indemography relates to differences in individual attitudes but also to groupmember shared perceptions about the direction of the climate for changeUnderstanding linkages between HR changes (in this case group demographywhich raises unique psychological processes related to social identity) andpositive consensus is important because it is relevant to cooperation workexperiences interpersonal conmacrict stress and norms in groups striving tomanage change

JOCM163

346

42 Greater work group demographic shift does not necessarily result infavorable climateOur second objective was to assess whether work groups with higherdemographic shifts tend to have more positive climates that is higherconsensus and valence toward diversity change for both organizational andgroup level change referents Our results only partially supported this beliefthat greater structural demographic shift would positively relate to a morefavorable climate for diversity Although at the organizational level theuniversity had experienced signiregcant change in organizational demographyan increase in the overall representation of white women (36 percent) andminorities (41 percent) over time work group members in units with thegreatest change did not have higher consensus or more favorable valenceWe found more favorable results for organizational than group levelreferents of change Members may regnd it easier to agree with organizationalreferents and more difregcult to agree on group level referents due to theirgreater psychological proximity to group processes that affect how theypersonally experience conmacrict and wellbeing from change in their dailywork Work groups with more demographic change had higher valence butlower consensus on whether the organization should pursue diversity as astrategic goal When organizational change strategies to increase diversityare enacted at the work group level the group demographic shift may bediluted across the organization to only slightly increase the representation ofwomen and minorities within actual groups Due to this dilution effectgroup membersrsquo social construction of the HR strategies may lack consensusor positive valence regarding the organizational diversity change strategiesIf the change is not large enough to create a critical mass to tiprepresentation of minorities to at least 35 percent according to Kanter (1977)negative intergroup dynamics such as increased resource competition andtokenism may still occur

43 Diluting diversity gradually increasing the proportion of senior womenminorities does not necessarily improve group climateOur third objective was to investigate whether increasing the proportion ofsenior women and minorities at the department level positively relates togroup valence and consensus There was no relation between increasing theproportion of senior women and minorities and group valence andconsensus Due to diluted saliency of organizational HR change strategies toincrease diversity at the work group level slight demographic shifts toproportionally increase the representation of senior women and minoritiesdid not relate to higher group valence or consensus This increase did littleto improve group climate and may have even hurt the climate in theshort run

HRM changestrategy

347

5 Implications for future research and practice51 Incremental structural change not sufregcient to improve climateWe hope this study will encourage researchers and practitioners interested indiversity change strategies to augment individual and organizational levels ofintervention to focus on work group climates supporting change Our resultshighlight some disparities between the underlying assumptions of HRstrategies to manage organizational change (increasing structural diversitywill be positive for climate) and reality at the department level (incrementalincreases did not improve climate) Organizational strategists implementingdiversity recruitment initiatives typically assume that change strategies (in thiscase hiring diversity) would improve the group social climate due to a positivechange in member attitudes based on increased group diversity and positivesocial contact Yet our study does not support this view Although theorganization we studied had had signiregcant success in enacting the formal HRstrategy plusmn increasing the representation of women and minorities at theorganizational level plusmn our research shows that gradually shifting thedemography of work groups may not necessarily lead to increased consensusor valence about the social climate within work groups For example it did notlead to improved climate regarding the perceived fairness of resourceallocation or good mixing in social interaction In fact the incremental increasein the proportion of minority groups lowered consensus and agreement in theshort term as subgroups sorted out shifting and socially ambiguous powerrelationships within work units These results suggest the importance ofmeasuring change initiatives impacts at the group level and having group levelinterventions such as team-building and group-focused resource allocation tosupport organizational level diversity strategies

52 Focus change strategies on tipping points within targeted work unitsFuture research should consider how HR strategies to manage organizationaldemographic change are socially constructed and enacted at the work grouplevel It is especially important to identify what constitutes a group leveltipping point or critical mass across hierarchical levels and differentdemographic groups to enable positive consensus supporting changePractice might regnd it more fruitful to focus future HR strategies to increasediversity through hiring to alter tipping points within targeted work groupsThis will dramatically alter the saliency of strategic organizationaldemographic change and intervene to shape the climate of speciregc units Ourstudy suggests that organizations may need demographic earthquakes withinwork groups to effect group consensus and valence on the climate regardingthe HR change (ie managing diversity) Dilution of change at the group levelmay ameliorate intended positive effects unless certain group level tippingpoints such as at least 35 percent of the work group are effected in each unit by

JOCM163

348

the HR strategy After eight years of hiring new faculty in our sample womenhovered at about 15 percent in 89 percent of groups and minorities had amaximum representation of 25 percent The demographic shifts whileincreasing were not dramatic enough to create imbalances that markedlyincreased the salience and power of new demographic subgroups created bythe HR strategies Instead of being balanced many of our groups were stilllargely white male and remain tilted or skewed over time which Kanter (1977)notes perpetuates stereotypical negative dynamics

53 Avoiding backlash weaving tipping points across the hierarchyRecent work at MIT by Bailyn (2000) suggests that backlash and resistance toorganizational change may be occurring among senior white male faculty inresponse to HR diversity strategies Initially women and minorities now enterdepartments with relatively equal resources to white men as supported by theHR system design However by the time women and minorities reach theassociate and full professor levels subtle discrimination processes occur in howHR policies are applied such as lower access to resources promotions andpoor climate Despite the increase in upward mobility the representation ofwomen and minorities at the top is still relatively low Ely (1994) found that ifwomen are under-represented at higher organizational levels relations betweenwomen at lower levels might be of lower quality due to increased competitionand negative gender dynamics in the larger social system If HR changestrategies that alter the distribution of existing resources across organizationaland demographic groups are to be successful they must take into account hownot only tipping points in terms of overall departmental representation butmore importantly how tipping points are distributed in demographies acrossthe hierarchies of work groups or departments

HR diversity change strategies that bring in many new hires of differentbackgrounds may have negative ramiregcations for social functioning in groupsparticularly if new members are not supported by additional HR strategies thatallow them to enter work groups on an equal footing or if their work groups donot have leadership that mirror their demographic identities As Kanter (1983)found identity groups must be equal in their access to power resources ifcontact is to produce improvement in attitudes There must also be time for thenew members to be socialized and integrated into the culture Harrison et al(1998) found that the longer the length of time that members worked togetherthe greater the effects of deep level attitude similarity Organizations areincreasing surface (demographic) and deep level (attitudinal) diversitysimultaneously which is likely to have negative short-term ramiregcations forthe consensus and direction of climate We argue HR strategies must not onlyfocus on the surface level plusmn remacrecting structural and demographic attributes plusmnbut also simultaneously to deep level characteristics remacrecting values within

HRM changestrategy

349

the context of speciregc groups at all organizational levels In sum HR strategistsand change agents should focus on how HR change strategies relate to the jointalignment of group memberrsquos multiple characteristics (eg demographichierarchical status task) in work groups plusmn that is to examine group contextualinmacruences (Wharton 1992)

54 LimitationsOur study while making contributions to organizational research is notwithout limitations Though large in individual respondents by utilizing thegroup as the level of analysis our sample size includes only 74 groups Thiscould reduce the level of statistical power of our regndings Additionally becausethe sample was based on one organization it is important to replicate thisanalysis across multiple organizations to substantiate our results Althoughchanges in department heterogeneity across an eight-year period wereincluded this is not a longitudinal design Future research should employrepeated measures over a length of time across all variables measuringdemographic change in heterogeneity as well as consensus and valence Ourstudy points to some of the challenges in measuring ordfsuccess in increasingdiversityordm as we did not look at annual turnover or the effects of the glassceiling at the group level in this study which is a gap we hope futureresearchers will be motivated to study spurred by our research

Since little published work has been done on group consensus regarding theeffects of HR policies to increase diversity through demographic change weextend many social science literatures ranging from organizationaldevelopment to demography and HR Future research on diversity andorganizational change also should strive to be interdisciplinary as theliterature on HR strategy and change has been largely developed in the HRliterature separate from other social science disciplines reducing theoreticalintegration Without such integration and increased scholarly and practitionerattention to climate and consensus many organizations may fail to supportgroup membersrsquo explicit use of their cultural experiences to advanceorganizational learning from enacted diversity change strategies despite thegroup and intergroup social tensions that will inevitably occur (Thomas andEly 1996)

References

Aiken LA and West SG (1993) Multiple Regression Testing and Interpreting InteractionsSage Publications Newbury Park CA

Alderfer CP and Smith KK (1982) ordfStudying embedded intergroup relations inorganizationsordm Administrative Science Quarterly Vol 27 pp 35-65

Allport G (1954) The Nature of Prejudice Addison Wesley Cambridge MA

Bailyn L (2000) ordfGender and diversity in organizations preconference presentationordm paperpresented at the National Academy of Management meeting Toronto

JOCM163

350

Blalock HM (1967) Toward a Theory of Minority Group Relations Wiley New York NY

Bleise PD and Halverson RR (1998) ordfGroup consensus and psychological well-being a largeregeld studyordm Journal of Applied Social Psychology Vol 28 pp 563-80

Brown KG Kozlowski SWJ and Hattrup K (1996) Theory Issues and Recommendations inConceptualizing Agreement as a Construct in Organizational Research The Search forConsensus Regarding Consensus National Academy of Management Cincinnati OH

Cox T Jr (1993) Cultural Diversity in Organizations Berrett-Koehler San Francisco CA

Ely R (1994) ordfThe effects of organizational demographics and social identity on relationshipsamong professional womenordm Administrative Science Quarterly Vol 39 pp 203-38

Ely R (1995) ordfThe power in demography womenrsquos social constructions of gender identity atworkordm Academy of Management Journal Vol 38 pp 589-634

Festinger L (1954) ordfA theory of social comparison processesordm Human Relations Vol 7pp 117-40

Forsyth DR (1990) Group Dynamics BrooksCole Publishing Paciregc Grove CA

Harrison DA Price KH and Bell MP (1998) ordfBeyond relational demography time and theeffects of surface and deep-level diversity on work group cohesionordm Academy ofManagement Journal Vol 41 pp 96-107

James K Lovato C and Khoo G (1994) ordfSocial identity correlates of minority workersrsquo healthordmAcademy of Management Journal Vol 37 p 383+

Kanter R (1977) ordfSome effects of proportions on group life skewed sex ratios and reactions totoken womenordm American Journal of Sociology Vol 82 pp 965-90

Kanter R (1983) Men and Women of the Corporation Basic Books New York NY

Kossek EE and Lobel SA (1996) ordfIntroduction transforming human resource systems tomanage diversityordm in Managing Diversity Human Resource Strategies for Transforming

the Workplace Blackwell Oxford pp 1-19

Kossek EE and Zonia SC (1993) ordfAssessing diversity climate a regeld study of reactions toemployer efforts to promote diversityordm Journal of Organizational Behavior Vol 14pp 61-81

Larkey LK (1996) ordfToward a theory of communicative interactions in culturally diverseworkgroupsordm Academy of Management Review Vol 21 pp 463-91

Lau DC and Murnighan JK (1998) ordfDemographic diversity and faultlines the compositionaldynamics of organizational groupsordm Academy of Management Review Vol 29 pp 325-40

Lindell MK and Brandt CJ (2000) ordfClimate quality and climate consensus as mediators of therelationship between organizational antecedents and outcomesordm Journal of AppliedPsychology Vol 85 pp 331-48

McGrath JE Berdahl JL and Arrow H (1995) ordfTraits expectations culture and clout thedynamics of diversity in work groupsordm in Jackson SE and Ruderman MN (Eds)Diversity in Work Teams Research Paradigms for a Changing Workplace AmericanPsychological Association Washington DC

Payne RL and Pugh DS (1976) ordfOrganizational structure and climateordm in Dunnette MD(Ed) Handbook of Industrial and Organizational Psychology Rand McNally Chicago ILpp 1125-73

Reichers A and Schneider B (1990) ordfClimate and culture an evolution of constructsordm inSchneider B (Ed) Organizational Climate and Culture Jossey-Bass San Francisco CA

HRM changestrategy

351

Var

iable

67

89

1011

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

plusmn7

Pro

por

tion

wit

hte

nure

Tim

e1

20

20plusmn

8P

ropor

tion

wit

hte

nure

Tim

e2

20

190

31plusmn

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

054

2

058

2

029

plusmn

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

048

2

018

20

32

073

plusmn

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

004

009

012

006

003

plusmn12

C

han

ge

inte

nure

Tim

e1-

Tim

e2

20

062

016

022

014

006

008

13C

han

ge

ingen

der

Tim

e1-

Tim

e2

032

2

019

20

130

68

055

0

0514

C

han

ge

inw

hit

esT

ime

1-T

ime

22

018

023

014

001

034

0

1315

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s0

010

020

080

140

110

0816

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s2

013

001

000

009

003

007

17P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s2

008

20

170

070

170

092

011

18E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s2

005

20

162

009

007

20

080

1519

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s2

011

011

008

20

102

023

004

20D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n0

57

20

022

014

036

0

48

014

21C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n0

030

182

001

016

026

0

0122

P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

007

20

040

050

32

030

0

0423

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n2

013

001

012

008

011

005

24E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

20

33

20

052

007

20

072

012

20

0225

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

048

0

022

004

020

027

0

1626

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

20

110

140

010

030

172

003

27P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n0

050

030

190

210

150

0628

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

20

092

004

022

20

012

002

004

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n2

026

2

018

20

102

007

20

162

024

(con

tinued

)

Table I

HRM changestrategy

339

Var

iable

1213

1415

1617

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

7P

ropor

tion

wit

hte

nure

Tim

e1

8P

ropor

tion

wit

hte

nure

Tim

e2

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

12C

han

ge

inte

nure

Tim

e1-

Tim

e2

plusmn13

C

han

ge

ingen

der

Tim

e1-

Tim

e2

001

plusmn14

C

han

ge

inw

hit

esT

ime

1-T

ime

20

162

007

plusmn15

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s0

150

110

14plusmn

16C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s2

013

021

006

20

07plusmn

17P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s0

150

130

060

38

034

plusmn

18E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s0

130

182

001

20

110

45

008

19E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s0

000

040

042

011

042

0

0020

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n0

000

050

150

34

20

32

20

0721

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n2

020

022

022

022

023

002

22P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

010

011

018

017

010

016

23E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n0

022

009

015

20

090

062

007

24E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

001

007

011

20

062

019

007

25D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

005

20

060

100

080

012

010

26C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

20

27

009

021

024

20

030

0227

P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n0

070

040

120

232

003

013

28E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

010

20

190

102

003

20

142

014

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n0

022

002

009

000

001

020

(con

tinued

)

Table I

JOCM163

340

Var

iable

1819

2021

2223

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

7P

ropor

tion

wit

hte

nure

Tim

e1

8P

ropor

tion

wit

hte

nure

Tim

e2

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

12C

han

ge

inte

nure

Tim

e1-

Tim

e2

13C

han

ge

ingen

der

Tim

e1-

Tim

e2

14C

han

ge

inw

hit

esT

ime

1-T

ime

215

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s16

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s17

P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s18

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

splusmn

19E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s0

57

plusmn20

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n2

035

2

027

plusmn

21C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n0

180

180

08plusmn

22P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

010

20

050

150

27

plusmn23

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n0

010

002

008

008

014

plusmn24

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

030

0

34

20

26

026

0

192

021

25D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

20

222

013

083

0

040

082

007

26C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

20

070

020

150

79

018

000

27P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n2

002

20

040

140

120

84

012

28E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

20

052

009

20

020

022

004

086

29

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n0

140

31

20

210

100

042

008

(con

tinued

)

Table I

HRM changestrategy

341

Table I

Var

iable

2425

2627

2829

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

7P

ropor

tion

wit

hte

nure

Tim

e1

8P

ropor

tion

wit

hte

nure

Tim

e2

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

12C

han

ge

inte

nure

Tim

e1-

Tim

e2

13C

han

ge

ingen

der

Tim

e1-

Tim

e2

14C

han

ge

inw

hit

esT

ime

1-T

ime

215

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s16

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s17

P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s18

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s19

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s20

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n21

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n22

P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

23E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n24

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

plusmn25

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

20

20plusmn

26C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

025

002

plusmn27

P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n0

080

010

19plusmn

28E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

20

32

000

20

010

05plusmn

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n0

58

20

240

180

132

003

plusmn

Note

p

001

orgre

ater

JOCM163

342

support of resources for women and men and whites and non-whites asthese measures hovered at plusmn024 and 2 025 or close to 0 which would betotal consensus There was the least consensus that minorities and majoritiesmixed well (plusmn125) with slightly more consensus that diversity should bepursued as organizational (plusmn10) The most positive valence was theperception of equality of department support for minorities (393) and thatdiversity should be pursued as a goal (39)

34 Regression results on demographic shift and climateTo address our second research objective which was to examine whetherwork groups with higher demographic shifts had more positive groupclimates for each of the regve dependent variables separate hierarchicalregressions were run to examine the relation between group demographicchange and group consensus (see Table II) or valence (see Table III) In theregrst step of all regressions proportional change in size was entered as acontrol For the consensus regressions the current mean of each dependentvariable was also entered as a control In step 2 the group proportionalchange in women and the group proportional change in minorities wereentered (In early analyses for the second step of each regression interactionterms change in gender or race by tenure were included following Aikenand Westrsquos (1993) procedures No signiregcant interactions were found) Usingstandardized beta weights as seen in step 2 of Tables II and III ourhypothesis that work groups with higher demographic shifts may havehigher consensus and valence toward diversity change received partialsupport Group referents of diversity climate were unchanged bydemographic shifts remacrecting possible dilution of HR strategiesOrganizational referents of valence were effected but consensus wasreduced in terms of whether the organization should pursue diversity as anorganizational goal Departments with higher shifts in the proportion ofwomen had lower consensus on whether the organization should pursuediversity as an organizational goal Both regressions with organizationalreferents as dependent variables were signiregcant Groups with higherincreases in the proportion of women over time tended to have morefavorable valence regarding the desirability of promoting diversity as anorganizational goal but were less sure of managementrsquos commitment asthey had more experience with the reality of implementing theorganizational change

35 Regression results on positive senior women and minority effect on climateOur third research objective was to examine whether increasing the proportionof senior women and minorities at the department level positively relates togroup valence and consensus In our data analysis the proportional change inoverall tenured faculty was entered in step 2 and then the proportional change

HRM changestrategy

343

Org

aniz

atio

nD

epar

tmen

tD

iver

sity

asor

gan

izat

ional

goa

l

Com

mit

men

tof

man

agem

ent

todiv

ersi

ty

Per

ceiv

edw

ork-g

roup

mix

Equal

ity

ofdep

artm

ent

suppor

tof

wom

enE

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

Ste

p1

Dep

tsi

zeT

ime

2-T

ime

10

020

080

180

000

06M

ean

ofD

V0

35

022

017

002

029

D

R2

012

0

060

060

000

09T

otal

R2

012

0

060

060

000

09Ste

p2

Wom

enT

ime

2-T

ime

12

035

2

017

20

182

007

20

02W

hit

esT

ime

2-T

ime

12

002

20

052

003

004

021

Ten

ure

Tim

e2-

Tim

e1

009

003

022

010

000

DR

20

11

003

008

001

004

Tot

alR

20

23

009

014

001

013

Ste

p3

Whit

ete

nure

dm

ales

Tim

e2-

Tim

e1

20

092

012

005

20

302

031

DR

20

010

010

000

040

05T

otal

R2

024

0

100

140

050

18

Note

p

005

orgre

ater

p0

01or

gre

ater

Table IIRegressionpredictingconsensus onclimate for diversity

JOCM163

344

Org

aniz

atio

nD

epar

tmen

tD

iver

sity

asor

gan

izat

ional

goa

l

Com

mit

men

tof

man

agem

ent

todiv

ersi

ty

Per

ceiv

edw

ork-g

roup

mix

Equal

ity

ofdep

artm

ent

suppor

tof

wom

enE

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

Ste

p1

Dep

tsi

zeT

ime

2-T

ime

10

080

152

006

021

0

05D

R2

001

002

000

004

0

00T

otal

R2

001

002

000

004

0

00Ste

p2

Wom

enT

ime

2-T

ime

10

40

20

38

20

21

20

062

013

Whit

esT

ime

2-T

ime

12

007

019

008

20

172

001

Ten

ure

Tim

e2-

Tim

e1

20

072

013

012

016

019

DR

20

16

017

0

060

070

05T

otal

R2

017

0

19

006

011

0

05Ste

p3

Whit

ete

nure

dm

ales

Tim

e2-

Tim

e1

005

011

000

007

035

D

R2

000

001

000

000

004

T

otal

R2

017

0

20

006

011

009

Note

p

005

orgre

ater

p0

01or

gre

ater

Table IIIRegression

predicting valenceof climate for

diversity

HRM changestrategy

345

in tenured white males was entered in step 3 The results are that the thirdblock was not statistically signiregcant in the regressions showing that smallincreases in the group proportions of senior women and minorities did notincrease group consensus or valence regarding climate for change

4 DiscussionThis study explored links between actual change in organizational structureorganizational demography and work group interpretation of this change Ourresults showed that HR strategies that focus on structural change withoutworking to develop supportive group norms and positive climate areinadequate change strategies Current employer efforts to increase diversity asan isolated HR strategy may not necessarily lead to increased memberagreement that change is favorable Most change efforts like the diversitychange effort we studied are focused at the organizational or individualintervention levels and under-emphasize the work group level and theimportance of group tipping points Our study highlights a disconnect betweenorganizational level diversity practice which often assumes that hiring moreminorities and women will improve climate and what our research found at thegroup level a climate that was mixed and ambivalent at best and sometimesnegative

41 Group level measures of demographic change consensus and valence ondiversity climateThe regrst objective of this study was to develop new theory and measures ofgroup level change in diversity We provided new ways to measure and groupsocial construction of change These included consensus the degree to whichmembers held common perceptions of diversity climate and valence thedirection of those perceptions Our measures identify group and organizationallevel referents of HR strategies to increase demographic diversity Group levelreferents of diversity change were how well members perceived their workgroup mix and the perceived equality of resource allocation acrossdemographic groups Organizational referents were managementcommitment to change and whether diversity should be an organizationalstrategy assessed general support for change Our study highlights howimportant it is for research and practice to not only examine how change indemography relates to differences in individual attitudes but also to groupmember shared perceptions about the direction of the climate for changeUnderstanding linkages between HR changes (in this case group demographywhich raises unique psychological processes related to social identity) andpositive consensus is important because it is relevant to cooperation workexperiences interpersonal conmacrict stress and norms in groups striving tomanage change

JOCM163

346

42 Greater work group demographic shift does not necessarily result infavorable climateOur second objective was to assess whether work groups with higherdemographic shifts tend to have more positive climates that is higherconsensus and valence toward diversity change for both organizational andgroup level change referents Our results only partially supported this beliefthat greater structural demographic shift would positively relate to a morefavorable climate for diversity Although at the organizational level theuniversity had experienced signiregcant change in organizational demographyan increase in the overall representation of white women (36 percent) andminorities (41 percent) over time work group members in units with thegreatest change did not have higher consensus or more favorable valenceWe found more favorable results for organizational than group levelreferents of change Members may regnd it easier to agree with organizationalreferents and more difregcult to agree on group level referents due to theirgreater psychological proximity to group processes that affect how theypersonally experience conmacrict and wellbeing from change in their dailywork Work groups with more demographic change had higher valence butlower consensus on whether the organization should pursue diversity as astrategic goal When organizational change strategies to increase diversityare enacted at the work group level the group demographic shift may bediluted across the organization to only slightly increase the representation ofwomen and minorities within actual groups Due to this dilution effectgroup membersrsquo social construction of the HR strategies may lack consensusor positive valence regarding the organizational diversity change strategiesIf the change is not large enough to create a critical mass to tiprepresentation of minorities to at least 35 percent according to Kanter (1977)negative intergroup dynamics such as increased resource competition andtokenism may still occur

43 Diluting diversity gradually increasing the proportion of senior womenminorities does not necessarily improve group climateOur third objective was to investigate whether increasing the proportion ofsenior women and minorities at the department level positively relates togroup valence and consensus There was no relation between increasing theproportion of senior women and minorities and group valence andconsensus Due to diluted saliency of organizational HR change strategies toincrease diversity at the work group level slight demographic shifts toproportionally increase the representation of senior women and minoritiesdid not relate to higher group valence or consensus This increase did littleto improve group climate and may have even hurt the climate in theshort run

HRM changestrategy

347

5 Implications for future research and practice51 Incremental structural change not sufregcient to improve climateWe hope this study will encourage researchers and practitioners interested indiversity change strategies to augment individual and organizational levels ofintervention to focus on work group climates supporting change Our resultshighlight some disparities between the underlying assumptions of HRstrategies to manage organizational change (increasing structural diversitywill be positive for climate) and reality at the department level (incrementalincreases did not improve climate) Organizational strategists implementingdiversity recruitment initiatives typically assume that change strategies (in thiscase hiring diversity) would improve the group social climate due to a positivechange in member attitudes based on increased group diversity and positivesocial contact Yet our study does not support this view Although theorganization we studied had had signiregcant success in enacting the formal HRstrategy plusmn increasing the representation of women and minorities at theorganizational level plusmn our research shows that gradually shifting thedemography of work groups may not necessarily lead to increased consensusor valence about the social climate within work groups For example it did notlead to improved climate regarding the perceived fairness of resourceallocation or good mixing in social interaction In fact the incremental increasein the proportion of minority groups lowered consensus and agreement in theshort term as subgroups sorted out shifting and socially ambiguous powerrelationships within work units These results suggest the importance ofmeasuring change initiatives impacts at the group level and having group levelinterventions such as team-building and group-focused resource allocation tosupport organizational level diversity strategies

52 Focus change strategies on tipping points within targeted work unitsFuture research should consider how HR strategies to manage organizationaldemographic change are socially constructed and enacted at the work grouplevel It is especially important to identify what constitutes a group leveltipping point or critical mass across hierarchical levels and differentdemographic groups to enable positive consensus supporting changePractice might regnd it more fruitful to focus future HR strategies to increasediversity through hiring to alter tipping points within targeted work groupsThis will dramatically alter the saliency of strategic organizationaldemographic change and intervene to shape the climate of speciregc units Ourstudy suggests that organizations may need demographic earthquakes withinwork groups to effect group consensus and valence on the climate regardingthe HR change (ie managing diversity) Dilution of change at the group levelmay ameliorate intended positive effects unless certain group level tippingpoints such as at least 35 percent of the work group are effected in each unit by

JOCM163

348

the HR strategy After eight years of hiring new faculty in our sample womenhovered at about 15 percent in 89 percent of groups and minorities had amaximum representation of 25 percent The demographic shifts whileincreasing were not dramatic enough to create imbalances that markedlyincreased the salience and power of new demographic subgroups created bythe HR strategies Instead of being balanced many of our groups were stilllargely white male and remain tilted or skewed over time which Kanter (1977)notes perpetuates stereotypical negative dynamics

53 Avoiding backlash weaving tipping points across the hierarchyRecent work at MIT by Bailyn (2000) suggests that backlash and resistance toorganizational change may be occurring among senior white male faculty inresponse to HR diversity strategies Initially women and minorities now enterdepartments with relatively equal resources to white men as supported by theHR system design However by the time women and minorities reach theassociate and full professor levels subtle discrimination processes occur in howHR policies are applied such as lower access to resources promotions andpoor climate Despite the increase in upward mobility the representation ofwomen and minorities at the top is still relatively low Ely (1994) found that ifwomen are under-represented at higher organizational levels relations betweenwomen at lower levels might be of lower quality due to increased competitionand negative gender dynamics in the larger social system If HR changestrategies that alter the distribution of existing resources across organizationaland demographic groups are to be successful they must take into account hownot only tipping points in terms of overall departmental representation butmore importantly how tipping points are distributed in demographies acrossthe hierarchies of work groups or departments

HR diversity change strategies that bring in many new hires of differentbackgrounds may have negative ramiregcations for social functioning in groupsparticularly if new members are not supported by additional HR strategies thatallow them to enter work groups on an equal footing or if their work groups donot have leadership that mirror their demographic identities As Kanter (1983)found identity groups must be equal in their access to power resources ifcontact is to produce improvement in attitudes There must also be time for thenew members to be socialized and integrated into the culture Harrison et al(1998) found that the longer the length of time that members worked togetherthe greater the effects of deep level attitude similarity Organizations areincreasing surface (demographic) and deep level (attitudinal) diversitysimultaneously which is likely to have negative short-term ramiregcations forthe consensus and direction of climate We argue HR strategies must not onlyfocus on the surface level plusmn remacrecting structural and demographic attributes plusmnbut also simultaneously to deep level characteristics remacrecting values within

HRM changestrategy

349

the context of speciregc groups at all organizational levels In sum HR strategistsand change agents should focus on how HR change strategies relate to the jointalignment of group memberrsquos multiple characteristics (eg demographichierarchical status task) in work groups plusmn that is to examine group contextualinmacruences (Wharton 1992)

54 LimitationsOur study while making contributions to organizational research is notwithout limitations Though large in individual respondents by utilizing thegroup as the level of analysis our sample size includes only 74 groups Thiscould reduce the level of statistical power of our regndings Additionally becausethe sample was based on one organization it is important to replicate thisanalysis across multiple organizations to substantiate our results Althoughchanges in department heterogeneity across an eight-year period wereincluded this is not a longitudinal design Future research should employrepeated measures over a length of time across all variables measuringdemographic change in heterogeneity as well as consensus and valence Ourstudy points to some of the challenges in measuring ordfsuccess in increasingdiversityordm as we did not look at annual turnover or the effects of the glassceiling at the group level in this study which is a gap we hope futureresearchers will be motivated to study spurred by our research

Since little published work has been done on group consensus regarding theeffects of HR policies to increase diversity through demographic change weextend many social science literatures ranging from organizationaldevelopment to demography and HR Future research on diversity andorganizational change also should strive to be interdisciplinary as theliterature on HR strategy and change has been largely developed in the HRliterature separate from other social science disciplines reducing theoreticalintegration Without such integration and increased scholarly and practitionerattention to climate and consensus many organizations may fail to supportgroup membersrsquo explicit use of their cultural experiences to advanceorganizational learning from enacted diversity change strategies despite thegroup and intergroup social tensions that will inevitably occur (Thomas andEly 1996)

References

Aiken LA and West SG (1993) Multiple Regression Testing and Interpreting InteractionsSage Publications Newbury Park CA

Alderfer CP and Smith KK (1982) ordfStudying embedded intergroup relations inorganizationsordm Administrative Science Quarterly Vol 27 pp 35-65

Allport G (1954) The Nature of Prejudice Addison Wesley Cambridge MA

Bailyn L (2000) ordfGender and diversity in organizations preconference presentationordm paperpresented at the National Academy of Management meeting Toronto

JOCM163

350

Blalock HM (1967) Toward a Theory of Minority Group Relations Wiley New York NY

Bleise PD and Halverson RR (1998) ordfGroup consensus and psychological well-being a largeregeld studyordm Journal of Applied Social Psychology Vol 28 pp 563-80

Brown KG Kozlowski SWJ and Hattrup K (1996) Theory Issues and Recommendations inConceptualizing Agreement as a Construct in Organizational Research The Search forConsensus Regarding Consensus National Academy of Management Cincinnati OH

Cox T Jr (1993) Cultural Diversity in Organizations Berrett-Koehler San Francisco CA

Ely R (1994) ordfThe effects of organizational demographics and social identity on relationshipsamong professional womenordm Administrative Science Quarterly Vol 39 pp 203-38

Ely R (1995) ordfThe power in demography womenrsquos social constructions of gender identity atworkordm Academy of Management Journal Vol 38 pp 589-634

Festinger L (1954) ordfA theory of social comparison processesordm Human Relations Vol 7pp 117-40

Forsyth DR (1990) Group Dynamics BrooksCole Publishing Paciregc Grove CA

Harrison DA Price KH and Bell MP (1998) ordfBeyond relational demography time and theeffects of surface and deep-level diversity on work group cohesionordm Academy ofManagement Journal Vol 41 pp 96-107

James K Lovato C and Khoo G (1994) ordfSocial identity correlates of minority workersrsquo healthordmAcademy of Management Journal Vol 37 p 383+

Kanter R (1977) ordfSome effects of proportions on group life skewed sex ratios and reactions totoken womenordm American Journal of Sociology Vol 82 pp 965-90

Kanter R (1983) Men and Women of the Corporation Basic Books New York NY

Kossek EE and Lobel SA (1996) ordfIntroduction transforming human resource systems tomanage diversityordm in Managing Diversity Human Resource Strategies for Transforming

the Workplace Blackwell Oxford pp 1-19

Kossek EE and Zonia SC (1993) ordfAssessing diversity climate a regeld study of reactions toemployer efforts to promote diversityordm Journal of Organizational Behavior Vol 14pp 61-81

Larkey LK (1996) ordfToward a theory of communicative interactions in culturally diverseworkgroupsordm Academy of Management Review Vol 21 pp 463-91

Lau DC and Murnighan JK (1998) ordfDemographic diversity and faultlines the compositionaldynamics of organizational groupsordm Academy of Management Review Vol 29 pp 325-40

Lindell MK and Brandt CJ (2000) ordfClimate quality and climate consensus as mediators of therelationship between organizational antecedents and outcomesordm Journal of AppliedPsychology Vol 85 pp 331-48

McGrath JE Berdahl JL and Arrow H (1995) ordfTraits expectations culture and clout thedynamics of diversity in work groupsordm in Jackson SE and Ruderman MN (Eds)Diversity in Work Teams Research Paradigms for a Changing Workplace AmericanPsychological Association Washington DC

Payne RL and Pugh DS (1976) ordfOrganizational structure and climateordm in Dunnette MD(Ed) Handbook of Industrial and Organizational Psychology Rand McNally Chicago ILpp 1125-73

Reichers A and Schneider B (1990) ordfClimate and culture an evolution of constructsordm inSchneider B (Ed) Organizational Climate and Culture Jossey-Bass San Francisco CA

HRM changestrategy

351

Var

iable

1213

1415

1617

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

7P

ropor

tion

wit

hte

nure

Tim

e1

8P

ropor

tion

wit

hte

nure

Tim

e2

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

12C

han

ge

inte

nure

Tim

e1-

Tim

e2

plusmn13

C

han

ge

ingen

der

Tim

e1-

Tim

e2

001

plusmn14

C

han

ge

inw

hit

esT

ime

1-T

ime

20

162

007

plusmn15

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s0

150

110

14plusmn

16C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s2

013

021

006

20

07plusmn

17P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s0

150

130

060

38

034

plusmn

18E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s0

130

182

001

20

110

45

008

19E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s0

000

040

042

011

042

0

0020

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n0

000

050

150

34

20

32

20

0721

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n2

020

022

022

022

023

002

22P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

010

011

018

017

010

016

23E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n0

022

009

015

20

090

062

007

24E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

001

007

011

20

062

019

007

25D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

005

20

060

100

080

012

010

26C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

20

27

009

021

024

20

030

0227

P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n0

070

040

120

232

003

013

28E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

010

20

190

102

003

20

142

014

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n0

022

002

009

000

001

020

(con

tinued

)

Table I

JOCM163

340

Var

iable

1819

2021

2223

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

7P

ropor

tion

wit

hte

nure

Tim

e1

8P

ropor

tion

wit

hte

nure

Tim

e2

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

12C

han

ge

inte

nure

Tim

e1-

Tim

e2

13C

han

ge

ingen

der

Tim

e1-

Tim

e2

14C

han

ge

inw

hit

esT

ime

1-T

ime

215

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s16

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s17

P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s18

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

splusmn

19E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s0

57

plusmn20

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n2

035

2

027

plusmn

21C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n0

180

180

08plusmn

22P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

010

20

050

150

27

plusmn23

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n0

010

002

008

008

014

plusmn24

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

030

0

34

20

26

026

0

192

021

25D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

20

222

013

083

0

040

082

007

26C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

20

070

020

150

79

018

000

27P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n2

002

20

040

140

120

84

012

28E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

20

052

009

20

020

022

004

086

29

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n0

140

31

20

210

100

042

008

(con

tinued

)

Table I

HRM changestrategy

341

Table I

Var

iable

2425

2627

2829

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

7P

ropor

tion

wit

hte

nure

Tim

e1

8P

ropor

tion

wit

hte

nure

Tim

e2

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

12C

han

ge

inte

nure

Tim

e1-

Tim

e2

13C

han

ge

ingen

der

Tim

e1-

Tim

e2

14C

han

ge

inw

hit

esT

ime

1-T

ime

215

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s16

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s17

P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s18

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s19

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s20

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n21

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n22

P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

23E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n24

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

plusmn25

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

20

20plusmn

26C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

025

002

plusmn27

P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n0

080

010

19plusmn

28E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

20

32

000

20

010

05plusmn

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n0

58

20

240

180

132

003

plusmn

Note

p

001

orgre

ater

JOCM163

342

support of resources for women and men and whites and non-whites asthese measures hovered at plusmn024 and 2 025 or close to 0 which would betotal consensus There was the least consensus that minorities and majoritiesmixed well (plusmn125) with slightly more consensus that diversity should bepursued as organizational (plusmn10) The most positive valence was theperception of equality of department support for minorities (393) and thatdiversity should be pursued as a goal (39)

34 Regression results on demographic shift and climateTo address our second research objective which was to examine whetherwork groups with higher demographic shifts had more positive groupclimates for each of the regve dependent variables separate hierarchicalregressions were run to examine the relation between group demographicchange and group consensus (see Table II) or valence (see Table III) In theregrst step of all regressions proportional change in size was entered as acontrol For the consensus regressions the current mean of each dependentvariable was also entered as a control In step 2 the group proportionalchange in women and the group proportional change in minorities wereentered (In early analyses for the second step of each regression interactionterms change in gender or race by tenure were included following Aikenand Westrsquos (1993) procedures No signiregcant interactions were found) Usingstandardized beta weights as seen in step 2 of Tables II and III ourhypothesis that work groups with higher demographic shifts may havehigher consensus and valence toward diversity change received partialsupport Group referents of diversity climate were unchanged bydemographic shifts remacrecting possible dilution of HR strategiesOrganizational referents of valence were effected but consensus wasreduced in terms of whether the organization should pursue diversity as anorganizational goal Departments with higher shifts in the proportion ofwomen had lower consensus on whether the organization should pursuediversity as an organizational goal Both regressions with organizationalreferents as dependent variables were signiregcant Groups with higherincreases in the proportion of women over time tended to have morefavorable valence regarding the desirability of promoting diversity as anorganizational goal but were less sure of managementrsquos commitment asthey had more experience with the reality of implementing theorganizational change

35 Regression results on positive senior women and minority effect on climateOur third research objective was to examine whether increasing the proportionof senior women and minorities at the department level positively relates togroup valence and consensus In our data analysis the proportional change inoverall tenured faculty was entered in step 2 and then the proportional change

HRM changestrategy

343

Org

aniz

atio

nD

epar

tmen

tD

iver

sity

asor

gan

izat

ional

goa

l

Com

mit

men

tof

man

agem

ent

todiv

ersi

ty

Per

ceiv

edw

ork-g

roup

mix

Equal

ity

ofdep

artm

ent

suppor

tof

wom

enE

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

Ste

p1

Dep

tsi

zeT

ime

2-T

ime

10

020

080

180

000

06M

ean

ofD

V0

35

022

017

002

029

D

R2

012

0

060

060

000

09T

otal

R2

012

0

060

060

000

09Ste

p2

Wom

enT

ime

2-T

ime

12

035

2

017

20

182

007

20

02W

hit

esT

ime

2-T

ime

12

002

20

052

003

004

021

Ten

ure

Tim

e2-

Tim

e1

009

003

022

010

000

DR

20

11

003

008

001

004

Tot

alR

20

23

009

014

001

013

Ste

p3

Whit

ete

nure

dm

ales

Tim

e2-

Tim

e1

20

092

012

005

20

302

031

DR

20

010

010

000

040

05T

otal

R2

024

0

100

140

050

18

Note

p

005

orgre

ater

p0

01or

gre

ater

Table IIRegressionpredictingconsensus onclimate for diversity

JOCM163

344

Org

aniz

atio

nD

epar

tmen

tD

iver

sity

asor

gan

izat

ional

goa

l

Com

mit

men

tof

man

agem

ent

todiv

ersi

ty

Per

ceiv

edw

ork-g

roup

mix

Equal

ity

ofdep

artm

ent

suppor

tof

wom

enE

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

Ste

p1

Dep

tsi

zeT

ime

2-T

ime

10

080

152

006

021

0

05D

R2

001

002

000

004

0

00T

otal

R2

001

002

000

004

0

00Ste

p2

Wom

enT

ime

2-T

ime

10

40

20

38

20

21

20

062

013

Whit

esT

ime

2-T

ime

12

007

019

008

20

172

001

Ten

ure

Tim

e2-

Tim

e1

20

072

013

012

016

019

DR

20

16

017

0

060

070

05T

otal

R2

017

0

19

006

011

0

05Ste

p3

Whit

ete

nure

dm

ales

Tim

e2-

Tim

e1

005

011

000

007

035

D

R2

000

001

000

000

004

T

otal

R2

017

0

20

006

011

009

Note

p

005

orgre

ater

p0

01or

gre

ater

Table IIIRegression

predicting valenceof climate for

diversity

HRM changestrategy

345

in tenured white males was entered in step 3 The results are that the thirdblock was not statistically signiregcant in the regressions showing that smallincreases in the group proportions of senior women and minorities did notincrease group consensus or valence regarding climate for change

4 DiscussionThis study explored links between actual change in organizational structureorganizational demography and work group interpretation of this change Ourresults showed that HR strategies that focus on structural change withoutworking to develop supportive group norms and positive climate areinadequate change strategies Current employer efforts to increase diversity asan isolated HR strategy may not necessarily lead to increased memberagreement that change is favorable Most change efforts like the diversitychange effort we studied are focused at the organizational or individualintervention levels and under-emphasize the work group level and theimportance of group tipping points Our study highlights a disconnect betweenorganizational level diversity practice which often assumes that hiring moreminorities and women will improve climate and what our research found at thegroup level a climate that was mixed and ambivalent at best and sometimesnegative

41 Group level measures of demographic change consensus and valence ondiversity climateThe regrst objective of this study was to develop new theory and measures ofgroup level change in diversity We provided new ways to measure and groupsocial construction of change These included consensus the degree to whichmembers held common perceptions of diversity climate and valence thedirection of those perceptions Our measures identify group and organizationallevel referents of HR strategies to increase demographic diversity Group levelreferents of diversity change were how well members perceived their workgroup mix and the perceived equality of resource allocation acrossdemographic groups Organizational referents were managementcommitment to change and whether diversity should be an organizationalstrategy assessed general support for change Our study highlights howimportant it is for research and practice to not only examine how change indemography relates to differences in individual attitudes but also to groupmember shared perceptions about the direction of the climate for changeUnderstanding linkages between HR changes (in this case group demographywhich raises unique psychological processes related to social identity) andpositive consensus is important because it is relevant to cooperation workexperiences interpersonal conmacrict stress and norms in groups striving tomanage change

JOCM163

346

42 Greater work group demographic shift does not necessarily result infavorable climateOur second objective was to assess whether work groups with higherdemographic shifts tend to have more positive climates that is higherconsensus and valence toward diversity change for both organizational andgroup level change referents Our results only partially supported this beliefthat greater structural demographic shift would positively relate to a morefavorable climate for diversity Although at the organizational level theuniversity had experienced signiregcant change in organizational demographyan increase in the overall representation of white women (36 percent) andminorities (41 percent) over time work group members in units with thegreatest change did not have higher consensus or more favorable valenceWe found more favorable results for organizational than group levelreferents of change Members may regnd it easier to agree with organizationalreferents and more difregcult to agree on group level referents due to theirgreater psychological proximity to group processes that affect how theypersonally experience conmacrict and wellbeing from change in their dailywork Work groups with more demographic change had higher valence butlower consensus on whether the organization should pursue diversity as astrategic goal When organizational change strategies to increase diversityare enacted at the work group level the group demographic shift may bediluted across the organization to only slightly increase the representation ofwomen and minorities within actual groups Due to this dilution effectgroup membersrsquo social construction of the HR strategies may lack consensusor positive valence regarding the organizational diversity change strategiesIf the change is not large enough to create a critical mass to tiprepresentation of minorities to at least 35 percent according to Kanter (1977)negative intergroup dynamics such as increased resource competition andtokenism may still occur

43 Diluting diversity gradually increasing the proportion of senior womenminorities does not necessarily improve group climateOur third objective was to investigate whether increasing the proportion ofsenior women and minorities at the department level positively relates togroup valence and consensus There was no relation between increasing theproportion of senior women and minorities and group valence andconsensus Due to diluted saliency of organizational HR change strategies toincrease diversity at the work group level slight demographic shifts toproportionally increase the representation of senior women and minoritiesdid not relate to higher group valence or consensus This increase did littleto improve group climate and may have even hurt the climate in theshort run

HRM changestrategy

347

5 Implications for future research and practice51 Incremental structural change not sufregcient to improve climateWe hope this study will encourage researchers and practitioners interested indiversity change strategies to augment individual and organizational levels ofintervention to focus on work group climates supporting change Our resultshighlight some disparities between the underlying assumptions of HRstrategies to manage organizational change (increasing structural diversitywill be positive for climate) and reality at the department level (incrementalincreases did not improve climate) Organizational strategists implementingdiversity recruitment initiatives typically assume that change strategies (in thiscase hiring diversity) would improve the group social climate due to a positivechange in member attitudes based on increased group diversity and positivesocial contact Yet our study does not support this view Although theorganization we studied had had signiregcant success in enacting the formal HRstrategy plusmn increasing the representation of women and minorities at theorganizational level plusmn our research shows that gradually shifting thedemography of work groups may not necessarily lead to increased consensusor valence about the social climate within work groups For example it did notlead to improved climate regarding the perceived fairness of resourceallocation or good mixing in social interaction In fact the incremental increasein the proportion of minority groups lowered consensus and agreement in theshort term as subgroups sorted out shifting and socially ambiguous powerrelationships within work units These results suggest the importance ofmeasuring change initiatives impacts at the group level and having group levelinterventions such as team-building and group-focused resource allocation tosupport organizational level diversity strategies

52 Focus change strategies on tipping points within targeted work unitsFuture research should consider how HR strategies to manage organizationaldemographic change are socially constructed and enacted at the work grouplevel It is especially important to identify what constitutes a group leveltipping point or critical mass across hierarchical levels and differentdemographic groups to enable positive consensus supporting changePractice might regnd it more fruitful to focus future HR strategies to increasediversity through hiring to alter tipping points within targeted work groupsThis will dramatically alter the saliency of strategic organizationaldemographic change and intervene to shape the climate of speciregc units Ourstudy suggests that organizations may need demographic earthquakes withinwork groups to effect group consensus and valence on the climate regardingthe HR change (ie managing diversity) Dilution of change at the group levelmay ameliorate intended positive effects unless certain group level tippingpoints such as at least 35 percent of the work group are effected in each unit by

JOCM163

348

the HR strategy After eight years of hiring new faculty in our sample womenhovered at about 15 percent in 89 percent of groups and minorities had amaximum representation of 25 percent The demographic shifts whileincreasing were not dramatic enough to create imbalances that markedlyincreased the salience and power of new demographic subgroups created bythe HR strategies Instead of being balanced many of our groups were stilllargely white male and remain tilted or skewed over time which Kanter (1977)notes perpetuates stereotypical negative dynamics

53 Avoiding backlash weaving tipping points across the hierarchyRecent work at MIT by Bailyn (2000) suggests that backlash and resistance toorganizational change may be occurring among senior white male faculty inresponse to HR diversity strategies Initially women and minorities now enterdepartments with relatively equal resources to white men as supported by theHR system design However by the time women and minorities reach theassociate and full professor levels subtle discrimination processes occur in howHR policies are applied such as lower access to resources promotions andpoor climate Despite the increase in upward mobility the representation ofwomen and minorities at the top is still relatively low Ely (1994) found that ifwomen are under-represented at higher organizational levels relations betweenwomen at lower levels might be of lower quality due to increased competitionand negative gender dynamics in the larger social system If HR changestrategies that alter the distribution of existing resources across organizationaland demographic groups are to be successful they must take into account hownot only tipping points in terms of overall departmental representation butmore importantly how tipping points are distributed in demographies acrossthe hierarchies of work groups or departments

HR diversity change strategies that bring in many new hires of differentbackgrounds may have negative ramiregcations for social functioning in groupsparticularly if new members are not supported by additional HR strategies thatallow them to enter work groups on an equal footing or if their work groups donot have leadership that mirror their demographic identities As Kanter (1983)found identity groups must be equal in their access to power resources ifcontact is to produce improvement in attitudes There must also be time for thenew members to be socialized and integrated into the culture Harrison et al(1998) found that the longer the length of time that members worked togetherthe greater the effects of deep level attitude similarity Organizations areincreasing surface (demographic) and deep level (attitudinal) diversitysimultaneously which is likely to have negative short-term ramiregcations forthe consensus and direction of climate We argue HR strategies must not onlyfocus on the surface level plusmn remacrecting structural and demographic attributes plusmnbut also simultaneously to deep level characteristics remacrecting values within

HRM changestrategy

349

the context of speciregc groups at all organizational levels In sum HR strategistsand change agents should focus on how HR change strategies relate to the jointalignment of group memberrsquos multiple characteristics (eg demographichierarchical status task) in work groups plusmn that is to examine group contextualinmacruences (Wharton 1992)

54 LimitationsOur study while making contributions to organizational research is notwithout limitations Though large in individual respondents by utilizing thegroup as the level of analysis our sample size includes only 74 groups Thiscould reduce the level of statistical power of our regndings Additionally becausethe sample was based on one organization it is important to replicate thisanalysis across multiple organizations to substantiate our results Althoughchanges in department heterogeneity across an eight-year period wereincluded this is not a longitudinal design Future research should employrepeated measures over a length of time across all variables measuringdemographic change in heterogeneity as well as consensus and valence Ourstudy points to some of the challenges in measuring ordfsuccess in increasingdiversityordm as we did not look at annual turnover or the effects of the glassceiling at the group level in this study which is a gap we hope futureresearchers will be motivated to study spurred by our research

Since little published work has been done on group consensus regarding theeffects of HR policies to increase diversity through demographic change weextend many social science literatures ranging from organizationaldevelopment to demography and HR Future research on diversity andorganizational change also should strive to be interdisciplinary as theliterature on HR strategy and change has been largely developed in the HRliterature separate from other social science disciplines reducing theoreticalintegration Without such integration and increased scholarly and practitionerattention to climate and consensus many organizations may fail to supportgroup membersrsquo explicit use of their cultural experiences to advanceorganizational learning from enacted diversity change strategies despite thegroup and intergroup social tensions that will inevitably occur (Thomas andEly 1996)

References

Aiken LA and West SG (1993) Multiple Regression Testing and Interpreting InteractionsSage Publications Newbury Park CA

Alderfer CP and Smith KK (1982) ordfStudying embedded intergroup relations inorganizationsordm Administrative Science Quarterly Vol 27 pp 35-65

Allport G (1954) The Nature of Prejudice Addison Wesley Cambridge MA

Bailyn L (2000) ordfGender and diversity in organizations preconference presentationordm paperpresented at the National Academy of Management meeting Toronto

JOCM163

350

Blalock HM (1967) Toward a Theory of Minority Group Relations Wiley New York NY

Bleise PD and Halverson RR (1998) ordfGroup consensus and psychological well-being a largeregeld studyordm Journal of Applied Social Psychology Vol 28 pp 563-80

Brown KG Kozlowski SWJ and Hattrup K (1996) Theory Issues and Recommendations inConceptualizing Agreement as a Construct in Organizational Research The Search forConsensus Regarding Consensus National Academy of Management Cincinnati OH

Cox T Jr (1993) Cultural Diversity in Organizations Berrett-Koehler San Francisco CA

Ely R (1994) ordfThe effects of organizational demographics and social identity on relationshipsamong professional womenordm Administrative Science Quarterly Vol 39 pp 203-38

Ely R (1995) ordfThe power in demography womenrsquos social constructions of gender identity atworkordm Academy of Management Journal Vol 38 pp 589-634

Festinger L (1954) ordfA theory of social comparison processesordm Human Relations Vol 7pp 117-40

Forsyth DR (1990) Group Dynamics BrooksCole Publishing Paciregc Grove CA

Harrison DA Price KH and Bell MP (1998) ordfBeyond relational demography time and theeffects of surface and deep-level diversity on work group cohesionordm Academy ofManagement Journal Vol 41 pp 96-107

James K Lovato C and Khoo G (1994) ordfSocial identity correlates of minority workersrsquo healthordmAcademy of Management Journal Vol 37 p 383+

Kanter R (1977) ordfSome effects of proportions on group life skewed sex ratios and reactions totoken womenordm American Journal of Sociology Vol 82 pp 965-90

Kanter R (1983) Men and Women of the Corporation Basic Books New York NY

Kossek EE and Lobel SA (1996) ordfIntroduction transforming human resource systems tomanage diversityordm in Managing Diversity Human Resource Strategies for Transforming

the Workplace Blackwell Oxford pp 1-19

Kossek EE and Zonia SC (1993) ordfAssessing diversity climate a regeld study of reactions toemployer efforts to promote diversityordm Journal of Organizational Behavior Vol 14pp 61-81

Larkey LK (1996) ordfToward a theory of communicative interactions in culturally diverseworkgroupsordm Academy of Management Review Vol 21 pp 463-91

Lau DC and Murnighan JK (1998) ordfDemographic diversity and faultlines the compositionaldynamics of organizational groupsordm Academy of Management Review Vol 29 pp 325-40

Lindell MK and Brandt CJ (2000) ordfClimate quality and climate consensus as mediators of therelationship between organizational antecedents and outcomesordm Journal of AppliedPsychology Vol 85 pp 331-48

McGrath JE Berdahl JL and Arrow H (1995) ordfTraits expectations culture and clout thedynamics of diversity in work groupsordm in Jackson SE and Ruderman MN (Eds)Diversity in Work Teams Research Paradigms for a Changing Workplace AmericanPsychological Association Washington DC

Payne RL and Pugh DS (1976) ordfOrganizational structure and climateordm in Dunnette MD(Ed) Handbook of Industrial and Organizational Psychology Rand McNally Chicago ILpp 1125-73

Reichers A and Schneider B (1990) ordfClimate and culture an evolution of constructsordm inSchneider B (Ed) Organizational Climate and Culture Jossey-Bass San Francisco CA

HRM changestrategy

351

Var

iable

1819

2021

2223

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

7P

ropor

tion

wit

hte

nure

Tim

e1

8P

ropor

tion

wit

hte

nure

Tim

e2

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

12C

han

ge

inte

nure

Tim

e1-

Tim

e2

13C

han

ge

ingen

der

Tim

e1-

Tim

e2

14C

han

ge

inw

hit

esT

ime

1-T

ime

215

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s16

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s17

P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s18

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

splusmn

19E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s0

57

plusmn20

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n2

035

2

027

plusmn

21C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n0

180

180

08plusmn

22P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

010

20

050

150

27

plusmn23

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n0

010

002

008

008

014

plusmn24

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

030

0

34

20

26

026

0

192

021

25D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

20

222

013

083

0

040

082

007

26C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

20

070

020

150

79

018

000

27P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n2

002

20

040

140

120

84

012

28E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

20

052

009

20

020

022

004

086

29

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n0

140

31

20

210

100

042

008

(con

tinued

)

Table I

HRM changestrategy

341

Table I

Var

iable

2425

2627

2829

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

7P

ropor

tion

wit

hte

nure

Tim

e1

8P

ropor

tion

wit

hte

nure

Tim

e2

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

12C

han

ge

inte

nure

Tim

e1-

Tim

e2

13C

han

ge

ingen

der

Tim

e1-

Tim

e2

14C

han

ge

inw

hit

esT

ime

1-T

ime

215

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s16

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s17

P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s18

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s19

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s20

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n21

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n22

P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

23E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n24

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

plusmn25

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

20

20plusmn

26C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

025

002

plusmn27

P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n0

080

010

19plusmn

28E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

20

32

000

20

010

05plusmn

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n0

58

20

240

180

132

003

plusmn

Note

p

001

orgre

ater

JOCM163

342

support of resources for women and men and whites and non-whites asthese measures hovered at plusmn024 and 2 025 or close to 0 which would betotal consensus There was the least consensus that minorities and majoritiesmixed well (plusmn125) with slightly more consensus that diversity should bepursued as organizational (plusmn10) The most positive valence was theperception of equality of department support for minorities (393) and thatdiversity should be pursued as a goal (39)

34 Regression results on demographic shift and climateTo address our second research objective which was to examine whetherwork groups with higher demographic shifts had more positive groupclimates for each of the regve dependent variables separate hierarchicalregressions were run to examine the relation between group demographicchange and group consensus (see Table II) or valence (see Table III) In theregrst step of all regressions proportional change in size was entered as acontrol For the consensus regressions the current mean of each dependentvariable was also entered as a control In step 2 the group proportionalchange in women and the group proportional change in minorities wereentered (In early analyses for the second step of each regression interactionterms change in gender or race by tenure were included following Aikenand Westrsquos (1993) procedures No signiregcant interactions were found) Usingstandardized beta weights as seen in step 2 of Tables II and III ourhypothesis that work groups with higher demographic shifts may havehigher consensus and valence toward diversity change received partialsupport Group referents of diversity climate were unchanged bydemographic shifts remacrecting possible dilution of HR strategiesOrganizational referents of valence were effected but consensus wasreduced in terms of whether the organization should pursue diversity as anorganizational goal Departments with higher shifts in the proportion ofwomen had lower consensus on whether the organization should pursuediversity as an organizational goal Both regressions with organizationalreferents as dependent variables were signiregcant Groups with higherincreases in the proportion of women over time tended to have morefavorable valence regarding the desirability of promoting diversity as anorganizational goal but were less sure of managementrsquos commitment asthey had more experience with the reality of implementing theorganizational change

35 Regression results on positive senior women and minority effect on climateOur third research objective was to examine whether increasing the proportionof senior women and minorities at the department level positively relates togroup valence and consensus In our data analysis the proportional change inoverall tenured faculty was entered in step 2 and then the proportional change

HRM changestrategy

343

Org

aniz

atio

nD

epar

tmen

tD

iver

sity

asor

gan

izat

ional

goa

l

Com

mit

men

tof

man

agem

ent

todiv

ersi

ty

Per

ceiv

edw

ork-g

roup

mix

Equal

ity

ofdep

artm

ent

suppor

tof

wom

enE

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

Ste

p1

Dep

tsi

zeT

ime

2-T

ime

10

020

080

180

000

06M

ean

ofD

V0

35

022

017

002

029

D

R2

012

0

060

060

000

09T

otal

R2

012

0

060

060

000

09Ste

p2

Wom

enT

ime

2-T

ime

12

035

2

017

20

182

007

20

02W

hit

esT

ime

2-T

ime

12

002

20

052

003

004

021

Ten

ure

Tim

e2-

Tim

e1

009

003

022

010

000

DR

20

11

003

008

001

004

Tot

alR

20

23

009

014

001

013

Ste

p3

Whit

ete

nure

dm

ales

Tim

e2-

Tim

e1

20

092

012

005

20

302

031

DR

20

010

010

000

040

05T

otal

R2

024

0

100

140

050

18

Note

p

005

orgre

ater

p0

01or

gre

ater

Table IIRegressionpredictingconsensus onclimate for diversity

JOCM163

344

Org

aniz

atio

nD

epar

tmen

tD

iver

sity

asor

gan

izat

ional

goa

l

Com

mit

men

tof

man

agem

ent

todiv

ersi

ty

Per

ceiv

edw

ork-g

roup

mix

Equal

ity

ofdep

artm

ent

suppor

tof

wom

enE

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

Ste

p1

Dep

tsi

zeT

ime

2-T

ime

10

080

152

006

021

0

05D

R2

001

002

000

004

0

00T

otal

R2

001

002

000

004

0

00Ste

p2

Wom

enT

ime

2-T

ime

10

40

20

38

20

21

20

062

013

Whit

esT

ime

2-T

ime

12

007

019

008

20

172

001

Ten

ure

Tim

e2-

Tim

e1

20

072

013

012

016

019

DR

20

16

017

0

060

070

05T

otal

R2

017

0

19

006

011

0

05Ste

p3

Whit

ete

nure

dm

ales

Tim

e2-

Tim

e1

005

011

000

007

035

D

R2

000

001

000

000

004

T

otal

R2

017

0

20

006

011

009

Note

p

005

orgre

ater

p0

01or

gre

ater

Table IIIRegression

predicting valenceof climate for

diversity

HRM changestrategy

345

in tenured white males was entered in step 3 The results are that the thirdblock was not statistically signiregcant in the regressions showing that smallincreases in the group proportions of senior women and minorities did notincrease group consensus or valence regarding climate for change

4 DiscussionThis study explored links between actual change in organizational structureorganizational demography and work group interpretation of this change Ourresults showed that HR strategies that focus on structural change withoutworking to develop supportive group norms and positive climate areinadequate change strategies Current employer efforts to increase diversity asan isolated HR strategy may not necessarily lead to increased memberagreement that change is favorable Most change efforts like the diversitychange effort we studied are focused at the organizational or individualintervention levels and under-emphasize the work group level and theimportance of group tipping points Our study highlights a disconnect betweenorganizational level diversity practice which often assumes that hiring moreminorities and women will improve climate and what our research found at thegroup level a climate that was mixed and ambivalent at best and sometimesnegative

41 Group level measures of demographic change consensus and valence ondiversity climateThe regrst objective of this study was to develop new theory and measures ofgroup level change in diversity We provided new ways to measure and groupsocial construction of change These included consensus the degree to whichmembers held common perceptions of diversity climate and valence thedirection of those perceptions Our measures identify group and organizationallevel referents of HR strategies to increase demographic diversity Group levelreferents of diversity change were how well members perceived their workgroup mix and the perceived equality of resource allocation acrossdemographic groups Organizational referents were managementcommitment to change and whether diversity should be an organizationalstrategy assessed general support for change Our study highlights howimportant it is for research and practice to not only examine how change indemography relates to differences in individual attitudes but also to groupmember shared perceptions about the direction of the climate for changeUnderstanding linkages between HR changes (in this case group demographywhich raises unique psychological processes related to social identity) andpositive consensus is important because it is relevant to cooperation workexperiences interpersonal conmacrict stress and norms in groups striving tomanage change

JOCM163

346

42 Greater work group demographic shift does not necessarily result infavorable climateOur second objective was to assess whether work groups with higherdemographic shifts tend to have more positive climates that is higherconsensus and valence toward diversity change for both organizational andgroup level change referents Our results only partially supported this beliefthat greater structural demographic shift would positively relate to a morefavorable climate for diversity Although at the organizational level theuniversity had experienced signiregcant change in organizational demographyan increase in the overall representation of white women (36 percent) andminorities (41 percent) over time work group members in units with thegreatest change did not have higher consensus or more favorable valenceWe found more favorable results for organizational than group levelreferents of change Members may regnd it easier to agree with organizationalreferents and more difregcult to agree on group level referents due to theirgreater psychological proximity to group processes that affect how theypersonally experience conmacrict and wellbeing from change in their dailywork Work groups with more demographic change had higher valence butlower consensus on whether the organization should pursue diversity as astrategic goal When organizational change strategies to increase diversityare enacted at the work group level the group demographic shift may bediluted across the organization to only slightly increase the representation ofwomen and minorities within actual groups Due to this dilution effectgroup membersrsquo social construction of the HR strategies may lack consensusor positive valence regarding the organizational diversity change strategiesIf the change is not large enough to create a critical mass to tiprepresentation of minorities to at least 35 percent according to Kanter (1977)negative intergroup dynamics such as increased resource competition andtokenism may still occur

43 Diluting diversity gradually increasing the proportion of senior womenminorities does not necessarily improve group climateOur third objective was to investigate whether increasing the proportion ofsenior women and minorities at the department level positively relates togroup valence and consensus There was no relation between increasing theproportion of senior women and minorities and group valence andconsensus Due to diluted saliency of organizational HR change strategies toincrease diversity at the work group level slight demographic shifts toproportionally increase the representation of senior women and minoritiesdid not relate to higher group valence or consensus This increase did littleto improve group climate and may have even hurt the climate in theshort run

HRM changestrategy

347

5 Implications for future research and practice51 Incremental structural change not sufregcient to improve climateWe hope this study will encourage researchers and practitioners interested indiversity change strategies to augment individual and organizational levels ofintervention to focus on work group climates supporting change Our resultshighlight some disparities between the underlying assumptions of HRstrategies to manage organizational change (increasing structural diversitywill be positive for climate) and reality at the department level (incrementalincreases did not improve climate) Organizational strategists implementingdiversity recruitment initiatives typically assume that change strategies (in thiscase hiring diversity) would improve the group social climate due to a positivechange in member attitudes based on increased group diversity and positivesocial contact Yet our study does not support this view Although theorganization we studied had had signiregcant success in enacting the formal HRstrategy plusmn increasing the representation of women and minorities at theorganizational level plusmn our research shows that gradually shifting thedemography of work groups may not necessarily lead to increased consensusor valence about the social climate within work groups For example it did notlead to improved climate regarding the perceived fairness of resourceallocation or good mixing in social interaction In fact the incremental increasein the proportion of minority groups lowered consensus and agreement in theshort term as subgroups sorted out shifting and socially ambiguous powerrelationships within work units These results suggest the importance ofmeasuring change initiatives impacts at the group level and having group levelinterventions such as team-building and group-focused resource allocation tosupport organizational level diversity strategies

52 Focus change strategies on tipping points within targeted work unitsFuture research should consider how HR strategies to manage organizationaldemographic change are socially constructed and enacted at the work grouplevel It is especially important to identify what constitutes a group leveltipping point or critical mass across hierarchical levels and differentdemographic groups to enable positive consensus supporting changePractice might regnd it more fruitful to focus future HR strategies to increasediversity through hiring to alter tipping points within targeted work groupsThis will dramatically alter the saliency of strategic organizationaldemographic change and intervene to shape the climate of speciregc units Ourstudy suggests that organizations may need demographic earthquakes withinwork groups to effect group consensus and valence on the climate regardingthe HR change (ie managing diversity) Dilution of change at the group levelmay ameliorate intended positive effects unless certain group level tippingpoints such as at least 35 percent of the work group are effected in each unit by

JOCM163

348

the HR strategy After eight years of hiring new faculty in our sample womenhovered at about 15 percent in 89 percent of groups and minorities had amaximum representation of 25 percent The demographic shifts whileincreasing were not dramatic enough to create imbalances that markedlyincreased the salience and power of new demographic subgroups created bythe HR strategies Instead of being balanced many of our groups were stilllargely white male and remain tilted or skewed over time which Kanter (1977)notes perpetuates stereotypical negative dynamics

53 Avoiding backlash weaving tipping points across the hierarchyRecent work at MIT by Bailyn (2000) suggests that backlash and resistance toorganizational change may be occurring among senior white male faculty inresponse to HR diversity strategies Initially women and minorities now enterdepartments with relatively equal resources to white men as supported by theHR system design However by the time women and minorities reach theassociate and full professor levels subtle discrimination processes occur in howHR policies are applied such as lower access to resources promotions andpoor climate Despite the increase in upward mobility the representation ofwomen and minorities at the top is still relatively low Ely (1994) found that ifwomen are under-represented at higher organizational levels relations betweenwomen at lower levels might be of lower quality due to increased competitionand negative gender dynamics in the larger social system If HR changestrategies that alter the distribution of existing resources across organizationaland demographic groups are to be successful they must take into account hownot only tipping points in terms of overall departmental representation butmore importantly how tipping points are distributed in demographies acrossthe hierarchies of work groups or departments

HR diversity change strategies that bring in many new hires of differentbackgrounds may have negative ramiregcations for social functioning in groupsparticularly if new members are not supported by additional HR strategies thatallow them to enter work groups on an equal footing or if their work groups donot have leadership that mirror their demographic identities As Kanter (1983)found identity groups must be equal in their access to power resources ifcontact is to produce improvement in attitudes There must also be time for thenew members to be socialized and integrated into the culture Harrison et al(1998) found that the longer the length of time that members worked togetherthe greater the effects of deep level attitude similarity Organizations areincreasing surface (demographic) and deep level (attitudinal) diversitysimultaneously which is likely to have negative short-term ramiregcations forthe consensus and direction of climate We argue HR strategies must not onlyfocus on the surface level plusmn remacrecting structural and demographic attributes plusmnbut also simultaneously to deep level characteristics remacrecting values within

HRM changestrategy

349

the context of speciregc groups at all organizational levels In sum HR strategistsand change agents should focus on how HR change strategies relate to the jointalignment of group memberrsquos multiple characteristics (eg demographichierarchical status task) in work groups plusmn that is to examine group contextualinmacruences (Wharton 1992)

54 LimitationsOur study while making contributions to organizational research is notwithout limitations Though large in individual respondents by utilizing thegroup as the level of analysis our sample size includes only 74 groups Thiscould reduce the level of statistical power of our regndings Additionally becausethe sample was based on one organization it is important to replicate thisanalysis across multiple organizations to substantiate our results Althoughchanges in department heterogeneity across an eight-year period wereincluded this is not a longitudinal design Future research should employrepeated measures over a length of time across all variables measuringdemographic change in heterogeneity as well as consensus and valence Ourstudy points to some of the challenges in measuring ordfsuccess in increasingdiversityordm as we did not look at annual turnover or the effects of the glassceiling at the group level in this study which is a gap we hope futureresearchers will be motivated to study spurred by our research

Since little published work has been done on group consensus regarding theeffects of HR policies to increase diversity through demographic change weextend many social science literatures ranging from organizationaldevelopment to demography and HR Future research on diversity andorganizational change also should strive to be interdisciplinary as theliterature on HR strategy and change has been largely developed in the HRliterature separate from other social science disciplines reducing theoreticalintegration Without such integration and increased scholarly and practitionerattention to climate and consensus many organizations may fail to supportgroup membersrsquo explicit use of their cultural experiences to advanceorganizational learning from enacted diversity change strategies despite thegroup and intergroup social tensions that will inevitably occur (Thomas andEly 1996)

References

Aiken LA and West SG (1993) Multiple Regression Testing and Interpreting InteractionsSage Publications Newbury Park CA

Alderfer CP and Smith KK (1982) ordfStudying embedded intergroup relations inorganizationsordm Administrative Science Quarterly Vol 27 pp 35-65

Allport G (1954) The Nature of Prejudice Addison Wesley Cambridge MA

Bailyn L (2000) ordfGender and diversity in organizations preconference presentationordm paperpresented at the National Academy of Management meeting Toronto

JOCM163

350

Blalock HM (1967) Toward a Theory of Minority Group Relations Wiley New York NY

Bleise PD and Halverson RR (1998) ordfGroup consensus and psychological well-being a largeregeld studyordm Journal of Applied Social Psychology Vol 28 pp 563-80

Brown KG Kozlowski SWJ and Hattrup K (1996) Theory Issues and Recommendations inConceptualizing Agreement as a Construct in Organizational Research The Search forConsensus Regarding Consensus National Academy of Management Cincinnati OH

Cox T Jr (1993) Cultural Diversity in Organizations Berrett-Koehler San Francisco CA

Ely R (1994) ordfThe effects of organizational demographics and social identity on relationshipsamong professional womenordm Administrative Science Quarterly Vol 39 pp 203-38

Ely R (1995) ordfThe power in demography womenrsquos social constructions of gender identity atworkordm Academy of Management Journal Vol 38 pp 589-634

Festinger L (1954) ordfA theory of social comparison processesordm Human Relations Vol 7pp 117-40

Forsyth DR (1990) Group Dynamics BrooksCole Publishing Paciregc Grove CA

Harrison DA Price KH and Bell MP (1998) ordfBeyond relational demography time and theeffects of surface and deep-level diversity on work group cohesionordm Academy ofManagement Journal Vol 41 pp 96-107

James K Lovato C and Khoo G (1994) ordfSocial identity correlates of minority workersrsquo healthordmAcademy of Management Journal Vol 37 p 383+

Kanter R (1977) ordfSome effects of proportions on group life skewed sex ratios and reactions totoken womenordm American Journal of Sociology Vol 82 pp 965-90

Kanter R (1983) Men and Women of the Corporation Basic Books New York NY

Kossek EE and Lobel SA (1996) ordfIntroduction transforming human resource systems tomanage diversityordm in Managing Diversity Human Resource Strategies for Transforming

the Workplace Blackwell Oxford pp 1-19

Kossek EE and Zonia SC (1993) ordfAssessing diversity climate a regeld study of reactions toemployer efforts to promote diversityordm Journal of Organizational Behavior Vol 14pp 61-81

Larkey LK (1996) ordfToward a theory of communicative interactions in culturally diverseworkgroupsordm Academy of Management Review Vol 21 pp 463-91

Lau DC and Murnighan JK (1998) ordfDemographic diversity and faultlines the compositionaldynamics of organizational groupsordm Academy of Management Review Vol 29 pp 325-40

Lindell MK and Brandt CJ (2000) ordfClimate quality and climate consensus as mediators of therelationship between organizational antecedents and outcomesordm Journal of AppliedPsychology Vol 85 pp 331-48

McGrath JE Berdahl JL and Arrow H (1995) ordfTraits expectations culture and clout thedynamics of diversity in work groupsordm in Jackson SE and Ruderman MN (Eds)Diversity in Work Teams Research Paradigms for a Changing Workplace AmericanPsychological Association Washington DC

Payne RL and Pugh DS (1976) ordfOrganizational structure and climateordm in Dunnette MD(Ed) Handbook of Industrial and Organizational Psychology Rand McNally Chicago ILpp 1125-73

Reichers A and Schneider B (1990) ordfClimate and culture an evolution of constructsordm inSchneider B (Ed) Organizational Climate and Culture Jossey-Bass San Francisco CA

HRM changestrategy

351

Table I

Var

iable

2425

2627

2829

1D

epar

tmen

tsi

zeT

ime

12

Dep

artm

ent

size

Tim

e2

3P

ropor

tion

ofw

hit

esT

ime

14

Pro

por

tion

ofw

hit

esT

ime

25

Pro

por

tion

ofw

omen

Tim

e1

6P

ropor

tion

ofw

omen

Tim

e2

7P

ropor

tion

wit

hte

nure

Tim

e1

8P

ropor

tion

wit

hte

nure

Tim

e2

9P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e1

10P

ropor

tion

ofw

hit

ete

nure

dm

ales

Tim

e2

11C

han

ge

indep

artm

ent

size

Tim

e1-

Tim

e2

12C

han

ge

inte

nure

Tim

e1-

Tim

e2

13C

han

ge

ingen

der

Tim

e1-

Tim

e2

14C

han

ge

inw

hit

esT

ime

1-T

ime

215

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Con

sensu

s16

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Con

sensu

s17

P

erce

ived

wor

k-g

roup

mix

plusmnC

onse

nsu

s18

E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Con

sensu

s19

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnC

onse

nsu

s20

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Mea

n21

C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Mea

n22

P

erce

ived

wor

k-g

roup

mix

plusmnM

ean

23E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Mea

n24

E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

ean

plusmn25

D

iver

sity

asor

gan

izat

ional

goa

lplusmn

Med

ian

20

20plusmn

26C

omm

itm

ent

ofm

anag

emen

tto

div

ersi

typlusmn

Med

ian

025

002

plusmn27

P

erce

ived

wor

k-g

roup

mix

plusmnM

edia

n0

080

010

19plusmn

28E

qual

ity

ofdep

artm

ent

suppor

tof

wom

enplusmn

Med

ian

20

32

000

20

010

05plusmn

29E

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

plusmnM

edia

n0

58

20

240

180

132

003

plusmn

Note

p

001

orgre

ater

JOCM163

342

support of resources for women and men and whites and non-whites asthese measures hovered at plusmn024 and 2 025 or close to 0 which would betotal consensus There was the least consensus that minorities and majoritiesmixed well (plusmn125) with slightly more consensus that diversity should bepursued as organizational (plusmn10) The most positive valence was theperception of equality of department support for minorities (393) and thatdiversity should be pursued as a goal (39)

34 Regression results on demographic shift and climateTo address our second research objective which was to examine whetherwork groups with higher demographic shifts had more positive groupclimates for each of the regve dependent variables separate hierarchicalregressions were run to examine the relation between group demographicchange and group consensus (see Table II) or valence (see Table III) In theregrst step of all regressions proportional change in size was entered as acontrol For the consensus regressions the current mean of each dependentvariable was also entered as a control In step 2 the group proportionalchange in women and the group proportional change in minorities wereentered (In early analyses for the second step of each regression interactionterms change in gender or race by tenure were included following Aikenand Westrsquos (1993) procedures No signiregcant interactions were found) Usingstandardized beta weights as seen in step 2 of Tables II and III ourhypothesis that work groups with higher demographic shifts may havehigher consensus and valence toward diversity change received partialsupport Group referents of diversity climate were unchanged bydemographic shifts remacrecting possible dilution of HR strategiesOrganizational referents of valence were effected but consensus wasreduced in terms of whether the organization should pursue diversity as anorganizational goal Departments with higher shifts in the proportion ofwomen had lower consensus on whether the organization should pursuediversity as an organizational goal Both regressions with organizationalreferents as dependent variables were signiregcant Groups with higherincreases in the proportion of women over time tended to have morefavorable valence regarding the desirability of promoting diversity as anorganizational goal but were less sure of managementrsquos commitment asthey had more experience with the reality of implementing theorganizational change

35 Regression results on positive senior women and minority effect on climateOur third research objective was to examine whether increasing the proportionof senior women and minorities at the department level positively relates togroup valence and consensus In our data analysis the proportional change inoverall tenured faculty was entered in step 2 and then the proportional change

HRM changestrategy

343

Org

aniz

atio

nD

epar

tmen

tD

iver

sity

asor

gan

izat

ional

goa

l

Com

mit

men

tof

man

agem

ent

todiv

ersi

ty

Per

ceiv

edw

ork-g

roup

mix

Equal

ity

ofdep

artm

ent

suppor

tof

wom

enE

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

Ste

p1

Dep

tsi

zeT

ime

2-T

ime

10

020

080

180

000

06M

ean

ofD

V0

35

022

017

002

029

D

R2

012

0

060

060

000

09T

otal

R2

012

0

060

060

000

09Ste

p2

Wom

enT

ime

2-T

ime

12

035

2

017

20

182

007

20

02W

hit

esT

ime

2-T

ime

12

002

20

052

003

004

021

Ten

ure

Tim

e2-

Tim

e1

009

003

022

010

000

DR

20

11

003

008

001

004

Tot

alR

20

23

009

014

001

013

Ste

p3

Whit

ete

nure

dm

ales

Tim

e2-

Tim

e1

20

092

012

005

20

302

031

DR

20

010

010

000

040

05T

otal

R2

024

0

100

140

050

18

Note

p

005

orgre

ater

p0

01or

gre

ater

Table IIRegressionpredictingconsensus onclimate for diversity

JOCM163

344

Org

aniz

atio

nD

epar

tmen

tD

iver

sity

asor

gan

izat

ional

goa

l

Com

mit

men

tof

man

agem

ent

todiv

ersi

ty

Per

ceiv

edw

ork-g

roup

mix

Equal

ity

ofdep

artm

ent

suppor

tof

wom

enE

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

Ste

p1

Dep

tsi

zeT

ime

2-T

ime

10

080

152

006

021

0

05D

R2

001

002

000

004

0

00T

otal

R2

001

002

000

004

0

00Ste

p2

Wom

enT

ime

2-T

ime

10

40

20

38

20

21

20

062

013

Whit

esT

ime

2-T

ime

12

007

019

008

20

172

001

Ten

ure

Tim

e2-

Tim

e1

20

072

013

012

016

019

DR

20

16

017

0

060

070

05T

otal

R2

017

0

19

006

011

0

05Ste

p3

Whit

ete

nure

dm

ales

Tim

e2-

Tim

e1

005

011

000

007

035

D

R2

000

001

000

000

004

T

otal

R2

017

0

20

006

011

009

Note

p

005

orgre

ater

p0

01or

gre

ater

Table IIIRegression

predicting valenceof climate for

diversity

HRM changestrategy

345

in tenured white males was entered in step 3 The results are that the thirdblock was not statistically signiregcant in the regressions showing that smallincreases in the group proportions of senior women and minorities did notincrease group consensus or valence regarding climate for change

4 DiscussionThis study explored links between actual change in organizational structureorganizational demography and work group interpretation of this change Ourresults showed that HR strategies that focus on structural change withoutworking to develop supportive group norms and positive climate areinadequate change strategies Current employer efforts to increase diversity asan isolated HR strategy may not necessarily lead to increased memberagreement that change is favorable Most change efforts like the diversitychange effort we studied are focused at the organizational or individualintervention levels and under-emphasize the work group level and theimportance of group tipping points Our study highlights a disconnect betweenorganizational level diversity practice which often assumes that hiring moreminorities and women will improve climate and what our research found at thegroup level a climate that was mixed and ambivalent at best and sometimesnegative

41 Group level measures of demographic change consensus and valence ondiversity climateThe regrst objective of this study was to develop new theory and measures ofgroup level change in diversity We provided new ways to measure and groupsocial construction of change These included consensus the degree to whichmembers held common perceptions of diversity climate and valence thedirection of those perceptions Our measures identify group and organizationallevel referents of HR strategies to increase demographic diversity Group levelreferents of diversity change were how well members perceived their workgroup mix and the perceived equality of resource allocation acrossdemographic groups Organizational referents were managementcommitment to change and whether diversity should be an organizationalstrategy assessed general support for change Our study highlights howimportant it is for research and practice to not only examine how change indemography relates to differences in individual attitudes but also to groupmember shared perceptions about the direction of the climate for changeUnderstanding linkages between HR changes (in this case group demographywhich raises unique psychological processes related to social identity) andpositive consensus is important because it is relevant to cooperation workexperiences interpersonal conmacrict stress and norms in groups striving tomanage change

JOCM163

346

42 Greater work group demographic shift does not necessarily result infavorable climateOur second objective was to assess whether work groups with higherdemographic shifts tend to have more positive climates that is higherconsensus and valence toward diversity change for both organizational andgroup level change referents Our results only partially supported this beliefthat greater structural demographic shift would positively relate to a morefavorable climate for diversity Although at the organizational level theuniversity had experienced signiregcant change in organizational demographyan increase in the overall representation of white women (36 percent) andminorities (41 percent) over time work group members in units with thegreatest change did not have higher consensus or more favorable valenceWe found more favorable results for organizational than group levelreferents of change Members may regnd it easier to agree with organizationalreferents and more difregcult to agree on group level referents due to theirgreater psychological proximity to group processes that affect how theypersonally experience conmacrict and wellbeing from change in their dailywork Work groups with more demographic change had higher valence butlower consensus on whether the organization should pursue diversity as astrategic goal When organizational change strategies to increase diversityare enacted at the work group level the group demographic shift may bediluted across the organization to only slightly increase the representation ofwomen and minorities within actual groups Due to this dilution effectgroup membersrsquo social construction of the HR strategies may lack consensusor positive valence regarding the organizational diversity change strategiesIf the change is not large enough to create a critical mass to tiprepresentation of minorities to at least 35 percent according to Kanter (1977)negative intergroup dynamics such as increased resource competition andtokenism may still occur

43 Diluting diversity gradually increasing the proportion of senior womenminorities does not necessarily improve group climateOur third objective was to investigate whether increasing the proportion ofsenior women and minorities at the department level positively relates togroup valence and consensus There was no relation between increasing theproportion of senior women and minorities and group valence andconsensus Due to diluted saliency of organizational HR change strategies toincrease diversity at the work group level slight demographic shifts toproportionally increase the representation of senior women and minoritiesdid not relate to higher group valence or consensus This increase did littleto improve group climate and may have even hurt the climate in theshort run

HRM changestrategy

347

5 Implications for future research and practice51 Incremental structural change not sufregcient to improve climateWe hope this study will encourage researchers and practitioners interested indiversity change strategies to augment individual and organizational levels ofintervention to focus on work group climates supporting change Our resultshighlight some disparities between the underlying assumptions of HRstrategies to manage organizational change (increasing structural diversitywill be positive for climate) and reality at the department level (incrementalincreases did not improve climate) Organizational strategists implementingdiversity recruitment initiatives typically assume that change strategies (in thiscase hiring diversity) would improve the group social climate due to a positivechange in member attitudes based on increased group diversity and positivesocial contact Yet our study does not support this view Although theorganization we studied had had signiregcant success in enacting the formal HRstrategy plusmn increasing the representation of women and minorities at theorganizational level plusmn our research shows that gradually shifting thedemography of work groups may not necessarily lead to increased consensusor valence about the social climate within work groups For example it did notlead to improved climate regarding the perceived fairness of resourceallocation or good mixing in social interaction In fact the incremental increasein the proportion of minority groups lowered consensus and agreement in theshort term as subgroups sorted out shifting and socially ambiguous powerrelationships within work units These results suggest the importance ofmeasuring change initiatives impacts at the group level and having group levelinterventions such as team-building and group-focused resource allocation tosupport organizational level diversity strategies

52 Focus change strategies on tipping points within targeted work unitsFuture research should consider how HR strategies to manage organizationaldemographic change are socially constructed and enacted at the work grouplevel It is especially important to identify what constitutes a group leveltipping point or critical mass across hierarchical levels and differentdemographic groups to enable positive consensus supporting changePractice might regnd it more fruitful to focus future HR strategies to increasediversity through hiring to alter tipping points within targeted work groupsThis will dramatically alter the saliency of strategic organizationaldemographic change and intervene to shape the climate of speciregc units Ourstudy suggests that organizations may need demographic earthquakes withinwork groups to effect group consensus and valence on the climate regardingthe HR change (ie managing diversity) Dilution of change at the group levelmay ameliorate intended positive effects unless certain group level tippingpoints such as at least 35 percent of the work group are effected in each unit by

JOCM163

348

the HR strategy After eight years of hiring new faculty in our sample womenhovered at about 15 percent in 89 percent of groups and minorities had amaximum representation of 25 percent The demographic shifts whileincreasing were not dramatic enough to create imbalances that markedlyincreased the salience and power of new demographic subgroups created bythe HR strategies Instead of being balanced many of our groups were stilllargely white male and remain tilted or skewed over time which Kanter (1977)notes perpetuates stereotypical negative dynamics

53 Avoiding backlash weaving tipping points across the hierarchyRecent work at MIT by Bailyn (2000) suggests that backlash and resistance toorganizational change may be occurring among senior white male faculty inresponse to HR diversity strategies Initially women and minorities now enterdepartments with relatively equal resources to white men as supported by theHR system design However by the time women and minorities reach theassociate and full professor levels subtle discrimination processes occur in howHR policies are applied such as lower access to resources promotions andpoor climate Despite the increase in upward mobility the representation ofwomen and minorities at the top is still relatively low Ely (1994) found that ifwomen are under-represented at higher organizational levels relations betweenwomen at lower levels might be of lower quality due to increased competitionand negative gender dynamics in the larger social system If HR changestrategies that alter the distribution of existing resources across organizationaland demographic groups are to be successful they must take into account hownot only tipping points in terms of overall departmental representation butmore importantly how tipping points are distributed in demographies acrossthe hierarchies of work groups or departments

HR diversity change strategies that bring in many new hires of differentbackgrounds may have negative ramiregcations for social functioning in groupsparticularly if new members are not supported by additional HR strategies thatallow them to enter work groups on an equal footing or if their work groups donot have leadership that mirror their demographic identities As Kanter (1983)found identity groups must be equal in their access to power resources ifcontact is to produce improvement in attitudes There must also be time for thenew members to be socialized and integrated into the culture Harrison et al(1998) found that the longer the length of time that members worked togetherthe greater the effects of deep level attitude similarity Organizations areincreasing surface (demographic) and deep level (attitudinal) diversitysimultaneously which is likely to have negative short-term ramiregcations forthe consensus and direction of climate We argue HR strategies must not onlyfocus on the surface level plusmn remacrecting structural and demographic attributes plusmnbut also simultaneously to deep level characteristics remacrecting values within

HRM changestrategy

349

the context of speciregc groups at all organizational levels In sum HR strategistsand change agents should focus on how HR change strategies relate to the jointalignment of group memberrsquos multiple characteristics (eg demographichierarchical status task) in work groups plusmn that is to examine group contextualinmacruences (Wharton 1992)

54 LimitationsOur study while making contributions to organizational research is notwithout limitations Though large in individual respondents by utilizing thegroup as the level of analysis our sample size includes only 74 groups Thiscould reduce the level of statistical power of our regndings Additionally becausethe sample was based on one organization it is important to replicate thisanalysis across multiple organizations to substantiate our results Althoughchanges in department heterogeneity across an eight-year period wereincluded this is not a longitudinal design Future research should employrepeated measures over a length of time across all variables measuringdemographic change in heterogeneity as well as consensus and valence Ourstudy points to some of the challenges in measuring ordfsuccess in increasingdiversityordm as we did not look at annual turnover or the effects of the glassceiling at the group level in this study which is a gap we hope futureresearchers will be motivated to study spurred by our research

Since little published work has been done on group consensus regarding theeffects of HR policies to increase diversity through demographic change weextend many social science literatures ranging from organizationaldevelopment to demography and HR Future research on diversity andorganizational change also should strive to be interdisciplinary as theliterature on HR strategy and change has been largely developed in the HRliterature separate from other social science disciplines reducing theoreticalintegration Without such integration and increased scholarly and practitionerattention to climate and consensus many organizations may fail to supportgroup membersrsquo explicit use of their cultural experiences to advanceorganizational learning from enacted diversity change strategies despite thegroup and intergroup social tensions that will inevitably occur (Thomas andEly 1996)

References

Aiken LA and West SG (1993) Multiple Regression Testing and Interpreting InteractionsSage Publications Newbury Park CA

Alderfer CP and Smith KK (1982) ordfStudying embedded intergroup relations inorganizationsordm Administrative Science Quarterly Vol 27 pp 35-65

Allport G (1954) The Nature of Prejudice Addison Wesley Cambridge MA

Bailyn L (2000) ordfGender and diversity in organizations preconference presentationordm paperpresented at the National Academy of Management meeting Toronto

JOCM163

350

Blalock HM (1967) Toward a Theory of Minority Group Relations Wiley New York NY

Bleise PD and Halverson RR (1998) ordfGroup consensus and psychological well-being a largeregeld studyordm Journal of Applied Social Psychology Vol 28 pp 563-80

Brown KG Kozlowski SWJ and Hattrup K (1996) Theory Issues and Recommendations inConceptualizing Agreement as a Construct in Organizational Research The Search forConsensus Regarding Consensus National Academy of Management Cincinnati OH

Cox T Jr (1993) Cultural Diversity in Organizations Berrett-Koehler San Francisco CA

Ely R (1994) ordfThe effects of organizational demographics and social identity on relationshipsamong professional womenordm Administrative Science Quarterly Vol 39 pp 203-38

Ely R (1995) ordfThe power in demography womenrsquos social constructions of gender identity atworkordm Academy of Management Journal Vol 38 pp 589-634

Festinger L (1954) ordfA theory of social comparison processesordm Human Relations Vol 7pp 117-40

Forsyth DR (1990) Group Dynamics BrooksCole Publishing Paciregc Grove CA

Harrison DA Price KH and Bell MP (1998) ordfBeyond relational demography time and theeffects of surface and deep-level diversity on work group cohesionordm Academy ofManagement Journal Vol 41 pp 96-107

James K Lovato C and Khoo G (1994) ordfSocial identity correlates of minority workersrsquo healthordmAcademy of Management Journal Vol 37 p 383+

Kanter R (1977) ordfSome effects of proportions on group life skewed sex ratios and reactions totoken womenordm American Journal of Sociology Vol 82 pp 965-90

Kanter R (1983) Men and Women of the Corporation Basic Books New York NY

Kossek EE and Lobel SA (1996) ordfIntroduction transforming human resource systems tomanage diversityordm in Managing Diversity Human Resource Strategies for Transforming

the Workplace Blackwell Oxford pp 1-19

Kossek EE and Zonia SC (1993) ordfAssessing diversity climate a regeld study of reactions toemployer efforts to promote diversityordm Journal of Organizational Behavior Vol 14pp 61-81

Larkey LK (1996) ordfToward a theory of communicative interactions in culturally diverseworkgroupsordm Academy of Management Review Vol 21 pp 463-91

Lau DC and Murnighan JK (1998) ordfDemographic diversity and faultlines the compositionaldynamics of organizational groupsordm Academy of Management Review Vol 29 pp 325-40

Lindell MK and Brandt CJ (2000) ordfClimate quality and climate consensus as mediators of therelationship between organizational antecedents and outcomesordm Journal of AppliedPsychology Vol 85 pp 331-48

McGrath JE Berdahl JL and Arrow H (1995) ordfTraits expectations culture and clout thedynamics of diversity in work groupsordm in Jackson SE and Ruderman MN (Eds)Diversity in Work Teams Research Paradigms for a Changing Workplace AmericanPsychological Association Washington DC

Payne RL and Pugh DS (1976) ordfOrganizational structure and climateordm in Dunnette MD(Ed) Handbook of Industrial and Organizational Psychology Rand McNally Chicago ILpp 1125-73

Reichers A and Schneider B (1990) ordfClimate and culture an evolution of constructsordm inSchneider B (Ed) Organizational Climate and Culture Jossey-Bass San Francisco CA

HRM changestrategy

351

support of resources for women and men and whites and non-whites asthese measures hovered at plusmn024 and 2 025 or close to 0 which would betotal consensus There was the least consensus that minorities and majoritiesmixed well (plusmn125) with slightly more consensus that diversity should bepursued as organizational (plusmn10) The most positive valence was theperception of equality of department support for minorities (393) and thatdiversity should be pursued as a goal (39)

34 Regression results on demographic shift and climateTo address our second research objective which was to examine whetherwork groups with higher demographic shifts had more positive groupclimates for each of the regve dependent variables separate hierarchicalregressions were run to examine the relation between group demographicchange and group consensus (see Table II) or valence (see Table III) In theregrst step of all regressions proportional change in size was entered as acontrol For the consensus regressions the current mean of each dependentvariable was also entered as a control In step 2 the group proportionalchange in women and the group proportional change in minorities wereentered (In early analyses for the second step of each regression interactionterms change in gender or race by tenure were included following Aikenand Westrsquos (1993) procedures No signiregcant interactions were found) Usingstandardized beta weights as seen in step 2 of Tables II and III ourhypothesis that work groups with higher demographic shifts may havehigher consensus and valence toward diversity change received partialsupport Group referents of diversity climate were unchanged bydemographic shifts remacrecting possible dilution of HR strategiesOrganizational referents of valence were effected but consensus wasreduced in terms of whether the organization should pursue diversity as anorganizational goal Departments with higher shifts in the proportion ofwomen had lower consensus on whether the organization should pursuediversity as an organizational goal Both regressions with organizationalreferents as dependent variables were signiregcant Groups with higherincreases in the proportion of women over time tended to have morefavorable valence regarding the desirability of promoting diversity as anorganizational goal but were less sure of managementrsquos commitment asthey had more experience with the reality of implementing theorganizational change

35 Regression results on positive senior women and minority effect on climateOur third research objective was to examine whether increasing the proportionof senior women and minorities at the department level positively relates togroup valence and consensus In our data analysis the proportional change inoverall tenured faculty was entered in step 2 and then the proportional change

HRM changestrategy

343

Org

aniz

atio

nD

epar

tmen

tD

iver

sity

asor

gan

izat

ional

goa

l

Com

mit

men

tof

man

agem

ent

todiv

ersi

ty

Per

ceiv

edw

ork-g

roup

mix

Equal

ity

ofdep

artm

ent

suppor

tof

wom

enE

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

Ste

p1

Dep

tsi

zeT

ime

2-T

ime

10

020

080

180

000

06M

ean

ofD

V0

35

022

017

002

029

D

R2

012

0

060

060

000

09T

otal

R2

012

0

060

060

000

09Ste

p2

Wom

enT

ime

2-T

ime

12

035

2

017

20

182

007

20

02W

hit

esT

ime

2-T

ime

12

002

20

052

003

004

021

Ten

ure

Tim

e2-

Tim

e1

009

003

022

010

000

DR

20

11

003

008

001

004

Tot

alR

20

23

009

014

001

013

Ste

p3

Whit

ete

nure

dm

ales

Tim

e2-

Tim

e1

20

092

012

005

20

302

031

DR

20

010

010

000

040

05T

otal

R2

024

0

100

140

050

18

Note

p

005

orgre

ater

p0

01or

gre

ater

Table IIRegressionpredictingconsensus onclimate for diversity

JOCM163

344

Org

aniz

atio

nD

epar

tmen

tD

iver

sity

asor

gan

izat

ional

goa

l

Com

mit

men

tof

man

agem

ent

todiv

ersi

ty

Per

ceiv

edw

ork-g

roup

mix

Equal

ity

ofdep

artm

ent

suppor

tof

wom

enE

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

Ste

p1

Dep

tsi

zeT

ime

2-T

ime

10

080

152

006

021

0

05D

R2

001

002

000

004

0

00T

otal

R2

001

002

000

004

0

00Ste

p2

Wom

enT

ime

2-T

ime

10

40

20

38

20

21

20

062

013

Whit

esT

ime

2-T

ime

12

007

019

008

20

172

001

Ten

ure

Tim

e2-

Tim

e1

20

072

013

012

016

019

DR

20

16

017

0

060

070

05T

otal

R2

017

0

19

006

011

0

05Ste

p3

Whit

ete

nure

dm

ales

Tim

e2-

Tim

e1

005

011

000

007

035

D

R2

000

001

000

000

004

T

otal

R2

017

0

20

006

011

009

Note

p

005

orgre

ater

p0

01or

gre

ater

Table IIIRegression

predicting valenceof climate for

diversity

HRM changestrategy

345

in tenured white males was entered in step 3 The results are that the thirdblock was not statistically signiregcant in the regressions showing that smallincreases in the group proportions of senior women and minorities did notincrease group consensus or valence regarding climate for change

4 DiscussionThis study explored links between actual change in organizational structureorganizational demography and work group interpretation of this change Ourresults showed that HR strategies that focus on structural change withoutworking to develop supportive group norms and positive climate areinadequate change strategies Current employer efforts to increase diversity asan isolated HR strategy may not necessarily lead to increased memberagreement that change is favorable Most change efforts like the diversitychange effort we studied are focused at the organizational or individualintervention levels and under-emphasize the work group level and theimportance of group tipping points Our study highlights a disconnect betweenorganizational level diversity practice which often assumes that hiring moreminorities and women will improve climate and what our research found at thegroup level a climate that was mixed and ambivalent at best and sometimesnegative

41 Group level measures of demographic change consensus and valence ondiversity climateThe regrst objective of this study was to develop new theory and measures ofgroup level change in diversity We provided new ways to measure and groupsocial construction of change These included consensus the degree to whichmembers held common perceptions of diversity climate and valence thedirection of those perceptions Our measures identify group and organizationallevel referents of HR strategies to increase demographic diversity Group levelreferents of diversity change were how well members perceived their workgroup mix and the perceived equality of resource allocation acrossdemographic groups Organizational referents were managementcommitment to change and whether diversity should be an organizationalstrategy assessed general support for change Our study highlights howimportant it is for research and practice to not only examine how change indemography relates to differences in individual attitudes but also to groupmember shared perceptions about the direction of the climate for changeUnderstanding linkages between HR changes (in this case group demographywhich raises unique psychological processes related to social identity) andpositive consensus is important because it is relevant to cooperation workexperiences interpersonal conmacrict stress and norms in groups striving tomanage change

JOCM163

346

42 Greater work group demographic shift does not necessarily result infavorable climateOur second objective was to assess whether work groups with higherdemographic shifts tend to have more positive climates that is higherconsensus and valence toward diversity change for both organizational andgroup level change referents Our results only partially supported this beliefthat greater structural demographic shift would positively relate to a morefavorable climate for diversity Although at the organizational level theuniversity had experienced signiregcant change in organizational demographyan increase in the overall representation of white women (36 percent) andminorities (41 percent) over time work group members in units with thegreatest change did not have higher consensus or more favorable valenceWe found more favorable results for organizational than group levelreferents of change Members may regnd it easier to agree with organizationalreferents and more difregcult to agree on group level referents due to theirgreater psychological proximity to group processes that affect how theypersonally experience conmacrict and wellbeing from change in their dailywork Work groups with more demographic change had higher valence butlower consensus on whether the organization should pursue diversity as astrategic goal When organizational change strategies to increase diversityare enacted at the work group level the group demographic shift may bediluted across the organization to only slightly increase the representation ofwomen and minorities within actual groups Due to this dilution effectgroup membersrsquo social construction of the HR strategies may lack consensusor positive valence regarding the organizational diversity change strategiesIf the change is not large enough to create a critical mass to tiprepresentation of minorities to at least 35 percent according to Kanter (1977)negative intergroup dynamics such as increased resource competition andtokenism may still occur

43 Diluting diversity gradually increasing the proportion of senior womenminorities does not necessarily improve group climateOur third objective was to investigate whether increasing the proportion ofsenior women and minorities at the department level positively relates togroup valence and consensus There was no relation between increasing theproportion of senior women and minorities and group valence andconsensus Due to diluted saliency of organizational HR change strategies toincrease diversity at the work group level slight demographic shifts toproportionally increase the representation of senior women and minoritiesdid not relate to higher group valence or consensus This increase did littleto improve group climate and may have even hurt the climate in theshort run

HRM changestrategy

347

5 Implications for future research and practice51 Incremental structural change not sufregcient to improve climateWe hope this study will encourage researchers and practitioners interested indiversity change strategies to augment individual and organizational levels ofintervention to focus on work group climates supporting change Our resultshighlight some disparities between the underlying assumptions of HRstrategies to manage organizational change (increasing structural diversitywill be positive for climate) and reality at the department level (incrementalincreases did not improve climate) Organizational strategists implementingdiversity recruitment initiatives typically assume that change strategies (in thiscase hiring diversity) would improve the group social climate due to a positivechange in member attitudes based on increased group diversity and positivesocial contact Yet our study does not support this view Although theorganization we studied had had signiregcant success in enacting the formal HRstrategy plusmn increasing the representation of women and minorities at theorganizational level plusmn our research shows that gradually shifting thedemography of work groups may not necessarily lead to increased consensusor valence about the social climate within work groups For example it did notlead to improved climate regarding the perceived fairness of resourceallocation or good mixing in social interaction In fact the incremental increasein the proportion of minority groups lowered consensus and agreement in theshort term as subgroups sorted out shifting and socially ambiguous powerrelationships within work units These results suggest the importance ofmeasuring change initiatives impacts at the group level and having group levelinterventions such as team-building and group-focused resource allocation tosupport organizational level diversity strategies

52 Focus change strategies on tipping points within targeted work unitsFuture research should consider how HR strategies to manage organizationaldemographic change are socially constructed and enacted at the work grouplevel It is especially important to identify what constitutes a group leveltipping point or critical mass across hierarchical levels and differentdemographic groups to enable positive consensus supporting changePractice might regnd it more fruitful to focus future HR strategies to increasediversity through hiring to alter tipping points within targeted work groupsThis will dramatically alter the saliency of strategic organizationaldemographic change and intervene to shape the climate of speciregc units Ourstudy suggests that organizations may need demographic earthquakes withinwork groups to effect group consensus and valence on the climate regardingthe HR change (ie managing diversity) Dilution of change at the group levelmay ameliorate intended positive effects unless certain group level tippingpoints such as at least 35 percent of the work group are effected in each unit by

JOCM163

348

the HR strategy After eight years of hiring new faculty in our sample womenhovered at about 15 percent in 89 percent of groups and minorities had amaximum representation of 25 percent The demographic shifts whileincreasing were not dramatic enough to create imbalances that markedlyincreased the salience and power of new demographic subgroups created bythe HR strategies Instead of being balanced many of our groups were stilllargely white male and remain tilted or skewed over time which Kanter (1977)notes perpetuates stereotypical negative dynamics

53 Avoiding backlash weaving tipping points across the hierarchyRecent work at MIT by Bailyn (2000) suggests that backlash and resistance toorganizational change may be occurring among senior white male faculty inresponse to HR diversity strategies Initially women and minorities now enterdepartments with relatively equal resources to white men as supported by theHR system design However by the time women and minorities reach theassociate and full professor levels subtle discrimination processes occur in howHR policies are applied such as lower access to resources promotions andpoor climate Despite the increase in upward mobility the representation ofwomen and minorities at the top is still relatively low Ely (1994) found that ifwomen are under-represented at higher organizational levels relations betweenwomen at lower levels might be of lower quality due to increased competitionand negative gender dynamics in the larger social system If HR changestrategies that alter the distribution of existing resources across organizationaland demographic groups are to be successful they must take into account hownot only tipping points in terms of overall departmental representation butmore importantly how tipping points are distributed in demographies acrossthe hierarchies of work groups or departments

HR diversity change strategies that bring in many new hires of differentbackgrounds may have negative ramiregcations for social functioning in groupsparticularly if new members are not supported by additional HR strategies thatallow them to enter work groups on an equal footing or if their work groups donot have leadership that mirror their demographic identities As Kanter (1983)found identity groups must be equal in their access to power resources ifcontact is to produce improvement in attitudes There must also be time for thenew members to be socialized and integrated into the culture Harrison et al(1998) found that the longer the length of time that members worked togetherthe greater the effects of deep level attitude similarity Organizations areincreasing surface (demographic) and deep level (attitudinal) diversitysimultaneously which is likely to have negative short-term ramiregcations forthe consensus and direction of climate We argue HR strategies must not onlyfocus on the surface level plusmn remacrecting structural and demographic attributes plusmnbut also simultaneously to deep level characteristics remacrecting values within

HRM changestrategy

349

the context of speciregc groups at all organizational levels In sum HR strategistsand change agents should focus on how HR change strategies relate to the jointalignment of group memberrsquos multiple characteristics (eg demographichierarchical status task) in work groups plusmn that is to examine group contextualinmacruences (Wharton 1992)

54 LimitationsOur study while making contributions to organizational research is notwithout limitations Though large in individual respondents by utilizing thegroup as the level of analysis our sample size includes only 74 groups Thiscould reduce the level of statistical power of our regndings Additionally becausethe sample was based on one organization it is important to replicate thisanalysis across multiple organizations to substantiate our results Althoughchanges in department heterogeneity across an eight-year period wereincluded this is not a longitudinal design Future research should employrepeated measures over a length of time across all variables measuringdemographic change in heterogeneity as well as consensus and valence Ourstudy points to some of the challenges in measuring ordfsuccess in increasingdiversityordm as we did not look at annual turnover or the effects of the glassceiling at the group level in this study which is a gap we hope futureresearchers will be motivated to study spurred by our research

Since little published work has been done on group consensus regarding theeffects of HR policies to increase diversity through demographic change weextend many social science literatures ranging from organizationaldevelopment to demography and HR Future research on diversity andorganizational change also should strive to be interdisciplinary as theliterature on HR strategy and change has been largely developed in the HRliterature separate from other social science disciplines reducing theoreticalintegration Without such integration and increased scholarly and practitionerattention to climate and consensus many organizations may fail to supportgroup membersrsquo explicit use of their cultural experiences to advanceorganizational learning from enacted diversity change strategies despite thegroup and intergroup social tensions that will inevitably occur (Thomas andEly 1996)

References

Aiken LA and West SG (1993) Multiple Regression Testing and Interpreting InteractionsSage Publications Newbury Park CA

Alderfer CP and Smith KK (1982) ordfStudying embedded intergroup relations inorganizationsordm Administrative Science Quarterly Vol 27 pp 35-65

Allport G (1954) The Nature of Prejudice Addison Wesley Cambridge MA

Bailyn L (2000) ordfGender and diversity in organizations preconference presentationordm paperpresented at the National Academy of Management meeting Toronto

JOCM163

350

Blalock HM (1967) Toward a Theory of Minority Group Relations Wiley New York NY

Bleise PD and Halverson RR (1998) ordfGroup consensus and psychological well-being a largeregeld studyordm Journal of Applied Social Psychology Vol 28 pp 563-80

Brown KG Kozlowski SWJ and Hattrup K (1996) Theory Issues and Recommendations inConceptualizing Agreement as a Construct in Organizational Research The Search forConsensus Regarding Consensus National Academy of Management Cincinnati OH

Cox T Jr (1993) Cultural Diversity in Organizations Berrett-Koehler San Francisco CA

Ely R (1994) ordfThe effects of organizational demographics and social identity on relationshipsamong professional womenordm Administrative Science Quarterly Vol 39 pp 203-38

Ely R (1995) ordfThe power in demography womenrsquos social constructions of gender identity atworkordm Academy of Management Journal Vol 38 pp 589-634

Festinger L (1954) ordfA theory of social comparison processesordm Human Relations Vol 7pp 117-40

Forsyth DR (1990) Group Dynamics BrooksCole Publishing Paciregc Grove CA

Harrison DA Price KH and Bell MP (1998) ordfBeyond relational demography time and theeffects of surface and deep-level diversity on work group cohesionordm Academy ofManagement Journal Vol 41 pp 96-107

James K Lovato C and Khoo G (1994) ordfSocial identity correlates of minority workersrsquo healthordmAcademy of Management Journal Vol 37 p 383+

Kanter R (1977) ordfSome effects of proportions on group life skewed sex ratios and reactions totoken womenordm American Journal of Sociology Vol 82 pp 965-90

Kanter R (1983) Men and Women of the Corporation Basic Books New York NY

Kossek EE and Lobel SA (1996) ordfIntroduction transforming human resource systems tomanage diversityordm in Managing Diversity Human Resource Strategies for Transforming

the Workplace Blackwell Oxford pp 1-19

Kossek EE and Zonia SC (1993) ordfAssessing diversity climate a regeld study of reactions toemployer efforts to promote diversityordm Journal of Organizational Behavior Vol 14pp 61-81

Larkey LK (1996) ordfToward a theory of communicative interactions in culturally diverseworkgroupsordm Academy of Management Review Vol 21 pp 463-91

Lau DC and Murnighan JK (1998) ordfDemographic diversity and faultlines the compositionaldynamics of organizational groupsordm Academy of Management Review Vol 29 pp 325-40

Lindell MK and Brandt CJ (2000) ordfClimate quality and climate consensus as mediators of therelationship between organizational antecedents and outcomesordm Journal of AppliedPsychology Vol 85 pp 331-48

McGrath JE Berdahl JL and Arrow H (1995) ordfTraits expectations culture and clout thedynamics of diversity in work groupsordm in Jackson SE and Ruderman MN (Eds)Diversity in Work Teams Research Paradigms for a Changing Workplace AmericanPsychological Association Washington DC

Payne RL and Pugh DS (1976) ordfOrganizational structure and climateordm in Dunnette MD(Ed) Handbook of Industrial and Organizational Psychology Rand McNally Chicago ILpp 1125-73

Reichers A and Schneider B (1990) ordfClimate and culture an evolution of constructsordm inSchneider B (Ed) Organizational Climate and Culture Jossey-Bass San Francisco CA

HRM changestrategy

351

Org

aniz

atio

nD

epar

tmen

tD

iver

sity

asor

gan

izat

ional

goa

l

Com

mit

men

tof

man

agem

ent

todiv

ersi

ty

Per

ceiv

edw

ork-g

roup

mix

Equal

ity

ofdep

artm

ent

suppor

tof

wom

enE

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

Ste

p1

Dep

tsi

zeT

ime

2-T

ime

10

020

080

180

000

06M

ean

ofD

V0

35

022

017

002

029

D

R2

012

0

060

060

000

09T

otal

R2

012

0

060

060

000

09Ste

p2

Wom

enT

ime

2-T

ime

12

035

2

017

20

182

007

20

02W

hit

esT

ime

2-T

ime

12

002

20

052

003

004

021

Ten

ure

Tim

e2-

Tim

e1

009

003

022

010

000

DR

20

11

003

008

001

004

Tot

alR

20

23

009

014

001

013

Ste

p3

Whit

ete

nure

dm

ales

Tim

e2-

Tim

e1

20

092

012

005

20

302

031

DR

20

010

010

000

040

05T

otal

R2

024

0

100

140

050

18

Note

p

005

orgre

ater

p0

01or

gre

ater

Table IIRegressionpredictingconsensus onclimate for diversity

JOCM163

344

Org

aniz

atio

nD

epar

tmen

tD

iver

sity

asor

gan

izat

ional

goa

l

Com

mit

men

tof

man

agem

ent

todiv

ersi

ty

Per

ceiv

edw

ork-g

roup

mix

Equal

ity

ofdep

artm

ent

suppor

tof

wom

enE

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

Ste

p1

Dep

tsi

zeT

ime

2-T

ime

10

080

152

006

021

0

05D

R2

001

002

000

004

0

00T

otal

R2

001

002

000

004

0

00Ste

p2

Wom

enT

ime

2-T

ime

10

40

20

38

20

21

20

062

013

Whit

esT

ime

2-T

ime

12

007

019

008

20

172

001

Ten

ure

Tim

e2-

Tim

e1

20

072

013

012

016

019

DR

20

16

017

0

060

070

05T

otal

R2

017

0

19

006

011

0

05Ste

p3

Whit

ete

nure

dm

ales

Tim

e2-

Tim

e1

005

011

000

007

035

D

R2

000

001

000

000

004

T

otal

R2

017

0

20

006

011

009

Note

p

005

orgre

ater

p0

01or

gre

ater

Table IIIRegression

predicting valenceof climate for

diversity

HRM changestrategy

345

in tenured white males was entered in step 3 The results are that the thirdblock was not statistically signiregcant in the regressions showing that smallincreases in the group proportions of senior women and minorities did notincrease group consensus or valence regarding climate for change

4 DiscussionThis study explored links between actual change in organizational structureorganizational demography and work group interpretation of this change Ourresults showed that HR strategies that focus on structural change withoutworking to develop supportive group norms and positive climate areinadequate change strategies Current employer efforts to increase diversity asan isolated HR strategy may not necessarily lead to increased memberagreement that change is favorable Most change efforts like the diversitychange effort we studied are focused at the organizational or individualintervention levels and under-emphasize the work group level and theimportance of group tipping points Our study highlights a disconnect betweenorganizational level diversity practice which often assumes that hiring moreminorities and women will improve climate and what our research found at thegroup level a climate that was mixed and ambivalent at best and sometimesnegative

41 Group level measures of demographic change consensus and valence ondiversity climateThe regrst objective of this study was to develop new theory and measures ofgroup level change in diversity We provided new ways to measure and groupsocial construction of change These included consensus the degree to whichmembers held common perceptions of diversity climate and valence thedirection of those perceptions Our measures identify group and organizationallevel referents of HR strategies to increase demographic diversity Group levelreferents of diversity change were how well members perceived their workgroup mix and the perceived equality of resource allocation acrossdemographic groups Organizational referents were managementcommitment to change and whether diversity should be an organizationalstrategy assessed general support for change Our study highlights howimportant it is for research and practice to not only examine how change indemography relates to differences in individual attitudes but also to groupmember shared perceptions about the direction of the climate for changeUnderstanding linkages between HR changes (in this case group demographywhich raises unique psychological processes related to social identity) andpositive consensus is important because it is relevant to cooperation workexperiences interpersonal conmacrict stress and norms in groups striving tomanage change

JOCM163

346

42 Greater work group demographic shift does not necessarily result infavorable climateOur second objective was to assess whether work groups with higherdemographic shifts tend to have more positive climates that is higherconsensus and valence toward diversity change for both organizational andgroup level change referents Our results only partially supported this beliefthat greater structural demographic shift would positively relate to a morefavorable climate for diversity Although at the organizational level theuniversity had experienced signiregcant change in organizational demographyan increase in the overall representation of white women (36 percent) andminorities (41 percent) over time work group members in units with thegreatest change did not have higher consensus or more favorable valenceWe found more favorable results for organizational than group levelreferents of change Members may regnd it easier to agree with organizationalreferents and more difregcult to agree on group level referents due to theirgreater psychological proximity to group processes that affect how theypersonally experience conmacrict and wellbeing from change in their dailywork Work groups with more demographic change had higher valence butlower consensus on whether the organization should pursue diversity as astrategic goal When organizational change strategies to increase diversityare enacted at the work group level the group demographic shift may bediluted across the organization to only slightly increase the representation ofwomen and minorities within actual groups Due to this dilution effectgroup membersrsquo social construction of the HR strategies may lack consensusor positive valence regarding the organizational diversity change strategiesIf the change is not large enough to create a critical mass to tiprepresentation of minorities to at least 35 percent according to Kanter (1977)negative intergroup dynamics such as increased resource competition andtokenism may still occur

43 Diluting diversity gradually increasing the proportion of senior womenminorities does not necessarily improve group climateOur third objective was to investigate whether increasing the proportion ofsenior women and minorities at the department level positively relates togroup valence and consensus There was no relation between increasing theproportion of senior women and minorities and group valence andconsensus Due to diluted saliency of organizational HR change strategies toincrease diversity at the work group level slight demographic shifts toproportionally increase the representation of senior women and minoritiesdid not relate to higher group valence or consensus This increase did littleto improve group climate and may have even hurt the climate in theshort run

HRM changestrategy

347

5 Implications for future research and practice51 Incremental structural change not sufregcient to improve climateWe hope this study will encourage researchers and practitioners interested indiversity change strategies to augment individual and organizational levels ofintervention to focus on work group climates supporting change Our resultshighlight some disparities between the underlying assumptions of HRstrategies to manage organizational change (increasing structural diversitywill be positive for climate) and reality at the department level (incrementalincreases did not improve climate) Organizational strategists implementingdiversity recruitment initiatives typically assume that change strategies (in thiscase hiring diversity) would improve the group social climate due to a positivechange in member attitudes based on increased group diversity and positivesocial contact Yet our study does not support this view Although theorganization we studied had had signiregcant success in enacting the formal HRstrategy plusmn increasing the representation of women and minorities at theorganizational level plusmn our research shows that gradually shifting thedemography of work groups may not necessarily lead to increased consensusor valence about the social climate within work groups For example it did notlead to improved climate regarding the perceived fairness of resourceallocation or good mixing in social interaction In fact the incremental increasein the proportion of minority groups lowered consensus and agreement in theshort term as subgroups sorted out shifting and socially ambiguous powerrelationships within work units These results suggest the importance ofmeasuring change initiatives impacts at the group level and having group levelinterventions such as team-building and group-focused resource allocation tosupport organizational level diversity strategies

52 Focus change strategies on tipping points within targeted work unitsFuture research should consider how HR strategies to manage organizationaldemographic change are socially constructed and enacted at the work grouplevel It is especially important to identify what constitutes a group leveltipping point or critical mass across hierarchical levels and differentdemographic groups to enable positive consensus supporting changePractice might regnd it more fruitful to focus future HR strategies to increasediversity through hiring to alter tipping points within targeted work groupsThis will dramatically alter the saliency of strategic organizationaldemographic change and intervene to shape the climate of speciregc units Ourstudy suggests that organizations may need demographic earthquakes withinwork groups to effect group consensus and valence on the climate regardingthe HR change (ie managing diversity) Dilution of change at the group levelmay ameliorate intended positive effects unless certain group level tippingpoints such as at least 35 percent of the work group are effected in each unit by

JOCM163

348

the HR strategy After eight years of hiring new faculty in our sample womenhovered at about 15 percent in 89 percent of groups and minorities had amaximum representation of 25 percent The demographic shifts whileincreasing were not dramatic enough to create imbalances that markedlyincreased the salience and power of new demographic subgroups created bythe HR strategies Instead of being balanced many of our groups were stilllargely white male and remain tilted or skewed over time which Kanter (1977)notes perpetuates stereotypical negative dynamics

53 Avoiding backlash weaving tipping points across the hierarchyRecent work at MIT by Bailyn (2000) suggests that backlash and resistance toorganizational change may be occurring among senior white male faculty inresponse to HR diversity strategies Initially women and minorities now enterdepartments with relatively equal resources to white men as supported by theHR system design However by the time women and minorities reach theassociate and full professor levels subtle discrimination processes occur in howHR policies are applied such as lower access to resources promotions andpoor climate Despite the increase in upward mobility the representation ofwomen and minorities at the top is still relatively low Ely (1994) found that ifwomen are under-represented at higher organizational levels relations betweenwomen at lower levels might be of lower quality due to increased competitionand negative gender dynamics in the larger social system If HR changestrategies that alter the distribution of existing resources across organizationaland demographic groups are to be successful they must take into account hownot only tipping points in terms of overall departmental representation butmore importantly how tipping points are distributed in demographies acrossthe hierarchies of work groups or departments

HR diversity change strategies that bring in many new hires of differentbackgrounds may have negative ramiregcations for social functioning in groupsparticularly if new members are not supported by additional HR strategies thatallow them to enter work groups on an equal footing or if their work groups donot have leadership that mirror their demographic identities As Kanter (1983)found identity groups must be equal in their access to power resources ifcontact is to produce improvement in attitudes There must also be time for thenew members to be socialized and integrated into the culture Harrison et al(1998) found that the longer the length of time that members worked togetherthe greater the effects of deep level attitude similarity Organizations areincreasing surface (demographic) and deep level (attitudinal) diversitysimultaneously which is likely to have negative short-term ramiregcations forthe consensus and direction of climate We argue HR strategies must not onlyfocus on the surface level plusmn remacrecting structural and demographic attributes plusmnbut also simultaneously to deep level characteristics remacrecting values within

HRM changestrategy

349

the context of speciregc groups at all organizational levels In sum HR strategistsand change agents should focus on how HR change strategies relate to the jointalignment of group memberrsquos multiple characteristics (eg demographichierarchical status task) in work groups plusmn that is to examine group contextualinmacruences (Wharton 1992)

54 LimitationsOur study while making contributions to organizational research is notwithout limitations Though large in individual respondents by utilizing thegroup as the level of analysis our sample size includes only 74 groups Thiscould reduce the level of statistical power of our regndings Additionally becausethe sample was based on one organization it is important to replicate thisanalysis across multiple organizations to substantiate our results Althoughchanges in department heterogeneity across an eight-year period wereincluded this is not a longitudinal design Future research should employrepeated measures over a length of time across all variables measuringdemographic change in heterogeneity as well as consensus and valence Ourstudy points to some of the challenges in measuring ordfsuccess in increasingdiversityordm as we did not look at annual turnover or the effects of the glassceiling at the group level in this study which is a gap we hope futureresearchers will be motivated to study spurred by our research

Since little published work has been done on group consensus regarding theeffects of HR policies to increase diversity through demographic change weextend many social science literatures ranging from organizationaldevelopment to demography and HR Future research on diversity andorganizational change also should strive to be interdisciplinary as theliterature on HR strategy and change has been largely developed in the HRliterature separate from other social science disciplines reducing theoreticalintegration Without such integration and increased scholarly and practitionerattention to climate and consensus many organizations may fail to supportgroup membersrsquo explicit use of their cultural experiences to advanceorganizational learning from enacted diversity change strategies despite thegroup and intergroup social tensions that will inevitably occur (Thomas andEly 1996)

References

Aiken LA and West SG (1993) Multiple Regression Testing and Interpreting InteractionsSage Publications Newbury Park CA

Alderfer CP and Smith KK (1982) ordfStudying embedded intergroup relations inorganizationsordm Administrative Science Quarterly Vol 27 pp 35-65

Allport G (1954) The Nature of Prejudice Addison Wesley Cambridge MA

Bailyn L (2000) ordfGender and diversity in organizations preconference presentationordm paperpresented at the National Academy of Management meeting Toronto

JOCM163

350

Blalock HM (1967) Toward a Theory of Minority Group Relations Wiley New York NY

Bleise PD and Halverson RR (1998) ordfGroup consensus and psychological well-being a largeregeld studyordm Journal of Applied Social Psychology Vol 28 pp 563-80

Brown KG Kozlowski SWJ and Hattrup K (1996) Theory Issues and Recommendations inConceptualizing Agreement as a Construct in Organizational Research The Search forConsensus Regarding Consensus National Academy of Management Cincinnati OH

Cox T Jr (1993) Cultural Diversity in Organizations Berrett-Koehler San Francisco CA

Ely R (1994) ordfThe effects of organizational demographics and social identity on relationshipsamong professional womenordm Administrative Science Quarterly Vol 39 pp 203-38

Ely R (1995) ordfThe power in demography womenrsquos social constructions of gender identity atworkordm Academy of Management Journal Vol 38 pp 589-634

Festinger L (1954) ordfA theory of social comparison processesordm Human Relations Vol 7pp 117-40

Forsyth DR (1990) Group Dynamics BrooksCole Publishing Paciregc Grove CA

Harrison DA Price KH and Bell MP (1998) ordfBeyond relational demography time and theeffects of surface and deep-level diversity on work group cohesionordm Academy ofManagement Journal Vol 41 pp 96-107

James K Lovato C and Khoo G (1994) ordfSocial identity correlates of minority workersrsquo healthordmAcademy of Management Journal Vol 37 p 383+

Kanter R (1977) ordfSome effects of proportions on group life skewed sex ratios and reactions totoken womenordm American Journal of Sociology Vol 82 pp 965-90

Kanter R (1983) Men and Women of the Corporation Basic Books New York NY

Kossek EE and Lobel SA (1996) ordfIntroduction transforming human resource systems tomanage diversityordm in Managing Diversity Human Resource Strategies for Transforming

the Workplace Blackwell Oxford pp 1-19

Kossek EE and Zonia SC (1993) ordfAssessing diversity climate a regeld study of reactions toemployer efforts to promote diversityordm Journal of Organizational Behavior Vol 14pp 61-81

Larkey LK (1996) ordfToward a theory of communicative interactions in culturally diverseworkgroupsordm Academy of Management Review Vol 21 pp 463-91

Lau DC and Murnighan JK (1998) ordfDemographic diversity and faultlines the compositionaldynamics of organizational groupsordm Academy of Management Review Vol 29 pp 325-40

Lindell MK and Brandt CJ (2000) ordfClimate quality and climate consensus as mediators of therelationship between organizational antecedents and outcomesordm Journal of AppliedPsychology Vol 85 pp 331-48

McGrath JE Berdahl JL and Arrow H (1995) ordfTraits expectations culture and clout thedynamics of diversity in work groupsordm in Jackson SE and Ruderman MN (Eds)Diversity in Work Teams Research Paradigms for a Changing Workplace AmericanPsychological Association Washington DC

Payne RL and Pugh DS (1976) ordfOrganizational structure and climateordm in Dunnette MD(Ed) Handbook of Industrial and Organizational Psychology Rand McNally Chicago ILpp 1125-73

Reichers A and Schneider B (1990) ordfClimate and culture an evolution of constructsordm inSchneider B (Ed) Organizational Climate and Culture Jossey-Bass San Francisco CA

HRM changestrategy

351

Org

aniz

atio

nD

epar

tmen

tD

iver

sity

asor

gan

izat

ional

goa

l

Com

mit

men

tof

man

agem

ent

todiv

ersi

ty

Per

ceiv

edw

ork-g

roup

mix

Equal

ity

ofdep

artm

ent

suppor

tof

wom

enE

qual

ity

ofdep

artm

ent

suppor

tof

min

orit

ies

Ste

p1

Dep

tsi

zeT

ime

2-T

ime

10

080

152

006

021

0

05D

R2

001

002

000

004

0

00T

otal

R2

001

002

000

004

0

00Ste

p2

Wom

enT

ime

2-T

ime

10

40

20

38

20

21

20

062

013

Whit

esT

ime

2-T

ime

12

007

019

008

20

172

001

Ten

ure

Tim

e2-

Tim

e1

20

072

013

012

016

019

DR

20

16

017

0

060

070

05T

otal

R2

017

0

19

006

011

0

05Ste

p3

Whit

ete

nure

dm

ales

Tim

e2-

Tim

e1

005

011

000

007

035

D

R2

000

001

000

000

004

T

otal

R2

017

0

20

006

011

009

Note

p

005

orgre

ater

p0

01or

gre

ater

Table IIIRegression

predicting valenceof climate for

diversity

HRM changestrategy

345

in tenured white males was entered in step 3 The results are that the thirdblock was not statistically signiregcant in the regressions showing that smallincreases in the group proportions of senior women and minorities did notincrease group consensus or valence regarding climate for change

4 DiscussionThis study explored links between actual change in organizational structureorganizational demography and work group interpretation of this change Ourresults showed that HR strategies that focus on structural change withoutworking to develop supportive group norms and positive climate areinadequate change strategies Current employer efforts to increase diversity asan isolated HR strategy may not necessarily lead to increased memberagreement that change is favorable Most change efforts like the diversitychange effort we studied are focused at the organizational or individualintervention levels and under-emphasize the work group level and theimportance of group tipping points Our study highlights a disconnect betweenorganizational level diversity practice which often assumes that hiring moreminorities and women will improve climate and what our research found at thegroup level a climate that was mixed and ambivalent at best and sometimesnegative

41 Group level measures of demographic change consensus and valence ondiversity climateThe regrst objective of this study was to develop new theory and measures ofgroup level change in diversity We provided new ways to measure and groupsocial construction of change These included consensus the degree to whichmembers held common perceptions of diversity climate and valence thedirection of those perceptions Our measures identify group and organizationallevel referents of HR strategies to increase demographic diversity Group levelreferents of diversity change were how well members perceived their workgroup mix and the perceived equality of resource allocation acrossdemographic groups Organizational referents were managementcommitment to change and whether diversity should be an organizationalstrategy assessed general support for change Our study highlights howimportant it is for research and practice to not only examine how change indemography relates to differences in individual attitudes but also to groupmember shared perceptions about the direction of the climate for changeUnderstanding linkages between HR changes (in this case group demographywhich raises unique psychological processes related to social identity) andpositive consensus is important because it is relevant to cooperation workexperiences interpersonal conmacrict stress and norms in groups striving tomanage change

JOCM163

346

42 Greater work group demographic shift does not necessarily result infavorable climateOur second objective was to assess whether work groups with higherdemographic shifts tend to have more positive climates that is higherconsensus and valence toward diversity change for both organizational andgroup level change referents Our results only partially supported this beliefthat greater structural demographic shift would positively relate to a morefavorable climate for diversity Although at the organizational level theuniversity had experienced signiregcant change in organizational demographyan increase in the overall representation of white women (36 percent) andminorities (41 percent) over time work group members in units with thegreatest change did not have higher consensus or more favorable valenceWe found more favorable results for organizational than group levelreferents of change Members may regnd it easier to agree with organizationalreferents and more difregcult to agree on group level referents due to theirgreater psychological proximity to group processes that affect how theypersonally experience conmacrict and wellbeing from change in their dailywork Work groups with more demographic change had higher valence butlower consensus on whether the organization should pursue diversity as astrategic goal When organizational change strategies to increase diversityare enacted at the work group level the group demographic shift may bediluted across the organization to only slightly increase the representation ofwomen and minorities within actual groups Due to this dilution effectgroup membersrsquo social construction of the HR strategies may lack consensusor positive valence regarding the organizational diversity change strategiesIf the change is not large enough to create a critical mass to tiprepresentation of minorities to at least 35 percent according to Kanter (1977)negative intergroup dynamics such as increased resource competition andtokenism may still occur

43 Diluting diversity gradually increasing the proportion of senior womenminorities does not necessarily improve group climateOur third objective was to investigate whether increasing the proportion ofsenior women and minorities at the department level positively relates togroup valence and consensus There was no relation between increasing theproportion of senior women and minorities and group valence andconsensus Due to diluted saliency of organizational HR change strategies toincrease diversity at the work group level slight demographic shifts toproportionally increase the representation of senior women and minoritiesdid not relate to higher group valence or consensus This increase did littleto improve group climate and may have even hurt the climate in theshort run

HRM changestrategy

347

5 Implications for future research and practice51 Incremental structural change not sufregcient to improve climateWe hope this study will encourage researchers and practitioners interested indiversity change strategies to augment individual and organizational levels ofintervention to focus on work group climates supporting change Our resultshighlight some disparities between the underlying assumptions of HRstrategies to manage organizational change (increasing structural diversitywill be positive for climate) and reality at the department level (incrementalincreases did not improve climate) Organizational strategists implementingdiversity recruitment initiatives typically assume that change strategies (in thiscase hiring diversity) would improve the group social climate due to a positivechange in member attitudes based on increased group diversity and positivesocial contact Yet our study does not support this view Although theorganization we studied had had signiregcant success in enacting the formal HRstrategy plusmn increasing the representation of women and minorities at theorganizational level plusmn our research shows that gradually shifting thedemography of work groups may not necessarily lead to increased consensusor valence about the social climate within work groups For example it did notlead to improved climate regarding the perceived fairness of resourceallocation or good mixing in social interaction In fact the incremental increasein the proportion of minority groups lowered consensus and agreement in theshort term as subgroups sorted out shifting and socially ambiguous powerrelationships within work units These results suggest the importance ofmeasuring change initiatives impacts at the group level and having group levelinterventions such as team-building and group-focused resource allocation tosupport organizational level diversity strategies

52 Focus change strategies on tipping points within targeted work unitsFuture research should consider how HR strategies to manage organizationaldemographic change are socially constructed and enacted at the work grouplevel It is especially important to identify what constitutes a group leveltipping point or critical mass across hierarchical levels and differentdemographic groups to enable positive consensus supporting changePractice might regnd it more fruitful to focus future HR strategies to increasediversity through hiring to alter tipping points within targeted work groupsThis will dramatically alter the saliency of strategic organizationaldemographic change and intervene to shape the climate of speciregc units Ourstudy suggests that organizations may need demographic earthquakes withinwork groups to effect group consensus and valence on the climate regardingthe HR change (ie managing diversity) Dilution of change at the group levelmay ameliorate intended positive effects unless certain group level tippingpoints such as at least 35 percent of the work group are effected in each unit by

JOCM163

348

the HR strategy After eight years of hiring new faculty in our sample womenhovered at about 15 percent in 89 percent of groups and minorities had amaximum representation of 25 percent The demographic shifts whileincreasing were not dramatic enough to create imbalances that markedlyincreased the salience and power of new demographic subgroups created bythe HR strategies Instead of being balanced many of our groups were stilllargely white male and remain tilted or skewed over time which Kanter (1977)notes perpetuates stereotypical negative dynamics

53 Avoiding backlash weaving tipping points across the hierarchyRecent work at MIT by Bailyn (2000) suggests that backlash and resistance toorganizational change may be occurring among senior white male faculty inresponse to HR diversity strategies Initially women and minorities now enterdepartments with relatively equal resources to white men as supported by theHR system design However by the time women and minorities reach theassociate and full professor levels subtle discrimination processes occur in howHR policies are applied such as lower access to resources promotions andpoor climate Despite the increase in upward mobility the representation ofwomen and minorities at the top is still relatively low Ely (1994) found that ifwomen are under-represented at higher organizational levels relations betweenwomen at lower levels might be of lower quality due to increased competitionand negative gender dynamics in the larger social system If HR changestrategies that alter the distribution of existing resources across organizationaland demographic groups are to be successful they must take into account hownot only tipping points in terms of overall departmental representation butmore importantly how tipping points are distributed in demographies acrossthe hierarchies of work groups or departments

HR diversity change strategies that bring in many new hires of differentbackgrounds may have negative ramiregcations for social functioning in groupsparticularly if new members are not supported by additional HR strategies thatallow them to enter work groups on an equal footing or if their work groups donot have leadership that mirror their demographic identities As Kanter (1983)found identity groups must be equal in their access to power resources ifcontact is to produce improvement in attitudes There must also be time for thenew members to be socialized and integrated into the culture Harrison et al(1998) found that the longer the length of time that members worked togetherthe greater the effects of deep level attitude similarity Organizations areincreasing surface (demographic) and deep level (attitudinal) diversitysimultaneously which is likely to have negative short-term ramiregcations forthe consensus and direction of climate We argue HR strategies must not onlyfocus on the surface level plusmn remacrecting structural and demographic attributes plusmnbut also simultaneously to deep level characteristics remacrecting values within

HRM changestrategy

349

the context of speciregc groups at all organizational levels In sum HR strategistsand change agents should focus on how HR change strategies relate to the jointalignment of group memberrsquos multiple characteristics (eg demographichierarchical status task) in work groups plusmn that is to examine group contextualinmacruences (Wharton 1992)

54 LimitationsOur study while making contributions to organizational research is notwithout limitations Though large in individual respondents by utilizing thegroup as the level of analysis our sample size includes only 74 groups Thiscould reduce the level of statistical power of our regndings Additionally becausethe sample was based on one organization it is important to replicate thisanalysis across multiple organizations to substantiate our results Althoughchanges in department heterogeneity across an eight-year period wereincluded this is not a longitudinal design Future research should employrepeated measures over a length of time across all variables measuringdemographic change in heterogeneity as well as consensus and valence Ourstudy points to some of the challenges in measuring ordfsuccess in increasingdiversityordm as we did not look at annual turnover or the effects of the glassceiling at the group level in this study which is a gap we hope futureresearchers will be motivated to study spurred by our research

Since little published work has been done on group consensus regarding theeffects of HR policies to increase diversity through demographic change weextend many social science literatures ranging from organizationaldevelopment to demography and HR Future research on diversity andorganizational change also should strive to be interdisciplinary as theliterature on HR strategy and change has been largely developed in the HRliterature separate from other social science disciplines reducing theoreticalintegration Without such integration and increased scholarly and practitionerattention to climate and consensus many organizations may fail to supportgroup membersrsquo explicit use of their cultural experiences to advanceorganizational learning from enacted diversity change strategies despite thegroup and intergroup social tensions that will inevitably occur (Thomas andEly 1996)

References

Aiken LA and West SG (1993) Multiple Regression Testing and Interpreting InteractionsSage Publications Newbury Park CA

Alderfer CP and Smith KK (1982) ordfStudying embedded intergroup relations inorganizationsordm Administrative Science Quarterly Vol 27 pp 35-65

Allport G (1954) The Nature of Prejudice Addison Wesley Cambridge MA

Bailyn L (2000) ordfGender and diversity in organizations preconference presentationordm paperpresented at the National Academy of Management meeting Toronto

JOCM163

350

Blalock HM (1967) Toward a Theory of Minority Group Relations Wiley New York NY

Bleise PD and Halverson RR (1998) ordfGroup consensus and psychological well-being a largeregeld studyordm Journal of Applied Social Psychology Vol 28 pp 563-80

Brown KG Kozlowski SWJ and Hattrup K (1996) Theory Issues and Recommendations inConceptualizing Agreement as a Construct in Organizational Research The Search forConsensus Regarding Consensus National Academy of Management Cincinnati OH

Cox T Jr (1993) Cultural Diversity in Organizations Berrett-Koehler San Francisco CA

Ely R (1994) ordfThe effects of organizational demographics and social identity on relationshipsamong professional womenordm Administrative Science Quarterly Vol 39 pp 203-38

Ely R (1995) ordfThe power in demography womenrsquos social constructions of gender identity atworkordm Academy of Management Journal Vol 38 pp 589-634

Festinger L (1954) ordfA theory of social comparison processesordm Human Relations Vol 7pp 117-40

Forsyth DR (1990) Group Dynamics BrooksCole Publishing Paciregc Grove CA

Harrison DA Price KH and Bell MP (1998) ordfBeyond relational demography time and theeffects of surface and deep-level diversity on work group cohesionordm Academy ofManagement Journal Vol 41 pp 96-107

James K Lovato C and Khoo G (1994) ordfSocial identity correlates of minority workersrsquo healthordmAcademy of Management Journal Vol 37 p 383+

Kanter R (1977) ordfSome effects of proportions on group life skewed sex ratios and reactions totoken womenordm American Journal of Sociology Vol 82 pp 965-90

Kanter R (1983) Men and Women of the Corporation Basic Books New York NY

Kossek EE and Lobel SA (1996) ordfIntroduction transforming human resource systems tomanage diversityordm in Managing Diversity Human Resource Strategies for Transforming

the Workplace Blackwell Oxford pp 1-19

Kossek EE and Zonia SC (1993) ordfAssessing diversity climate a regeld study of reactions toemployer efforts to promote diversityordm Journal of Organizational Behavior Vol 14pp 61-81

Larkey LK (1996) ordfToward a theory of communicative interactions in culturally diverseworkgroupsordm Academy of Management Review Vol 21 pp 463-91

Lau DC and Murnighan JK (1998) ordfDemographic diversity and faultlines the compositionaldynamics of organizational groupsordm Academy of Management Review Vol 29 pp 325-40

Lindell MK and Brandt CJ (2000) ordfClimate quality and climate consensus as mediators of therelationship between organizational antecedents and outcomesordm Journal of AppliedPsychology Vol 85 pp 331-48

McGrath JE Berdahl JL and Arrow H (1995) ordfTraits expectations culture and clout thedynamics of diversity in work groupsordm in Jackson SE and Ruderman MN (Eds)Diversity in Work Teams Research Paradigms for a Changing Workplace AmericanPsychological Association Washington DC

Payne RL and Pugh DS (1976) ordfOrganizational structure and climateordm in Dunnette MD(Ed) Handbook of Industrial and Organizational Psychology Rand McNally Chicago ILpp 1125-73

Reichers A and Schneider B (1990) ordfClimate and culture an evolution of constructsordm inSchneider B (Ed) Organizational Climate and Culture Jossey-Bass San Francisco CA

HRM changestrategy

351

in tenured white males was entered in step 3 The results are that the thirdblock was not statistically signiregcant in the regressions showing that smallincreases in the group proportions of senior women and minorities did notincrease group consensus or valence regarding climate for change

4 DiscussionThis study explored links between actual change in organizational structureorganizational demography and work group interpretation of this change Ourresults showed that HR strategies that focus on structural change withoutworking to develop supportive group norms and positive climate areinadequate change strategies Current employer efforts to increase diversity asan isolated HR strategy may not necessarily lead to increased memberagreement that change is favorable Most change efforts like the diversitychange effort we studied are focused at the organizational or individualintervention levels and under-emphasize the work group level and theimportance of group tipping points Our study highlights a disconnect betweenorganizational level diversity practice which often assumes that hiring moreminorities and women will improve climate and what our research found at thegroup level a climate that was mixed and ambivalent at best and sometimesnegative

41 Group level measures of demographic change consensus and valence ondiversity climateThe regrst objective of this study was to develop new theory and measures ofgroup level change in diversity We provided new ways to measure and groupsocial construction of change These included consensus the degree to whichmembers held common perceptions of diversity climate and valence thedirection of those perceptions Our measures identify group and organizationallevel referents of HR strategies to increase demographic diversity Group levelreferents of diversity change were how well members perceived their workgroup mix and the perceived equality of resource allocation acrossdemographic groups Organizational referents were managementcommitment to change and whether diversity should be an organizationalstrategy assessed general support for change Our study highlights howimportant it is for research and practice to not only examine how change indemography relates to differences in individual attitudes but also to groupmember shared perceptions about the direction of the climate for changeUnderstanding linkages between HR changes (in this case group demographywhich raises unique psychological processes related to social identity) andpositive consensus is important because it is relevant to cooperation workexperiences interpersonal conmacrict stress and norms in groups striving tomanage change

JOCM163

346

42 Greater work group demographic shift does not necessarily result infavorable climateOur second objective was to assess whether work groups with higherdemographic shifts tend to have more positive climates that is higherconsensus and valence toward diversity change for both organizational andgroup level change referents Our results only partially supported this beliefthat greater structural demographic shift would positively relate to a morefavorable climate for diversity Although at the organizational level theuniversity had experienced signiregcant change in organizational demographyan increase in the overall representation of white women (36 percent) andminorities (41 percent) over time work group members in units with thegreatest change did not have higher consensus or more favorable valenceWe found more favorable results for organizational than group levelreferents of change Members may regnd it easier to agree with organizationalreferents and more difregcult to agree on group level referents due to theirgreater psychological proximity to group processes that affect how theypersonally experience conmacrict and wellbeing from change in their dailywork Work groups with more demographic change had higher valence butlower consensus on whether the organization should pursue diversity as astrategic goal When organizational change strategies to increase diversityare enacted at the work group level the group demographic shift may bediluted across the organization to only slightly increase the representation ofwomen and minorities within actual groups Due to this dilution effectgroup membersrsquo social construction of the HR strategies may lack consensusor positive valence regarding the organizational diversity change strategiesIf the change is not large enough to create a critical mass to tiprepresentation of minorities to at least 35 percent according to Kanter (1977)negative intergroup dynamics such as increased resource competition andtokenism may still occur

43 Diluting diversity gradually increasing the proportion of senior womenminorities does not necessarily improve group climateOur third objective was to investigate whether increasing the proportion ofsenior women and minorities at the department level positively relates togroup valence and consensus There was no relation between increasing theproportion of senior women and minorities and group valence andconsensus Due to diluted saliency of organizational HR change strategies toincrease diversity at the work group level slight demographic shifts toproportionally increase the representation of senior women and minoritiesdid not relate to higher group valence or consensus This increase did littleto improve group climate and may have even hurt the climate in theshort run

HRM changestrategy

347

5 Implications for future research and practice51 Incremental structural change not sufregcient to improve climateWe hope this study will encourage researchers and practitioners interested indiversity change strategies to augment individual and organizational levels ofintervention to focus on work group climates supporting change Our resultshighlight some disparities between the underlying assumptions of HRstrategies to manage organizational change (increasing structural diversitywill be positive for climate) and reality at the department level (incrementalincreases did not improve climate) Organizational strategists implementingdiversity recruitment initiatives typically assume that change strategies (in thiscase hiring diversity) would improve the group social climate due to a positivechange in member attitudes based on increased group diversity and positivesocial contact Yet our study does not support this view Although theorganization we studied had had signiregcant success in enacting the formal HRstrategy plusmn increasing the representation of women and minorities at theorganizational level plusmn our research shows that gradually shifting thedemography of work groups may not necessarily lead to increased consensusor valence about the social climate within work groups For example it did notlead to improved climate regarding the perceived fairness of resourceallocation or good mixing in social interaction In fact the incremental increasein the proportion of minority groups lowered consensus and agreement in theshort term as subgroups sorted out shifting and socially ambiguous powerrelationships within work units These results suggest the importance ofmeasuring change initiatives impacts at the group level and having group levelinterventions such as team-building and group-focused resource allocation tosupport organizational level diversity strategies

52 Focus change strategies on tipping points within targeted work unitsFuture research should consider how HR strategies to manage organizationaldemographic change are socially constructed and enacted at the work grouplevel It is especially important to identify what constitutes a group leveltipping point or critical mass across hierarchical levels and differentdemographic groups to enable positive consensus supporting changePractice might regnd it more fruitful to focus future HR strategies to increasediversity through hiring to alter tipping points within targeted work groupsThis will dramatically alter the saliency of strategic organizationaldemographic change and intervene to shape the climate of speciregc units Ourstudy suggests that organizations may need demographic earthquakes withinwork groups to effect group consensus and valence on the climate regardingthe HR change (ie managing diversity) Dilution of change at the group levelmay ameliorate intended positive effects unless certain group level tippingpoints such as at least 35 percent of the work group are effected in each unit by

JOCM163

348

the HR strategy After eight years of hiring new faculty in our sample womenhovered at about 15 percent in 89 percent of groups and minorities had amaximum representation of 25 percent The demographic shifts whileincreasing were not dramatic enough to create imbalances that markedlyincreased the salience and power of new demographic subgroups created bythe HR strategies Instead of being balanced many of our groups were stilllargely white male and remain tilted or skewed over time which Kanter (1977)notes perpetuates stereotypical negative dynamics

53 Avoiding backlash weaving tipping points across the hierarchyRecent work at MIT by Bailyn (2000) suggests that backlash and resistance toorganizational change may be occurring among senior white male faculty inresponse to HR diversity strategies Initially women and minorities now enterdepartments with relatively equal resources to white men as supported by theHR system design However by the time women and minorities reach theassociate and full professor levels subtle discrimination processes occur in howHR policies are applied such as lower access to resources promotions andpoor climate Despite the increase in upward mobility the representation ofwomen and minorities at the top is still relatively low Ely (1994) found that ifwomen are under-represented at higher organizational levels relations betweenwomen at lower levels might be of lower quality due to increased competitionand negative gender dynamics in the larger social system If HR changestrategies that alter the distribution of existing resources across organizationaland demographic groups are to be successful they must take into account hownot only tipping points in terms of overall departmental representation butmore importantly how tipping points are distributed in demographies acrossthe hierarchies of work groups or departments

HR diversity change strategies that bring in many new hires of differentbackgrounds may have negative ramiregcations for social functioning in groupsparticularly if new members are not supported by additional HR strategies thatallow them to enter work groups on an equal footing or if their work groups donot have leadership that mirror their demographic identities As Kanter (1983)found identity groups must be equal in their access to power resources ifcontact is to produce improvement in attitudes There must also be time for thenew members to be socialized and integrated into the culture Harrison et al(1998) found that the longer the length of time that members worked togetherthe greater the effects of deep level attitude similarity Organizations areincreasing surface (demographic) and deep level (attitudinal) diversitysimultaneously which is likely to have negative short-term ramiregcations forthe consensus and direction of climate We argue HR strategies must not onlyfocus on the surface level plusmn remacrecting structural and demographic attributes plusmnbut also simultaneously to deep level characteristics remacrecting values within

HRM changestrategy

349

the context of speciregc groups at all organizational levels In sum HR strategistsand change agents should focus on how HR change strategies relate to the jointalignment of group memberrsquos multiple characteristics (eg demographichierarchical status task) in work groups plusmn that is to examine group contextualinmacruences (Wharton 1992)

54 LimitationsOur study while making contributions to organizational research is notwithout limitations Though large in individual respondents by utilizing thegroup as the level of analysis our sample size includes only 74 groups Thiscould reduce the level of statistical power of our regndings Additionally becausethe sample was based on one organization it is important to replicate thisanalysis across multiple organizations to substantiate our results Althoughchanges in department heterogeneity across an eight-year period wereincluded this is not a longitudinal design Future research should employrepeated measures over a length of time across all variables measuringdemographic change in heterogeneity as well as consensus and valence Ourstudy points to some of the challenges in measuring ordfsuccess in increasingdiversityordm as we did not look at annual turnover or the effects of the glassceiling at the group level in this study which is a gap we hope futureresearchers will be motivated to study spurred by our research

Since little published work has been done on group consensus regarding theeffects of HR policies to increase diversity through demographic change weextend many social science literatures ranging from organizationaldevelopment to demography and HR Future research on diversity andorganizational change also should strive to be interdisciplinary as theliterature on HR strategy and change has been largely developed in the HRliterature separate from other social science disciplines reducing theoreticalintegration Without such integration and increased scholarly and practitionerattention to climate and consensus many organizations may fail to supportgroup membersrsquo explicit use of their cultural experiences to advanceorganizational learning from enacted diversity change strategies despite thegroup and intergroup social tensions that will inevitably occur (Thomas andEly 1996)

References

Aiken LA and West SG (1993) Multiple Regression Testing and Interpreting InteractionsSage Publications Newbury Park CA

Alderfer CP and Smith KK (1982) ordfStudying embedded intergroup relations inorganizationsordm Administrative Science Quarterly Vol 27 pp 35-65

Allport G (1954) The Nature of Prejudice Addison Wesley Cambridge MA

Bailyn L (2000) ordfGender and diversity in organizations preconference presentationordm paperpresented at the National Academy of Management meeting Toronto

JOCM163

350

Blalock HM (1967) Toward a Theory of Minority Group Relations Wiley New York NY

Bleise PD and Halverson RR (1998) ordfGroup consensus and psychological well-being a largeregeld studyordm Journal of Applied Social Psychology Vol 28 pp 563-80

Brown KG Kozlowski SWJ and Hattrup K (1996) Theory Issues and Recommendations inConceptualizing Agreement as a Construct in Organizational Research The Search forConsensus Regarding Consensus National Academy of Management Cincinnati OH

Cox T Jr (1993) Cultural Diversity in Organizations Berrett-Koehler San Francisco CA

Ely R (1994) ordfThe effects of organizational demographics and social identity on relationshipsamong professional womenordm Administrative Science Quarterly Vol 39 pp 203-38

Ely R (1995) ordfThe power in demography womenrsquos social constructions of gender identity atworkordm Academy of Management Journal Vol 38 pp 589-634

Festinger L (1954) ordfA theory of social comparison processesordm Human Relations Vol 7pp 117-40

Forsyth DR (1990) Group Dynamics BrooksCole Publishing Paciregc Grove CA

Harrison DA Price KH and Bell MP (1998) ordfBeyond relational demography time and theeffects of surface and deep-level diversity on work group cohesionordm Academy ofManagement Journal Vol 41 pp 96-107

James K Lovato C and Khoo G (1994) ordfSocial identity correlates of minority workersrsquo healthordmAcademy of Management Journal Vol 37 p 383+

Kanter R (1977) ordfSome effects of proportions on group life skewed sex ratios and reactions totoken womenordm American Journal of Sociology Vol 82 pp 965-90

Kanter R (1983) Men and Women of the Corporation Basic Books New York NY

Kossek EE and Lobel SA (1996) ordfIntroduction transforming human resource systems tomanage diversityordm in Managing Diversity Human Resource Strategies for Transforming

the Workplace Blackwell Oxford pp 1-19

Kossek EE and Zonia SC (1993) ordfAssessing diversity climate a regeld study of reactions toemployer efforts to promote diversityordm Journal of Organizational Behavior Vol 14pp 61-81

Larkey LK (1996) ordfToward a theory of communicative interactions in culturally diverseworkgroupsordm Academy of Management Review Vol 21 pp 463-91

Lau DC and Murnighan JK (1998) ordfDemographic diversity and faultlines the compositionaldynamics of organizational groupsordm Academy of Management Review Vol 29 pp 325-40

Lindell MK and Brandt CJ (2000) ordfClimate quality and climate consensus as mediators of therelationship between organizational antecedents and outcomesordm Journal of AppliedPsychology Vol 85 pp 331-48

McGrath JE Berdahl JL and Arrow H (1995) ordfTraits expectations culture and clout thedynamics of diversity in work groupsordm in Jackson SE and Ruderman MN (Eds)Diversity in Work Teams Research Paradigms for a Changing Workplace AmericanPsychological Association Washington DC

Payne RL and Pugh DS (1976) ordfOrganizational structure and climateordm in Dunnette MD(Ed) Handbook of Industrial and Organizational Psychology Rand McNally Chicago ILpp 1125-73

Reichers A and Schneider B (1990) ordfClimate and culture an evolution of constructsordm inSchneider B (Ed) Organizational Climate and Culture Jossey-Bass San Francisco CA

HRM changestrategy

351

42 Greater work group demographic shift does not necessarily result infavorable climateOur second objective was to assess whether work groups with higherdemographic shifts tend to have more positive climates that is higherconsensus and valence toward diversity change for both organizational andgroup level change referents Our results only partially supported this beliefthat greater structural demographic shift would positively relate to a morefavorable climate for diversity Although at the organizational level theuniversity had experienced signiregcant change in organizational demographyan increase in the overall representation of white women (36 percent) andminorities (41 percent) over time work group members in units with thegreatest change did not have higher consensus or more favorable valenceWe found more favorable results for organizational than group levelreferents of change Members may regnd it easier to agree with organizationalreferents and more difregcult to agree on group level referents due to theirgreater psychological proximity to group processes that affect how theypersonally experience conmacrict and wellbeing from change in their dailywork Work groups with more demographic change had higher valence butlower consensus on whether the organization should pursue diversity as astrategic goal When organizational change strategies to increase diversityare enacted at the work group level the group demographic shift may bediluted across the organization to only slightly increase the representation ofwomen and minorities within actual groups Due to this dilution effectgroup membersrsquo social construction of the HR strategies may lack consensusor positive valence regarding the organizational diversity change strategiesIf the change is not large enough to create a critical mass to tiprepresentation of minorities to at least 35 percent according to Kanter (1977)negative intergroup dynamics such as increased resource competition andtokenism may still occur

43 Diluting diversity gradually increasing the proportion of senior womenminorities does not necessarily improve group climateOur third objective was to investigate whether increasing the proportion ofsenior women and minorities at the department level positively relates togroup valence and consensus There was no relation between increasing theproportion of senior women and minorities and group valence andconsensus Due to diluted saliency of organizational HR change strategies toincrease diversity at the work group level slight demographic shifts toproportionally increase the representation of senior women and minoritiesdid not relate to higher group valence or consensus This increase did littleto improve group climate and may have even hurt the climate in theshort run

HRM changestrategy

347

5 Implications for future research and practice51 Incremental structural change not sufregcient to improve climateWe hope this study will encourage researchers and practitioners interested indiversity change strategies to augment individual and organizational levels ofintervention to focus on work group climates supporting change Our resultshighlight some disparities between the underlying assumptions of HRstrategies to manage organizational change (increasing structural diversitywill be positive for climate) and reality at the department level (incrementalincreases did not improve climate) Organizational strategists implementingdiversity recruitment initiatives typically assume that change strategies (in thiscase hiring diversity) would improve the group social climate due to a positivechange in member attitudes based on increased group diversity and positivesocial contact Yet our study does not support this view Although theorganization we studied had had signiregcant success in enacting the formal HRstrategy plusmn increasing the representation of women and minorities at theorganizational level plusmn our research shows that gradually shifting thedemography of work groups may not necessarily lead to increased consensusor valence about the social climate within work groups For example it did notlead to improved climate regarding the perceived fairness of resourceallocation or good mixing in social interaction In fact the incremental increasein the proportion of minority groups lowered consensus and agreement in theshort term as subgroups sorted out shifting and socially ambiguous powerrelationships within work units These results suggest the importance ofmeasuring change initiatives impacts at the group level and having group levelinterventions such as team-building and group-focused resource allocation tosupport organizational level diversity strategies

52 Focus change strategies on tipping points within targeted work unitsFuture research should consider how HR strategies to manage organizationaldemographic change are socially constructed and enacted at the work grouplevel It is especially important to identify what constitutes a group leveltipping point or critical mass across hierarchical levels and differentdemographic groups to enable positive consensus supporting changePractice might regnd it more fruitful to focus future HR strategies to increasediversity through hiring to alter tipping points within targeted work groupsThis will dramatically alter the saliency of strategic organizationaldemographic change and intervene to shape the climate of speciregc units Ourstudy suggests that organizations may need demographic earthquakes withinwork groups to effect group consensus and valence on the climate regardingthe HR change (ie managing diversity) Dilution of change at the group levelmay ameliorate intended positive effects unless certain group level tippingpoints such as at least 35 percent of the work group are effected in each unit by

JOCM163

348

the HR strategy After eight years of hiring new faculty in our sample womenhovered at about 15 percent in 89 percent of groups and minorities had amaximum representation of 25 percent The demographic shifts whileincreasing were not dramatic enough to create imbalances that markedlyincreased the salience and power of new demographic subgroups created bythe HR strategies Instead of being balanced many of our groups were stilllargely white male and remain tilted or skewed over time which Kanter (1977)notes perpetuates stereotypical negative dynamics

53 Avoiding backlash weaving tipping points across the hierarchyRecent work at MIT by Bailyn (2000) suggests that backlash and resistance toorganizational change may be occurring among senior white male faculty inresponse to HR diversity strategies Initially women and minorities now enterdepartments with relatively equal resources to white men as supported by theHR system design However by the time women and minorities reach theassociate and full professor levels subtle discrimination processes occur in howHR policies are applied such as lower access to resources promotions andpoor climate Despite the increase in upward mobility the representation ofwomen and minorities at the top is still relatively low Ely (1994) found that ifwomen are under-represented at higher organizational levels relations betweenwomen at lower levels might be of lower quality due to increased competitionand negative gender dynamics in the larger social system If HR changestrategies that alter the distribution of existing resources across organizationaland demographic groups are to be successful they must take into account hownot only tipping points in terms of overall departmental representation butmore importantly how tipping points are distributed in demographies acrossthe hierarchies of work groups or departments

HR diversity change strategies that bring in many new hires of differentbackgrounds may have negative ramiregcations for social functioning in groupsparticularly if new members are not supported by additional HR strategies thatallow them to enter work groups on an equal footing or if their work groups donot have leadership that mirror their demographic identities As Kanter (1983)found identity groups must be equal in their access to power resources ifcontact is to produce improvement in attitudes There must also be time for thenew members to be socialized and integrated into the culture Harrison et al(1998) found that the longer the length of time that members worked togetherthe greater the effects of deep level attitude similarity Organizations areincreasing surface (demographic) and deep level (attitudinal) diversitysimultaneously which is likely to have negative short-term ramiregcations forthe consensus and direction of climate We argue HR strategies must not onlyfocus on the surface level plusmn remacrecting structural and demographic attributes plusmnbut also simultaneously to deep level characteristics remacrecting values within

HRM changestrategy

349

the context of speciregc groups at all organizational levels In sum HR strategistsand change agents should focus on how HR change strategies relate to the jointalignment of group memberrsquos multiple characteristics (eg demographichierarchical status task) in work groups plusmn that is to examine group contextualinmacruences (Wharton 1992)

54 LimitationsOur study while making contributions to organizational research is notwithout limitations Though large in individual respondents by utilizing thegroup as the level of analysis our sample size includes only 74 groups Thiscould reduce the level of statistical power of our regndings Additionally becausethe sample was based on one organization it is important to replicate thisanalysis across multiple organizations to substantiate our results Althoughchanges in department heterogeneity across an eight-year period wereincluded this is not a longitudinal design Future research should employrepeated measures over a length of time across all variables measuringdemographic change in heterogeneity as well as consensus and valence Ourstudy points to some of the challenges in measuring ordfsuccess in increasingdiversityordm as we did not look at annual turnover or the effects of the glassceiling at the group level in this study which is a gap we hope futureresearchers will be motivated to study spurred by our research

Since little published work has been done on group consensus regarding theeffects of HR policies to increase diversity through demographic change weextend many social science literatures ranging from organizationaldevelopment to demography and HR Future research on diversity andorganizational change also should strive to be interdisciplinary as theliterature on HR strategy and change has been largely developed in the HRliterature separate from other social science disciplines reducing theoreticalintegration Without such integration and increased scholarly and practitionerattention to climate and consensus many organizations may fail to supportgroup membersrsquo explicit use of their cultural experiences to advanceorganizational learning from enacted diversity change strategies despite thegroup and intergroup social tensions that will inevitably occur (Thomas andEly 1996)

References

Aiken LA and West SG (1993) Multiple Regression Testing and Interpreting InteractionsSage Publications Newbury Park CA

Alderfer CP and Smith KK (1982) ordfStudying embedded intergroup relations inorganizationsordm Administrative Science Quarterly Vol 27 pp 35-65

Allport G (1954) The Nature of Prejudice Addison Wesley Cambridge MA

Bailyn L (2000) ordfGender and diversity in organizations preconference presentationordm paperpresented at the National Academy of Management meeting Toronto

JOCM163

350

Blalock HM (1967) Toward a Theory of Minority Group Relations Wiley New York NY

Bleise PD and Halverson RR (1998) ordfGroup consensus and psychological well-being a largeregeld studyordm Journal of Applied Social Psychology Vol 28 pp 563-80

Brown KG Kozlowski SWJ and Hattrup K (1996) Theory Issues and Recommendations inConceptualizing Agreement as a Construct in Organizational Research The Search forConsensus Regarding Consensus National Academy of Management Cincinnati OH

Cox T Jr (1993) Cultural Diversity in Organizations Berrett-Koehler San Francisco CA

Ely R (1994) ordfThe effects of organizational demographics and social identity on relationshipsamong professional womenordm Administrative Science Quarterly Vol 39 pp 203-38

Ely R (1995) ordfThe power in demography womenrsquos social constructions of gender identity atworkordm Academy of Management Journal Vol 38 pp 589-634

Festinger L (1954) ordfA theory of social comparison processesordm Human Relations Vol 7pp 117-40

Forsyth DR (1990) Group Dynamics BrooksCole Publishing Paciregc Grove CA

Harrison DA Price KH and Bell MP (1998) ordfBeyond relational demography time and theeffects of surface and deep-level diversity on work group cohesionordm Academy ofManagement Journal Vol 41 pp 96-107

James K Lovato C and Khoo G (1994) ordfSocial identity correlates of minority workersrsquo healthordmAcademy of Management Journal Vol 37 p 383+

Kanter R (1977) ordfSome effects of proportions on group life skewed sex ratios and reactions totoken womenordm American Journal of Sociology Vol 82 pp 965-90

Kanter R (1983) Men and Women of the Corporation Basic Books New York NY

Kossek EE and Lobel SA (1996) ordfIntroduction transforming human resource systems tomanage diversityordm in Managing Diversity Human Resource Strategies for Transforming

the Workplace Blackwell Oxford pp 1-19

Kossek EE and Zonia SC (1993) ordfAssessing diversity climate a regeld study of reactions toemployer efforts to promote diversityordm Journal of Organizational Behavior Vol 14pp 61-81

Larkey LK (1996) ordfToward a theory of communicative interactions in culturally diverseworkgroupsordm Academy of Management Review Vol 21 pp 463-91

Lau DC and Murnighan JK (1998) ordfDemographic diversity and faultlines the compositionaldynamics of organizational groupsordm Academy of Management Review Vol 29 pp 325-40

Lindell MK and Brandt CJ (2000) ordfClimate quality and climate consensus as mediators of therelationship between organizational antecedents and outcomesordm Journal of AppliedPsychology Vol 85 pp 331-48

McGrath JE Berdahl JL and Arrow H (1995) ordfTraits expectations culture and clout thedynamics of diversity in work groupsordm in Jackson SE and Ruderman MN (Eds)Diversity in Work Teams Research Paradigms for a Changing Workplace AmericanPsychological Association Washington DC

Payne RL and Pugh DS (1976) ordfOrganizational structure and climateordm in Dunnette MD(Ed) Handbook of Industrial and Organizational Psychology Rand McNally Chicago ILpp 1125-73

Reichers A and Schneider B (1990) ordfClimate and culture an evolution of constructsordm inSchneider B (Ed) Organizational Climate and Culture Jossey-Bass San Francisco CA

HRM changestrategy

351

5 Implications for future research and practice51 Incremental structural change not sufregcient to improve climateWe hope this study will encourage researchers and practitioners interested indiversity change strategies to augment individual and organizational levels ofintervention to focus on work group climates supporting change Our resultshighlight some disparities between the underlying assumptions of HRstrategies to manage organizational change (increasing structural diversitywill be positive for climate) and reality at the department level (incrementalincreases did not improve climate) Organizational strategists implementingdiversity recruitment initiatives typically assume that change strategies (in thiscase hiring diversity) would improve the group social climate due to a positivechange in member attitudes based on increased group diversity and positivesocial contact Yet our study does not support this view Although theorganization we studied had had signiregcant success in enacting the formal HRstrategy plusmn increasing the representation of women and minorities at theorganizational level plusmn our research shows that gradually shifting thedemography of work groups may not necessarily lead to increased consensusor valence about the social climate within work groups For example it did notlead to improved climate regarding the perceived fairness of resourceallocation or good mixing in social interaction In fact the incremental increasein the proportion of minority groups lowered consensus and agreement in theshort term as subgroups sorted out shifting and socially ambiguous powerrelationships within work units These results suggest the importance ofmeasuring change initiatives impacts at the group level and having group levelinterventions such as team-building and group-focused resource allocation tosupport organizational level diversity strategies

52 Focus change strategies on tipping points within targeted work unitsFuture research should consider how HR strategies to manage organizationaldemographic change are socially constructed and enacted at the work grouplevel It is especially important to identify what constitutes a group leveltipping point or critical mass across hierarchical levels and differentdemographic groups to enable positive consensus supporting changePractice might regnd it more fruitful to focus future HR strategies to increasediversity through hiring to alter tipping points within targeted work groupsThis will dramatically alter the saliency of strategic organizationaldemographic change and intervene to shape the climate of speciregc units Ourstudy suggests that organizations may need demographic earthquakes withinwork groups to effect group consensus and valence on the climate regardingthe HR change (ie managing diversity) Dilution of change at the group levelmay ameliorate intended positive effects unless certain group level tippingpoints such as at least 35 percent of the work group are effected in each unit by

JOCM163

348

the HR strategy After eight years of hiring new faculty in our sample womenhovered at about 15 percent in 89 percent of groups and minorities had amaximum representation of 25 percent The demographic shifts whileincreasing were not dramatic enough to create imbalances that markedlyincreased the salience and power of new demographic subgroups created bythe HR strategies Instead of being balanced many of our groups were stilllargely white male and remain tilted or skewed over time which Kanter (1977)notes perpetuates stereotypical negative dynamics

53 Avoiding backlash weaving tipping points across the hierarchyRecent work at MIT by Bailyn (2000) suggests that backlash and resistance toorganizational change may be occurring among senior white male faculty inresponse to HR diversity strategies Initially women and minorities now enterdepartments with relatively equal resources to white men as supported by theHR system design However by the time women and minorities reach theassociate and full professor levels subtle discrimination processes occur in howHR policies are applied such as lower access to resources promotions andpoor climate Despite the increase in upward mobility the representation ofwomen and minorities at the top is still relatively low Ely (1994) found that ifwomen are under-represented at higher organizational levels relations betweenwomen at lower levels might be of lower quality due to increased competitionand negative gender dynamics in the larger social system If HR changestrategies that alter the distribution of existing resources across organizationaland demographic groups are to be successful they must take into account hownot only tipping points in terms of overall departmental representation butmore importantly how tipping points are distributed in demographies acrossthe hierarchies of work groups or departments

HR diversity change strategies that bring in many new hires of differentbackgrounds may have negative ramiregcations for social functioning in groupsparticularly if new members are not supported by additional HR strategies thatallow them to enter work groups on an equal footing or if their work groups donot have leadership that mirror their demographic identities As Kanter (1983)found identity groups must be equal in their access to power resources ifcontact is to produce improvement in attitudes There must also be time for thenew members to be socialized and integrated into the culture Harrison et al(1998) found that the longer the length of time that members worked togetherthe greater the effects of deep level attitude similarity Organizations areincreasing surface (demographic) and deep level (attitudinal) diversitysimultaneously which is likely to have negative short-term ramiregcations forthe consensus and direction of climate We argue HR strategies must not onlyfocus on the surface level plusmn remacrecting structural and demographic attributes plusmnbut also simultaneously to deep level characteristics remacrecting values within

HRM changestrategy

349

the context of speciregc groups at all organizational levels In sum HR strategistsand change agents should focus on how HR change strategies relate to the jointalignment of group memberrsquos multiple characteristics (eg demographichierarchical status task) in work groups plusmn that is to examine group contextualinmacruences (Wharton 1992)

54 LimitationsOur study while making contributions to organizational research is notwithout limitations Though large in individual respondents by utilizing thegroup as the level of analysis our sample size includes only 74 groups Thiscould reduce the level of statistical power of our regndings Additionally becausethe sample was based on one organization it is important to replicate thisanalysis across multiple organizations to substantiate our results Althoughchanges in department heterogeneity across an eight-year period wereincluded this is not a longitudinal design Future research should employrepeated measures over a length of time across all variables measuringdemographic change in heterogeneity as well as consensus and valence Ourstudy points to some of the challenges in measuring ordfsuccess in increasingdiversityordm as we did not look at annual turnover or the effects of the glassceiling at the group level in this study which is a gap we hope futureresearchers will be motivated to study spurred by our research

Since little published work has been done on group consensus regarding theeffects of HR policies to increase diversity through demographic change weextend many social science literatures ranging from organizationaldevelopment to demography and HR Future research on diversity andorganizational change also should strive to be interdisciplinary as theliterature on HR strategy and change has been largely developed in the HRliterature separate from other social science disciplines reducing theoreticalintegration Without such integration and increased scholarly and practitionerattention to climate and consensus many organizations may fail to supportgroup membersrsquo explicit use of their cultural experiences to advanceorganizational learning from enacted diversity change strategies despite thegroup and intergroup social tensions that will inevitably occur (Thomas andEly 1996)

References

Aiken LA and West SG (1993) Multiple Regression Testing and Interpreting InteractionsSage Publications Newbury Park CA

Alderfer CP and Smith KK (1982) ordfStudying embedded intergroup relations inorganizationsordm Administrative Science Quarterly Vol 27 pp 35-65

Allport G (1954) The Nature of Prejudice Addison Wesley Cambridge MA

Bailyn L (2000) ordfGender and diversity in organizations preconference presentationordm paperpresented at the National Academy of Management meeting Toronto

JOCM163

350

Blalock HM (1967) Toward a Theory of Minority Group Relations Wiley New York NY

Bleise PD and Halverson RR (1998) ordfGroup consensus and psychological well-being a largeregeld studyordm Journal of Applied Social Psychology Vol 28 pp 563-80

Brown KG Kozlowski SWJ and Hattrup K (1996) Theory Issues and Recommendations inConceptualizing Agreement as a Construct in Organizational Research The Search forConsensus Regarding Consensus National Academy of Management Cincinnati OH

Cox T Jr (1993) Cultural Diversity in Organizations Berrett-Koehler San Francisco CA

Ely R (1994) ordfThe effects of organizational demographics and social identity on relationshipsamong professional womenordm Administrative Science Quarterly Vol 39 pp 203-38

Ely R (1995) ordfThe power in demography womenrsquos social constructions of gender identity atworkordm Academy of Management Journal Vol 38 pp 589-634

Festinger L (1954) ordfA theory of social comparison processesordm Human Relations Vol 7pp 117-40

Forsyth DR (1990) Group Dynamics BrooksCole Publishing Paciregc Grove CA

Harrison DA Price KH and Bell MP (1998) ordfBeyond relational demography time and theeffects of surface and deep-level diversity on work group cohesionordm Academy ofManagement Journal Vol 41 pp 96-107

James K Lovato C and Khoo G (1994) ordfSocial identity correlates of minority workersrsquo healthordmAcademy of Management Journal Vol 37 p 383+

Kanter R (1977) ordfSome effects of proportions on group life skewed sex ratios and reactions totoken womenordm American Journal of Sociology Vol 82 pp 965-90

Kanter R (1983) Men and Women of the Corporation Basic Books New York NY

Kossek EE and Lobel SA (1996) ordfIntroduction transforming human resource systems tomanage diversityordm in Managing Diversity Human Resource Strategies for Transforming

the Workplace Blackwell Oxford pp 1-19

Kossek EE and Zonia SC (1993) ordfAssessing diversity climate a regeld study of reactions toemployer efforts to promote diversityordm Journal of Organizational Behavior Vol 14pp 61-81

Larkey LK (1996) ordfToward a theory of communicative interactions in culturally diverseworkgroupsordm Academy of Management Review Vol 21 pp 463-91

Lau DC and Murnighan JK (1998) ordfDemographic diversity and faultlines the compositionaldynamics of organizational groupsordm Academy of Management Review Vol 29 pp 325-40

Lindell MK and Brandt CJ (2000) ordfClimate quality and climate consensus as mediators of therelationship between organizational antecedents and outcomesordm Journal of AppliedPsychology Vol 85 pp 331-48

McGrath JE Berdahl JL and Arrow H (1995) ordfTraits expectations culture and clout thedynamics of diversity in work groupsordm in Jackson SE and Ruderman MN (Eds)Diversity in Work Teams Research Paradigms for a Changing Workplace AmericanPsychological Association Washington DC

Payne RL and Pugh DS (1976) ordfOrganizational structure and climateordm in Dunnette MD(Ed) Handbook of Industrial and Organizational Psychology Rand McNally Chicago ILpp 1125-73

Reichers A and Schneider B (1990) ordfClimate and culture an evolution of constructsordm inSchneider B (Ed) Organizational Climate and Culture Jossey-Bass San Francisco CA

HRM changestrategy

351

the HR strategy After eight years of hiring new faculty in our sample womenhovered at about 15 percent in 89 percent of groups and minorities had amaximum representation of 25 percent The demographic shifts whileincreasing were not dramatic enough to create imbalances that markedlyincreased the salience and power of new demographic subgroups created bythe HR strategies Instead of being balanced many of our groups were stilllargely white male and remain tilted or skewed over time which Kanter (1977)notes perpetuates stereotypical negative dynamics

53 Avoiding backlash weaving tipping points across the hierarchyRecent work at MIT by Bailyn (2000) suggests that backlash and resistance toorganizational change may be occurring among senior white male faculty inresponse to HR diversity strategies Initially women and minorities now enterdepartments with relatively equal resources to white men as supported by theHR system design However by the time women and minorities reach theassociate and full professor levels subtle discrimination processes occur in howHR policies are applied such as lower access to resources promotions andpoor climate Despite the increase in upward mobility the representation ofwomen and minorities at the top is still relatively low Ely (1994) found that ifwomen are under-represented at higher organizational levels relations betweenwomen at lower levels might be of lower quality due to increased competitionand negative gender dynamics in the larger social system If HR changestrategies that alter the distribution of existing resources across organizationaland demographic groups are to be successful they must take into account hownot only tipping points in terms of overall departmental representation butmore importantly how tipping points are distributed in demographies acrossthe hierarchies of work groups or departments

HR diversity change strategies that bring in many new hires of differentbackgrounds may have negative ramiregcations for social functioning in groupsparticularly if new members are not supported by additional HR strategies thatallow them to enter work groups on an equal footing or if their work groups donot have leadership that mirror their demographic identities As Kanter (1983)found identity groups must be equal in their access to power resources ifcontact is to produce improvement in attitudes There must also be time for thenew members to be socialized and integrated into the culture Harrison et al(1998) found that the longer the length of time that members worked togetherthe greater the effects of deep level attitude similarity Organizations areincreasing surface (demographic) and deep level (attitudinal) diversitysimultaneously which is likely to have negative short-term ramiregcations forthe consensus and direction of climate We argue HR strategies must not onlyfocus on the surface level plusmn remacrecting structural and demographic attributes plusmnbut also simultaneously to deep level characteristics remacrecting values within

HRM changestrategy

349

the context of speciregc groups at all organizational levels In sum HR strategistsand change agents should focus on how HR change strategies relate to the jointalignment of group memberrsquos multiple characteristics (eg demographichierarchical status task) in work groups plusmn that is to examine group contextualinmacruences (Wharton 1992)

54 LimitationsOur study while making contributions to organizational research is notwithout limitations Though large in individual respondents by utilizing thegroup as the level of analysis our sample size includes only 74 groups Thiscould reduce the level of statistical power of our regndings Additionally becausethe sample was based on one organization it is important to replicate thisanalysis across multiple organizations to substantiate our results Althoughchanges in department heterogeneity across an eight-year period wereincluded this is not a longitudinal design Future research should employrepeated measures over a length of time across all variables measuringdemographic change in heterogeneity as well as consensus and valence Ourstudy points to some of the challenges in measuring ordfsuccess in increasingdiversityordm as we did not look at annual turnover or the effects of the glassceiling at the group level in this study which is a gap we hope futureresearchers will be motivated to study spurred by our research

Since little published work has been done on group consensus regarding theeffects of HR policies to increase diversity through demographic change weextend many social science literatures ranging from organizationaldevelopment to demography and HR Future research on diversity andorganizational change also should strive to be interdisciplinary as theliterature on HR strategy and change has been largely developed in the HRliterature separate from other social science disciplines reducing theoreticalintegration Without such integration and increased scholarly and practitionerattention to climate and consensus many organizations may fail to supportgroup membersrsquo explicit use of their cultural experiences to advanceorganizational learning from enacted diversity change strategies despite thegroup and intergroup social tensions that will inevitably occur (Thomas andEly 1996)

References

Aiken LA and West SG (1993) Multiple Regression Testing and Interpreting InteractionsSage Publications Newbury Park CA

Alderfer CP and Smith KK (1982) ordfStudying embedded intergroup relations inorganizationsordm Administrative Science Quarterly Vol 27 pp 35-65

Allport G (1954) The Nature of Prejudice Addison Wesley Cambridge MA

Bailyn L (2000) ordfGender and diversity in organizations preconference presentationordm paperpresented at the National Academy of Management meeting Toronto

JOCM163

350

Blalock HM (1967) Toward a Theory of Minority Group Relations Wiley New York NY

Bleise PD and Halverson RR (1998) ordfGroup consensus and psychological well-being a largeregeld studyordm Journal of Applied Social Psychology Vol 28 pp 563-80

Brown KG Kozlowski SWJ and Hattrup K (1996) Theory Issues and Recommendations inConceptualizing Agreement as a Construct in Organizational Research The Search forConsensus Regarding Consensus National Academy of Management Cincinnati OH

Cox T Jr (1993) Cultural Diversity in Organizations Berrett-Koehler San Francisco CA

Ely R (1994) ordfThe effects of organizational demographics and social identity on relationshipsamong professional womenordm Administrative Science Quarterly Vol 39 pp 203-38

Ely R (1995) ordfThe power in demography womenrsquos social constructions of gender identity atworkordm Academy of Management Journal Vol 38 pp 589-634

Festinger L (1954) ordfA theory of social comparison processesordm Human Relations Vol 7pp 117-40

Forsyth DR (1990) Group Dynamics BrooksCole Publishing Paciregc Grove CA

Harrison DA Price KH and Bell MP (1998) ordfBeyond relational demography time and theeffects of surface and deep-level diversity on work group cohesionordm Academy ofManagement Journal Vol 41 pp 96-107

James K Lovato C and Khoo G (1994) ordfSocial identity correlates of minority workersrsquo healthordmAcademy of Management Journal Vol 37 p 383+

Kanter R (1977) ordfSome effects of proportions on group life skewed sex ratios and reactions totoken womenordm American Journal of Sociology Vol 82 pp 965-90

Kanter R (1983) Men and Women of the Corporation Basic Books New York NY

Kossek EE and Lobel SA (1996) ordfIntroduction transforming human resource systems tomanage diversityordm in Managing Diversity Human Resource Strategies for Transforming

the Workplace Blackwell Oxford pp 1-19

Kossek EE and Zonia SC (1993) ordfAssessing diversity climate a regeld study of reactions toemployer efforts to promote diversityordm Journal of Organizational Behavior Vol 14pp 61-81

Larkey LK (1996) ordfToward a theory of communicative interactions in culturally diverseworkgroupsordm Academy of Management Review Vol 21 pp 463-91

Lau DC and Murnighan JK (1998) ordfDemographic diversity and faultlines the compositionaldynamics of organizational groupsordm Academy of Management Review Vol 29 pp 325-40

Lindell MK and Brandt CJ (2000) ordfClimate quality and climate consensus as mediators of therelationship between organizational antecedents and outcomesordm Journal of AppliedPsychology Vol 85 pp 331-48

McGrath JE Berdahl JL and Arrow H (1995) ordfTraits expectations culture and clout thedynamics of diversity in work groupsordm in Jackson SE and Ruderman MN (Eds)Diversity in Work Teams Research Paradigms for a Changing Workplace AmericanPsychological Association Washington DC

Payne RL and Pugh DS (1976) ordfOrganizational structure and climateordm in Dunnette MD(Ed) Handbook of Industrial and Organizational Psychology Rand McNally Chicago ILpp 1125-73

Reichers A and Schneider B (1990) ordfClimate and culture an evolution of constructsordm inSchneider B (Ed) Organizational Climate and Culture Jossey-Bass San Francisco CA

HRM changestrategy

351

the context of speciregc groups at all organizational levels In sum HR strategistsand change agents should focus on how HR change strategies relate to the jointalignment of group memberrsquos multiple characteristics (eg demographichierarchical status task) in work groups plusmn that is to examine group contextualinmacruences (Wharton 1992)

54 LimitationsOur study while making contributions to organizational research is notwithout limitations Though large in individual respondents by utilizing thegroup as the level of analysis our sample size includes only 74 groups Thiscould reduce the level of statistical power of our regndings Additionally becausethe sample was based on one organization it is important to replicate thisanalysis across multiple organizations to substantiate our results Althoughchanges in department heterogeneity across an eight-year period wereincluded this is not a longitudinal design Future research should employrepeated measures over a length of time across all variables measuringdemographic change in heterogeneity as well as consensus and valence Ourstudy points to some of the challenges in measuring ordfsuccess in increasingdiversityordm as we did not look at annual turnover or the effects of the glassceiling at the group level in this study which is a gap we hope futureresearchers will be motivated to study spurred by our research

Since little published work has been done on group consensus regarding theeffects of HR policies to increase diversity through demographic change weextend many social science literatures ranging from organizationaldevelopment to demography and HR Future research on diversity andorganizational change also should strive to be interdisciplinary as theliterature on HR strategy and change has been largely developed in the HRliterature separate from other social science disciplines reducing theoreticalintegration Without such integration and increased scholarly and practitionerattention to climate and consensus many organizations may fail to supportgroup membersrsquo explicit use of their cultural experiences to advanceorganizational learning from enacted diversity change strategies despite thegroup and intergroup social tensions that will inevitably occur (Thomas andEly 1996)

References

Aiken LA and West SG (1993) Multiple Regression Testing and Interpreting InteractionsSage Publications Newbury Park CA

Alderfer CP and Smith KK (1982) ordfStudying embedded intergroup relations inorganizationsordm Administrative Science Quarterly Vol 27 pp 35-65

Allport G (1954) The Nature of Prejudice Addison Wesley Cambridge MA

Bailyn L (2000) ordfGender and diversity in organizations preconference presentationordm paperpresented at the National Academy of Management meeting Toronto

JOCM163

350

Blalock HM (1967) Toward a Theory of Minority Group Relations Wiley New York NY

Bleise PD and Halverson RR (1998) ordfGroup consensus and psychological well-being a largeregeld studyordm Journal of Applied Social Psychology Vol 28 pp 563-80

Brown KG Kozlowski SWJ and Hattrup K (1996) Theory Issues and Recommendations inConceptualizing Agreement as a Construct in Organizational Research The Search forConsensus Regarding Consensus National Academy of Management Cincinnati OH

Cox T Jr (1993) Cultural Diversity in Organizations Berrett-Koehler San Francisco CA

Ely R (1994) ordfThe effects of organizational demographics and social identity on relationshipsamong professional womenordm Administrative Science Quarterly Vol 39 pp 203-38

Ely R (1995) ordfThe power in demography womenrsquos social constructions of gender identity atworkordm Academy of Management Journal Vol 38 pp 589-634

Festinger L (1954) ordfA theory of social comparison processesordm Human Relations Vol 7pp 117-40

Forsyth DR (1990) Group Dynamics BrooksCole Publishing Paciregc Grove CA

Harrison DA Price KH and Bell MP (1998) ordfBeyond relational demography time and theeffects of surface and deep-level diversity on work group cohesionordm Academy ofManagement Journal Vol 41 pp 96-107

James K Lovato C and Khoo G (1994) ordfSocial identity correlates of minority workersrsquo healthordmAcademy of Management Journal Vol 37 p 383+

Kanter R (1977) ordfSome effects of proportions on group life skewed sex ratios and reactions totoken womenordm American Journal of Sociology Vol 82 pp 965-90

Kanter R (1983) Men and Women of the Corporation Basic Books New York NY

Kossek EE and Lobel SA (1996) ordfIntroduction transforming human resource systems tomanage diversityordm in Managing Diversity Human Resource Strategies for Transforming

the Workplace Blackwell Oxford pp 1-19

Kossek EE and Zonia SC (1993) ordfAssessing diversity climate a regeld study of reactions toemployer efforts to promote diversityordm Journal of Organizational Behavior Vol 14pp 61-81

Larkey LK (1996) ordfToward a theory of communicative interactions in culturally diverseworkgroupsordm Academy of Management Review Vol 21 pp 463-91

Lau DC and Murnighan JK (1998) ordfDemographic diversity and faultlines the compositionaldynamics of organizational groupsordm Academy of Management Review Vol 29 pp 325-40

Lindell MK and Brandt CJ (2000) ordfClimate quality and climate consensus as mediators of therelationship between organizational antecedents and outcomesordm Journal of AppliedPsychology Vol 85 pp 331-48

McGrath JE Berdahl JL and Arrow H (1995) ordfTraits expectations culture and clout thedynamics of diversity in work groupsordm in Jackson SE and Ruderman MN (Eds)Diversity in Work Teams Research Paradigms for a Changing Workplace AmericanPsychological Association Washington DC

Payne RL and Pugh DS (1976) ordfOrganizational structure and climateordm in Dunnette MD(Ed) Handbook of Industrial and Organizational Psychology Rand McNally Chicago ILpp 1125-73

Reichers A and Schneider B (1990) ordfClimate and culture an evolution of constructsordm inSchneider B (Ed) Organizational Climate and Culture Jossey-Bass San Francisco CA

HRM changestrategy

351

Blalock HM (1967) Toward a Theory of Minority Group Relations Wiley New York NY

Bleise PD and Halverson RR (1998) ordfGroup consensus and psychological well-being a largeregeld studyordm Journal of Applied Social Psychology Vol 28 pp 563-80

Brown KG Kozlowski SWJ and Hattrup K (1996) Theory Issues and Recommendations inConceptualizing Agreement as a Construct in Organizational Research The Search forConsensus Regarding Consensus National Academy of Management Cincinnati OH

Cox T Jr (1993) Cultural Diversity in Organizations Berrett-Koehler San Francisco CA

Ely R (1994) ordfThe effects of organizational demographics and social identity on relationshipsamong professional womenordm Administrative Science Quarterly Vol 39 pp 203-38

Ely R (1995) ordfThe power in demography womenrsquos social constructions of gender identity atworkordm Academy of Management Journal Vol 38 pp 589-634

Festinger L (1954) ordfA theory of social comparison processesordm Human Relations Vol 7pp 117-40

Forsyth DR (1990) Group Dynamics BrooksCole Publishing Paciregc Grove CA

Harrison DA Price KH and Bell MP (1998) ordfBeyond relational demography time and theeffects of surface and deep-level diversity on work group cohesionordm Academy ofManagement Journal Vol 41 pp 96-107

James K Lovato C and Khoo G (1994) ordfSocial identity correlates of minority workersrsquo healthordmAcademy of Management Journal Vol 37 p 383+

Kanter R (1977) ordfSome effects of proportions on group life skewed sex ratios and reactions totoken womenordm American Journal of Sociology Vol 82 pp 965-90

Kanter R (1983) Men and Women of the Corporation Basic Books New York NY

Kossek EE and Lobel SA (1996) ordfIntroduction transforming human resource systems tomanage diversityordm in Managing Diversity Human Resource Strategies for Transforming

the Workplace Blackwell Oxford pp 1-19

Kossek EE and Zonia SC (1993) ordfAssessing diversity climate a regeld study of reactions toemployer efforts to promote diversityordm Journal of Organizational Behavior Vol 14pp 61-81

Larkey LK (1996) ordfToward a theory of communicative interactions in culturally diverseworkgroupsordm Academy of Management Review Vol 21 pp 463-91

Lau DC and Murnighan JK (1998) ordfDemographic diversity and faultlines the compositionaldynamics of organizational groupsordm Academy of Management Review Vol 29 pp 325-40

Lindell MK and Brandt CJ (2000) ordfClimate quality and climate consensus as mediators of therelationship between organizational antecedents and outcomesordm Journal of AppliedPsychology Vol 85 pp 331-48

McGrath JE Berdahl JL and Arrow H (1995) ordfTraits expectations culture and clout thedynamics of diversity in work groupsordm in Jackson SE and Ruderman MN (Eds)Diversity in Work Teams Research Paradigms for a Changing Workplace AmericanPsychological Association Washington DC

Payne RL and Pugh DS (1976) ordfOrganizational structure and climateordm in Dunnette MD(Ed) Handbook of Industrial and Organizational Psychology Rand McNally Chicago ILpp 1125-73

Reichers A and Schneider B (1990) ordfClimate and culture an evolution of constructsordm inSchneider B (Ed) Organizational Climate and Culture Jossey-Bass San Francisco CA

HRM changestrategy

351

Schein E (1988) Process Consultation Addison-Wesley Boston MA

Schneider B and Reichers AE (1983) ordfOn the etiology of climatesordm Personnel PsychologyVol 36 pp 19-39

Schneider KT Hitlan RT and Radhakrishnan P (2000) ordfAn examination of the nature andcorrelates of ethnic harassment experiences in multiple contextsordm Journal of AppliedPsychology Vol 85 pp 3-12

Thomas DA and Ely RJ (1996) ordfMaking differences matter a new paradigm for managingdiversityordm Harvard Business Review pp 79-90

Triandis HC Krowski LL and Gelfand MJ (1994) ordfWorkplace diversityordm in Triandis HCand Dunnette MD (Eds) Handbook of Industrial and Organizational PsychologyConsulting Psychologistrsquos Press Palo Alto CA Vol 4 pp 769-827

Tsui AS Egan TD and OrsquoReilly CA (1992) ordfBeing different relational demography andorganizational attachmentordm Administrative Science Quarterly Vol 37 pp 547-79

Tsui A and Gutek B (1999) Demographic Differences in Organizations Current Research amp

Future Directions Lexington Books New York NY

Wharton A (1992) ordfThe social construction of gender and race in organizations a social identityand group mobilization perspectiveordm in Tolbert PT and Bacharach SB (Eds) Researchin the Sociology of Organizations JAI Press Greenwich CT Vol 10 pp 55-84

Zander A (1994) Making Groups Effective Jossey-Bass San Francisco CA

Further reading

Kossek EE Zonia S and Young W (1996) ordfThe limitations of organizational demographycan diversity climate be enhanced in the absence of teamworkordm in Ruderman H Hughes-James MW and Jackson S (Eds) Selected Research on Work Team Diversity AmericanPsychological Association Washington DC pp 121-54

JOCM163

352