' Academy of Management Executive, 2000, Vol. 14. No.

Managing five paradoxes of360-degree feedback

Jai Ghorpade

Executive OverviewThe performance feedback method known as 360-degree feedback has gained wide

popularity in the corporate world to the point of being nearly universal among Fortune500 companies. A 360-degree feedback program enables organizational members toreceive feedback on their performance, usually anonymously, from all the majorconstituencies they serve. Unlike the traditional approach to performance counseling, the360-degree feedback concept does not rely solely on the supervisor as the source ofinformation. Instead, it enlists superiors, peers, subordinates, suppJiers, and customers inproviding individuals with feedback on different aspects of their performance. Feedbackrecipients can also rate their own performance and compare it with feedback providedby others. Although the 360-degree method is widely used, its application is filled withparadoxes. While it delivers valuable feedback, the 360-degree concept has seriousproblems relating to privacy, validity, and effectiveness. This article identifies fiveparadoxes of 360-degree feedback programs and offers suggestions for managing them.

Significant management problems, according toCharles Handy, come in the form of paradoxes.^ Aclean solution to a problem, without any accompa-nying dysfunctional effects, is possible only intheory. In the practical, dynamic world of organ-izations, solutions to significant managerialproblems inevitably bring with them conse-quences that are contradictory and inconsistent.^The art of management consists of accepting par-adoxes as endemic to organizational life and find-ing ways of coping with them to attain desiredends. Rather than shy away from a particular idea,theory, or method because it is accompanied by anundesirable consequence, the manager needs tofind ways of extracting the positive and minimiz-ing or controlling the negative.

This advice from one of the leading philosophersof management is well worth heeding with regardto 360-degree feedback, a performance improve-ment method that is rapidly gaining acceptance inthe corporate world but that is fraught with para-doxes. The rapidly expanding list of 360-degreefeedback users currently includes leaders of thecorporate sector such as AT&T, Exxon, GE, Amoco,IBM, Caterpillar, Levi Strauss, and Shell Oil. Twen-ty-two of Fortune's 32 most admired companieswere using upward or 360-degree feedback as of

1994. By 1996, 360-degree programs had becameneariy universal among Fortune 500 companies,which spend hundreds of millions of dollars annu-ally to support them.3

The attraction of the 360-degree concept forAmerican industry is easy to understand. Duringthe past two decades, U. S. corporations have beeninvolved in a massive program of restructuring tocope with the demands of the emerging globalmarketplace. Old products are being discardedand new ones are being created to service theneeds of increasingly discriminating consumers.Internally, there is a move from the old bureau-cratic style of management toward a flat organiza-tion that requires active participation from therank and file and that is more suited to the rapidresponse industry now has to make to changes inproduct demand and supply markets.'*

In this atmosphere, the 360-degree feedback con-cept has much to offer. Unlike the traditional per-formance appraisal model, in which superiorsevaluate subordinates, the 360-degree approachdoes not rely solely on the superior to providefeedback to the employee. Instead, it enlists mul-tiple constituencies to provide feedback to selectedorganizational members. These constituencies in-clude superiors, peers, coworkers in support areas.

140

2000 Ghorpade 141

subordinates, internal customers of the unit's work,and external customers of the organization's prod-ucts. In this process, the feedback recipient is ex-pected to evaluate his or her own performance onthe selected behavioral dimensions. This self-eval-uation is then compared with that provided by theother feedback providers. The recipient is encour-aged to use the feedback to improve performanceand to make a greater effort to blend his or hercontributions with the needs of the group. Thislinking ot individual performance with feedbackfrom all relevant constituencies fits well into theemerging team-based workplace.^

Another difference from traditional performanceappraisal is that 360-degree feedback is supposedto be given anonymously. Research has demon-strated that anonymous feedback is more honestand closer to what raters actually feel about thefeedback recipients. Appraisers whose identity isknown to the feedback recipients give higher rat-ings than those who are anonymous.^

Research has demonstrated thatanonymous feedback is more honest andcloser to what raters actually feel aboutthe feedback recipients.

Although any organizational member whose jobinteracts with others can participate in feedbackexercises, 360-degree feedback is typically pro-vided to managers, the only members with a fullcircle of 360-degree feedback providers. A surveyof the members of the Society of Human ResourceManagers found that 35 percent of the organiza-tions used 360-degree feedback primarily for exec-utives, and 37 percent for upper middle managers.Middle and first-level managers also were in-cluded, but to a lesser degree (23 percent and 18percent, respectively).'^

At this stage, no single corporate role is typicallyin charge of the 360-degree process. The person incharge could be the feedback receiver's superior,the human resources department, or even a com-mittee.^

While the 360-degree concept has much to offerand many successes have been documented,^there are also stories of confusion and disappoint-ment. Many of the 360-degree programs are carriedout in the absence of a strategic context, and fail tofocus on contributions that they can make to afirm's competitive advantage.'o There is little con-sistency to what is being done, and 360-degreefeedback programs can range from any deviationsfrom the traditional vertical form of performance

appraisal to highly sophisticated feedback sys-tems that systematically gather, analyze, and dis-seminate behavior data to managers, profession-als, and even rank-and-file workers functioning inteams.^^

Many organizations adopt 360-degree feedbackwithout clearly defining the mission and the scopeof the program. Consequently, employees who re-ceive the feedback are left to figure out for them-selves how to cope with the results and tend notto develop goals and action plans following 360-degree applications.'2 One study concluded thatwhile such programs are popular, in many caseslittle more than lip service is paid to them.'^ Fur-thermore, there is discouraging evidence regard-ing the effectiveness of feedback-intervention pro-grams as tools in bringing about improvements inperformance. A review of over 600 feedback studiesfound that only one-third reported improvementsin performance. Another third reported negativechanges in performance, while the final third re-ported no impact.'**

In their haste to gain the advertised benefits of360-degree feedback, organizations may not besufficiently aware of the problems that often ac-company its adoption.'^ Failure to recognize theparadoxes that often occur can lead to disillusion-ment, reduce the value of the exercise, and confirmthe lip service that tends to be paid to 360-degreeresults.

Five paradoxes are discussed in this article.They are organized around four issues involved inthe adoption of 360-degree feedback: Objectives,sources of behavior and performance data, meth-ods of data gathering and feedback, and selectionof a program administrator. In each case, the par-adoxes and their sources are identified, and sug-gestions are offered for aiding managers to copewith them.

Objectives of 360-Degree Feedback

A firmly established tenet of 360-degree feedbackis that such feedback is to be used for developmen-tal rather than appraisal purposes.'^ Yet a paradoxis associated with this tenet;

Employee Development Paradox

The primary objective of 360-degree feedback is todevelop rather than to appraise the participatingorganizational members. In practice, however, the360-degree process gets entangled with the ap-praisal process, thus creating the potential for con-fusion and erosion of its usefulness as a develop-mental tool. i I

142 Academy of Management Executive February

There are two reasons for this. First, the verynature of the process invites involvement andhence sets the stage for information seepage.Under a typical 360-degree program, groups ofindividuals from different levels are invited tocomment anonymously on the performance ofone or more individuals. However, such discus-sions sometimes go in directions other thanthose anticipated by the administrators. Datagathered for developmental purposes can leakout and become a part of the appraisal process inseveral ways. The feedback providers, particu-larly if they are untrained, may vent their ownfrustrations and problems, be unable to distin-guish between individual failings and organiza-tional defects, and generally raise issues that gobeyond the performance of the individuals in-volved.

The ratees themselves can be another source ofleaks. Feedback receivers who score high onachievement motivation and perceive high valueto the appraisal feedback are more likely to dis-cuss their appraisal results with appraisers." Ra-tees who are fearful of the process may attempt toforce the providers to reveal themselves by intim-idation. For example, at the Department of Ener-gy's Operations Office in Albuquerque, N. M., someemployees filed Federal Privacy Act requests todiscover the source of certain ratings.'^

The second reason why 360-degree feedback canget entangled with performance appraisal has itsroots in the economic and political realities of or-ganizations. Since 360-degree feedback appraisalsare expensive, companies want to get their mon-ey's worth by addressing multiple objectives.'^

This conflict between the developmental and ap-praisal objectives is a true paradox that has to bemanaged. The solutions depend on preferences ofmanagement with regard to objectives. Here aretwo solutions that assume different objectives: de-velopmental and appraisal.

Solution 1. If possible, keep the 3G0-degiee pro-gram as a developmental tool, and formulateclear rules for information sharing. To cope withthe problem of information seepage under anon-ymous feedback, managers must recognize thatindividuals will attempt to gain additional infor-mation and that leaving them to do so on theirown is counterproductive.^° One way to encour-age constructive information sharing is to allowfeedback receivers to solicit additional informa-tion on specific issues that concern them. Toprotect the providers' identity, the informationcan be given in writing.

A second way to promote constructive informa-tion sharing is to encourage peers and other co-

workers to share and clarify their views with thefeedback receivers on a voluntary basis, and underthe guidance of their supervisors. This can be doneface-to-face, or indirectly through the supervisor.Face-to-face meetings break the anonymity of theraters, but the confidentiality of ratings given byindividual raters can still be protected by requir-ing the attendees to focus on issues and themesthat emerged from the feedback rather than on theentire report. Such personal confrontations, how-ever, are risky, and should be undertaken onlywhere trust levels are high and the culture sup-ports openness.

One way to encourage constructiveinformation sharing is to allow feedbackreceivers to solicit additional informationon specific issues that concern them.

Sharing and clarifying may also be done indi-rectly by the supervisor by interviewing the raters,individually or in groups, and asking them to re-spond to specific questions raised by the recipi-ents. This offers the additional potential benefit ofcontrolling a serious problem facing the entire pro-cess—making the feedback providers act respon-sibly in performing their roles as raters.̂ -'

Solution 2: Combining 3GQ-degree feedback withperformance appraisal should be pursued gradu-ally, as part of a wider performance managementplan. If this course of action is chosen, manage-ment needs to face the fact that the political cli-mate often shifts when peers evaluate one anoth-er's performance. When 360-degree feedback istied to performance appraisal, and hence to pro-motions and wage increases, individuals and sub-groups may gang up to sabotage the process.^^Damage can be minimized by beginning the pro-cess with a 360-feedback program and then mov-ing gradually to appraisal. This is viable, however,only if the employees view the process as fair,non-threatening, and beneficial.

Alternatively, traditional performance appraisalprograms can gradually widen the base of ap-praisals by including raters other than superiors.For example, if peers are added to the list of ratersand the addition is accepted as beneficial by theratees, then they may be willing to broaden thebase further by including their subordinates.^^

Sources of 360-Degree Feedback Data

Two critical assumptions are made in the 360-degree feedback literature with regard to sources

2000 Ghorpade 143

of information. The first is that inciuding multiplesources widens the scope of information that isuncovered and sheds light on different facets ofperformance. The second is that anonymous rat-ings yield feedback that is more honest, and hencemay be more valid. While the practice of includingmultiple constituencies anonymously in the 360-degree process is now an integral part of the tech-nique, two paradoxes arise:

Multiple Constituents Paradox

Involvement of multiple constituents in 360-degreefeedback process broadens the scope of informa-tion provided to the receiver. However, more infor-mation does not necessarily yield better feedback.

Anonymous Ratings Paradox

Anonymous ratings are more honest than signedratings. However, honest ratings may not neces-sarily be more valid.

The need for broadening the sources of informa-tion in performance feedback and evaluation hasbeen recognized for some time.^^ There is no doubtthat involving multiple constituents broadens thescope of information that is gathered.^^ However, amere increase in the scope of information may notnecessarily yield data that are more accurate, im-partial, and competent than those provided by theindividual manager serving as evaluator. Honestratings, the major reason for making the processanonymous, are not necessarily more valid.

Inaccurate, biased, and even self-serving infor-mation can make its way into 360-degree feedbackbecause of informational, cognitive, and affectivecauses.26 At the informational level, discrepanciesin evaluation among multiple sources may arisewhen there is no clear idea what is expected ofratee roles, no clear criteria according to whichthey are judged, and no opportunities for observa-tion. Whether the ratings are honest is not theissue; these discrepancies arise simply becauseone or more of the parties lacks the informationneeded to provide accurate ratings, or has thewrong information about the roles in question.2'For example, managers from large corporationswho are recruited by small businesses may bringwith them expectations about roles that do notmatch those of their prospective bosses.

At the cognitive level, coping with the complex-ity of the task is often a problem. Appraising aperson in a role is a complex activity, requiring theevaluator to rationally acquire, store, retrieve, andintegrate complex sets of information about theperson, the job, the outputs, and time frames, and

then to use this information to pass judgment.Evaluators typically simplify the process by form-ing an overall impression of the ratee rather thanby focusing on specific behaviors; they rate ac-cording to this impression, discounting any incon-sistent behaviors.28 Honesty is not an issue in suchcognitive structuring: The raters might honestlybelieve that their impressions reflect reality.

Affective constraints, the third general cause ofevaluation discrepancies, stem from psychologicaland political factors. Individuals may distort rat-ings, unconsciously or consciously, to protect theirself-concept or to serve other personal ends. At theunconscious level, individuals may rate them-selves higher than their peers to preserve theirfavorable self-esteem. This in turn may lead to aself-serving bias—the tendency to take personalcredit for successful performance, but to assignresponsibility for failure to external causes.^a Un-conscious distortion of ratings of others may alsooccur because of hidden race, gender, age, or per-sonality biases held by the raters.^o

Conscious distortion of ratings, the second typeof affective bias, has its roots in the political as-pects of organizational life. As Burns pointed out 30years ago: "Members of a business concern are atone and the same time cooperators in a commonenterprise and rivals for the material and intangi-ble rewards of successful competition with eachother. The hierarchical order of rank and power,realized in the organization chart, which prevailsin all organizations, is both a control system and acareer ladder."^'

Given this duality of roles, it is normal for peoplein organizations to take into account the negativeconsequences for themselves of the ratings thatthey give to others. One of these consequencesmay be unpleasant future personal relations withcoworkers.22 Another might be a reduction in com-petitive status: The assignment of good ratings topeers, even when they are deserved, may reducethe relative ranking of the rater. Deliberate infla-tion or deflation of ratings may occur because offeelings of friendship or animosity, to preserve de-partmental territory and resources, or when eval-uators are confronted with decisions involving al-location of scarce resources, such as distributinglimited promotions and pay raises.^^

The issue of honesty in ratings needs to be morefully confronted. It is accepted axiomatically in the360-degree literature that anonymity enhanceshonesty in ratings,^^ and considerable support isdocumented for this proposition.^^

While conceding that anonymity can enhancehonesty, it would be risky to equate honesty withvalidity of ratings. Untrained raters might inadver-

144 Academy of Management Executive February

tently fall prey to the rating errors that industrialpsychologists have warned about over the years,namely halo, leniency, strictness, and central ten-dency. Halo refers to assigning ratings on specificcriteria on the basis of a general impression of theratee; leniency refers to assigning higher than de-served scores; strictness (also referred to as sever-ity) is the opposite of leniency and refers to assign-ing ratings lower than deserved; and centraltendency refers to clustering ratings near the cen-ter of the scale.^^

From a managerial perspective, the challengethat these paradoxes raise is twofold: how to evokeconstructive responses from the multiple constitu-ents, and how to incorporate the various facets ofinformation into a feedback package for the infor-mation receivers that will enable them to improvetheir performance. Some of the solutions to theseissues are technical and will be covered in the nextsection which deals with methods. At this stage,managers might consider the following sugges-tions for controlling these two paradoxes.

Solution 1: Get away from the assumption thatincreasing the number of rater constituencies au-tomatically improves the quality of the feedbackprovided, and focus instead on the relevance of thefeedback from a developmental perspective. AsLondon and Beatty point out, the customer may notalways be right or consistent in evaluating theperformance of a work unit and its leadership.^^ Inorder for the ieedback recipient to fully benefitfrom multiple data sources, the information gath-ered must add value beyond that acquired throughtraditional performance appraisal. Effective feed-back, regardless of source, should focus on the jobas the incumbent performs it within the context ofthe organization, rather than on a role in an ab-stract sense; should focus on concrete behaviorsthat are linked to specific events and incidents,rather than on personal traits; should be free ofracial, sexual, and other biases; and should pro-vide concrete recommendations for change.^^

Solution 2: Provide the raters with guidance andtraining, including a description of the major com-petencies expected of the role. This is particularlyrelevant in a full 360-degree application, wheredifferent constituencies might be expected to havevarying degrees of familiarity with the job. Forexample. United Jersey Bank Financial decided touse 360-degree feedback as part of a strategic re-structuring in pursuit of a vision of total customersatisfaction. They began by redefining five mana-gerial jobs in units that required close customercontact. Competencies judged to be relevant forthe future were incorporated into the 360-degreeprocess. The ratings were then correlated with ob-

jective measures of bank performance. The feed-back recipients were then provided with trainingto enable them to improve their behaviors within aspecified time frame.^^

Solution 3: Provide potential raters with opportu-nities to detect their own rating biases, if any, be-fore they actually rate an organizational member.This is particularly relevant in instances wherestructured numerically scored scales are used be-cause they are highly subject to unconscious rat-ing errors of halo, leniency, central tendency, andstrictness. These types of biases can be uncoveredby getting a group of potential raters to use theinstrument in a trial session by focusing on de-scriptions of hypothetical candidates who varywidely in their behavioral profiles. The commonpractice is to calculate means and standard devi-ations of the ratings and to display the results incomparative terms on a chalkboard or transpar-ency. Each rater knows his or her own numbers,but the group as a whole sees only the distribution.Both of these statistics enable individual raters toknow where they stand in reference to the groupand thus to detect whether they habitually gradehigh, low, or in the middle in comparison with theothers.*°

These suggestions are relevant only if it can beassumed that the raters are in fact willing to act ingood faith. No amount of training, however, is go-ing to be of any help if the organizational climateis politically charged and trust is low. In suchinstances, rather than focusing on individuals, theorganization would be better off using the logic ofthe 360-degree process to assess the standing ofthe organization itself in the eyes of its variousconstituencies.

No amount of training, however, is goingto be of any help if the organizationalclimate is politically charged and trustis low.

Methods of Data Gathering and Feedback

Data needed for 360-degree feedback can be gath-ered in a variety of ways. Programs can vary in theform of information gathered (e.g., quantitativeversus qualitative); the amount of structure in theprocedure (scaled or open-ended questions) usedto gather the data; and the extent to which theinformation gathered is context-specific versus ge-neric. A context-specific method would focus onbehaviors relevant to the immediate work unit andits culture, while a relatively generic 360-degree

2000 Ghorpade 145

method would focus on generalized behaviors thathave been judged to be common to the role inquestion, either through theory, industry experi-ence, or expert opinions.^^

The three choices with regard to method notedabove are not mutually exclusive. In theory, manycombinations are possible. For example, an instru-ment can be partially structured with a combina-tion of scaled and open-ended questions. Genericinformation {e.g., industry practices) can be com-bined with company variables in thinking aboutits relevance to the company.

Even though the three method choices are notmutually exclusive, a decision nevertheless has tobe made about which combination to use in aparticular context. In making this decision, a par-adox is encountered:

Sfrucfured Feedback Paradox

Quantitative and structured feedback based on ge-neric behaviors is easy to acquire, score, and dis-seminate. However, such data may not have muchrelevance to a particular workplace and may evenyield misleading results.

Structured, quantitative, and generic feedbackoffers two advantages. First, this approach is fairlyinexpensive. For example, the Center For CreativeLeadership sells its multirater assessment tool.Prospector, for $195 for a set of 12 questionnaires—one for self-rating and the other 11 for coworkerratings.^^ Second, the skills needed internally toadminister such a program are minimal. Basically,the role of the HR representative is to perform cler-ical tasks—ordering the form, administering it,and circulating the numerical scores.

Despite these benefits, this combination of quan-titative ratings acquired through a structuredquestionnaire containing generic role behavior in-formation can raise some very serious problemswith regard to fairness and accuracy. To beginwith, quantitative data of a generic type can bedifficult to interpret. Suppose a manager of anemergency ward in an inner city hospital receivesa mean score of 3.7 on a five-point scale of effec-tiveness in explaining to subordinates their dutiesand responsibilities. Suppose further that the in-dustry average for this behavior (supplied by theinstrument publisher) is 4.2. Is this manager to beviewed as inferior because of a lower score thanthe industry average? Our ability to pass any kindof judgment would hinge on the extent to which theworkplaces are comparable with regard to em-ployee readiness to receive direction, the commu-nication media made available to the manager,and a host of other environmental variables.

There is also the problem of controlling for therater tendencies discussed earlier. In order to beable to assign a substantive value with confidenceto a statistic resulting from a 360-degree adminis-tration, the ratings would need to be consideredfree from the informational and cognitive errorsnoted earlier. Even this is very difficult and is fre-quently not done in practice, as illustrated by thecase of a manager who was told by her peers thather informing behaviors, while frequent, were notsufficient. This created a problem of interpreta-tion— how much more can someone use a behaviorthat is already employed frequently? It was discov-ered at the information sharing and clarifyingmeeting that the weekly reports that the managerhad decided not to circulate because people werecomplaining about being overloaded with paperswere the very documents that they wanted to see.''^Their reservation was not with her ability to en-gage in informing behaviors, a generic variable,but a failure to circulate a particular document, acontext-specific variable.

Consider now what it would take for a companyto move toward the context-specific end of thecontinuum. As London and Beatty point out, 360-degree feedback ratings should be made on per-formance dimensions that are strategic to organ-izational success and thus relevant to the job.'''' TheHR representative would have to have both thetime and the ability to analyze tasks and behaviorsinvolved in the jobs of the persons selected to par-ticipate in the program, translate these into mea-surable behaviors, and provide feedback that isspecific to the context. This would no doubt in-crease the cost, both in terms of time and moneyneeded to develop and administer such a program.

How does a company balance these conflictingoptions? Rather than see them as black-and-whitechoices, compromises are possible. An operationalprocedure for constructing a 360-degree feedbackinstrument that balances the conflicting options ofthe structured feedback paradox is given in Figure1.̂ ^ The preliminary step, as in all 360-degree pro-grams, is to choose the feedback receivers andproviders. Simultaneously, a panel of Subject Mat-ter Experts (SMEs) is formed, consisting of the feed-back receivers and providers, their peers, manag-ers, subordinates, and others who have a stake inthe jobs held by the feedback receivers. This isfollowed by selection of generic and structuredbehavior inventories. These are examined andrated by all the SMEs, including the feedback re-ceivers and providers, for context relevancy. Onlythose that are judged to be context-relevant areretained and supplemented by other items of inter-est to the organization. This collection of context-

146 Academy of Management Executive February

Choose 360-degTee feedback receiversand providers and select subject

matter experts (SMEs).

Select behavior inventories.

Ask SMEs to select context-relevant itemsfrom behavior inventories and to add

other items that they judge to be relevant.

Format selected behavior items into asemi-structured questionnaire.

Require receivers and providers torate behaviors and explain their ratings.

Provide feedback to receiversand providers.

HGURE 1A Procedure for Developing and Administering aContext-Relevant 360-Degree Feedback Method

relevant items is then be formatted into a semi-structured questionnaire allowing for feedbackthat is both quantitative and qualitative, struc-tured, and context-specific. In keeping with theearlier discussion relating to the paradoxes deal-ing with employee development, multiple constit-uents, and anonymous ratings, raters are asked toprovide ratings as well as to explain them.

The relevance of the procedure in Figure 1 forcontrolling the conflicts of the paradox dealingwith structured feedback are evident in the abovediscussion. Nevertheless, here are some specificapplications:

Solution 1: The problem of balancing the quan-titative and qualitative information can be han-dled by asking the raters to provide both forms ofinformation on the performance dimensions in-cluded in the feedback instrument. Thus, for each

rating item on the instrument, space would be pro-vided for qualitative comments. It can be pointedout to potential raters during the training sessionsthat their feedback will be taken more seriously ifthey provide concrete examples of behaviors, inci-dents, and events that led them to the ratings. Infact, a more stringent rule can be imposed: Ratingswill not be counted unless they are explained.

Solution 2: The conflict stemming from the sec-ond continuum—confexf-speci/ic versus genericfeedback—can be managed by combining bothtypes of items in the instrument. A pure context-specific 360-degree instrument consists of a set ofcritical behaviors that have been arrived atthrough a systematic job analysis procedure fromscratch, without any use whatsoever of any exist-ing behavior inventories. While such a procedurehas its appeal, it is unnecessary {and even waste-ful) in today's context to begin from scratch forevery job analysis. An alternative approach is touse inventories that are available in publishedform as starting points for arriving at context-relevant behaviors. These can then be supple-mented with context-relevant behaviors not foundin the inventories.

Organizations willing to entertain this approachhave a wealth of materials to draw on. There are atleast five types of inventories available: Descrip-tions of generalized human behaviors relevant tothe work context;"^ descriptions of tasks typicallyperformed in thousands of jobs in today's econo-my; '̂' documentation of skills involved in perform-ing managerial and other complex jobs;''^ abstractdescriptions of human skills and abilities relevantto some degree in all work contexts;'*^ and inven-tories and instruments assembled specifically byvarious publishers for 360-degree feedback.^" It isrelevant to note that developers of the last set drawheavily on the other four for their materials.

The end result of this process is an instrumentcapable of generating feedback that is both con-text-relevant and inclusive of behavior items ofgeneral applicability. This procedure should alsoenlist the support of feedback receivers, as theywould have had a part in generating the instru-ment.^'

The final step in the process is actually provid-ing the feedback. Unlike the typical 360-degreeprogram, however, the procedure outlined in Fig-ure 1 provides for feedback both to the receiversand providers. The receivers get to know how theyare perceived by their constituents and the provid-ers learn about how well they performed their roleas raters. Data relating to undesirable rating ten-dencies, such as halo, central tendency, leniency,and strictness, are revealed at this stage. Also,

2000 Ghorpade 147

systematic and structured analysis of the contentsof the qualitative comments can be done underdifferent dimensions of performance (e.g. techni-cal, interpersonal, and managerial competencies)and the results relayed to the feedback receivers.

Appointing the 360-Degree Administrator

A host of administrative challenges arises whenthe decision is made to engage in 360-degree feed-back. A critical issue here is who should gather,process, and have custody of the feedback data. Aparadox is encountered relating to this issue:

Managerial Involvement Paradox

Managerial involvement in gathering and process-ing 360-degree feedback data is legitimate andinevitable. However, involving persons in author-ity may taint the process and reduce its credibility.

While recognizing the legitimacy of managerialinvolvement in 360-degree administrations, it isalso important to recognize that such programsmay cause fear. They invite judgments on individ-ual performance by a wide range of participants,and violations of confidentiality cannot be con-trolled. In fact, so great is the potential of theseprograms for causing fear that a Price Waterhousesurvey referred to earlier advised management toplan for paranoia.^^

Given this possibility, managers need to makethe program appear as impartial and fair as pos-sible. Here are two potential solutions to selectinga 360-degree administrator:

Solution 1: Assign the administrative tasks to amanagerial role such as manager, HB director, orsupervisor. While this is a common practice, itsappropriateness within a particular context hingesentirely on the extent to which the person selectedis trusted by the feedback receivers.

Solution 2: Assign the administrative tasks to aperson whose involvement in the process can en-hance trust. Given the inherent potential of 360-degree programs for causing fear, the person incharge of the data gathering and analysis must beperceived as trustworthy^^ by the feedback receiv-ers and providers and must have the technicalcompetencies to help the parties resolve themethod issues encountered in conducting 360-degree programs such as those depicted in Figure1. Such individuals are hard to find in the chain ofcommand of organizations. Persons who occupysuch conflict management roles as ombudsman,mediators, and organizational development pro-fessionals may have greater credibility, and hencemay be considered for this role. In one unionized

company where multisource feedback was to beprovided to supervisors, nominations for this rolewere sought directly from the supervisors them-selves. The supervisors chose to delegate the ad-ministration of their program to a committee con-sisting of an HR representative, the union steward,and a supervisor from another area who was not apart of the feedback program.

If a 360-degree administrator cannot be foundinternally, outside consultants can be hired. How-ever, the fees for such help are usually paid bymanagement and hence they may not be trusted byemployees.^"^ Therefore, if this avenue is chosen,the consultant should be retained on a long-termbasis and continued only if he or she is able tobuild trust among organizational members. This,in fact, was the avenue that was chosen by thecompany in the case presented below.

How One Company Coped with the FiveParadoxes

The organization in which the 360-degree feedbacktook place began as an assembly plant of an Asianmultinational electronics manufacturer located onthe West Coast, and then expanded its businessoperations to include assembly and manufactureof various components. Starting with about 100workers, the firm grew to over a thousand employ-ees within five years. The growth, however, camewith a price. The management team, which ini-tially consisted of an Asian general manager andtwo Americans overseeing various work units andtheir supervisors, quickly grew to include 15 man-agers, with a mix of Asian and U. S. managers.While the company continued to grow, there wasconsiderable tension within the managerial ranksat all levels. Rapid changes in technology, increas-ing ethnic mix of the labor force, and differingmanagement styles led to process inefficiencies,high turnover, and worker dissatisfaction. A laborunion initiated an organizing drive.

The general manager decided to act. He assem-bled the senior managers, the HR director, and anorganization development (OD) consultant whohad worked for the company for some time per-forming developmental and conflict resolution ac-tivities. Many options were explored, includingorganizational restructuring, executive develop-ment, and sensitivity training. After much deliber-ation, it was decided that a good starting pointwould be for the top managers to try to understanddifferences within their own rank. Multisourcefeedback was selected as a way of initiating thisprocess. The OD consultant was placed in chargeof this effort, as he had done projects for just about

14B Academy of Management Executive FebTuary

all the managers in the group and they had gottento know him and felt comfortable with him.

The 15 middle managers were selected as thefeedback receivers of this program. Right at thebeginning, the general manager committed him-self publicly to making the feedback effort devel-opmental rather than evaluative. This was an im-portant variable in unfreezing communicationlinks.

The OD consultant then began the task of assem-bling the behaviors that were to form the basis ofthe feedback. First, the 15 managers were inter-viewed individually in an open-ended mannerabout their personal view of the situation. Thisyielded about 30 context-specific behaviors focus-ing on peer relations and dynamics. Examples ofsuch behaviors were: Keeps commitments made topeers, keeps me in informed about changes thatconcern me, cooperates in getting things done,trustworthy, professionally competent, and knowsjob well. To this list were added 15 generic mana-gerial behaviors from the management and lead-ership literature. Examples were: Planning andorganizing, controlling, problem solving, and inter-facing.

The 15 managers were then asked to rate thebehaviors for relevance to their roles and situa-tions on a rating scale that posed the followingquestion: To what extent is this peer behavior ortrait of concern to you in your job situation? Therating was done on a five-point Likert-type scale,where 1 was linked to a descriptive anchor of "Noconcern at all," and 5 was anchored to "A greatdeal of concern."

The mean scores of each of the 45 behaviorswere tabulated and ranked. Consultation with themanagers and the general manager led to elimi-nation of 15 items that had received low scores(below 3) and that had high standard deviations,signifying low levels of agreement among raters.As might be expected from the discussion of Para-dox 4, in this highly charged emotional atmo-sphere, these turned out to be mostly generic man-agerial behaviors such as planning, organizing,and controlling. The 30 remaining behaviors werereflective of the real problems faced by the man-agers. These behaviors were then anchored to apeer description scale where the lead questionwas: How descriptive is the following behavior ortrait of this manager? Again, a five-point Likert-type scale was used, with 1 labelled as "Not at alldescriptive," and 5 as "Very descriptive."

Multisource ratings were acquired. The generalmanager was rated by all the senior managers,who in turn were rated by the general managerand all their peers. Along with rating their peers.

participants were asked to provide a self rating forthe 30 items. They were also asked to justify alltheir numerical ratings with qualitative commentsand explanations. Because much emotion hadbuilt up over the months preceding this exercise,participants poured out their feelings, both posi-tive and negative. However, as might be expected,not all feelings expressed were useful or construc-tive. The OD consultant placed the qualitativecomments on cards and sorted them into piles thatcontained similar sentiments. These were summa-rized and the wording was changed to protect in-dividual identities.

Feedback to the participants, which included thegeneral manager, was provided individually bythe OD consultant. Feedback sessions lasted fromone-and-a-half to three hours. During these meet-ings, the OD consultant began by providing therecipients with a statistical summary of the resultsthat covered mean scores the candidate receivedfrom peers on the 30 items, a comparison of peerand self scores, and a summary of qualitative feed-back from peers. Participants were asked not todiscuss the results personally with anybody and toponder the feedback for two weeks on their own,with the OD consultant standing by to help themwith clarification and process issues. A meetingwas called by the general manager at the end oftwo weeks. The purpose of this meeting was not todiscuss substantive findings of the exercise butrather to assess feelings about the value of theprocess and to decide on a future course of action.

All participants stated that they had gainedsome value from the exercise and that they wantedthe process to continue. Four follow-up actionswere taken: (1) Process meetings—all managersagreed to meet twice a month for a period of threemonths to explore quality issues, without gettinginto personal issues; (2) expansion of feedbacksources—all the managers asked for and were pro-vided with feedback from the unit heads and su-pervisors under their command using the processdescribed above; the supervisors in turn were pro-vided with feedback from their peers and employ-ees; (3) clarifying meetings—^in a few cases, meet-ings were arranged among peers to discusssubstantive issues raised in the feedback exercisewith the involvement of the OD consultant; and (4)individual development—managers used the peerand subordinate feedback to seek out specializedreading materials, training programs, and otherforms of professional development assistance.

The employee development paradox was re-solved by a firm commitment by the general man-ager to making development the objective of the360-degree feedback. The multiple constituents

2000 Gborpade 149

and anonymous ratings paradoxes were managedby the OD consultant's serving as the filter of thefeedback provided to the managers. While the fil-tering worked as intended in this case, it needs tobe recognized that the potential exists for suchpower to be misused. It may be useful to have morethan one person involved in this filtering processto enhance trust. The structured feedback paradoxwas managed by involving the recipients in select-ing the behavior items included in the feedbackinstrument. And the managerial involvement par-adox was resolved by delegating the administra-tion of the 360-degree feedback to an outsider whohad built trust with the participants.

Should Organizations Use 360-Degree Feedback?

It would be difficult to argue against the generalnotion of multisource feedback in today's businessclimate. Corporations have decentralized theirmanagement systems and considerable impor-tance is placed on teamwork with the role of themanager, particularly the middle manager, beingcloser to that of a team leader than that of anofficer in the traditional bureaucratic sense. In thiscompetitive context, it would be difficult for anymanager in any complex organization to go verylong without receiving some feedback from themultiple constituencies that the role serves. The360-degree concept enables such feedback at arelatively low operating economic cost. Researchindicates that the gains from 360-degree feedback,when used as a developmental tool, are substan-tial. Changes in behavior brought about by suchprograms tend to be immediate and frequentlydramatic.

Research indicates that the gains from360-degree feedback, when used as adevelopmental tool, are substantial.

The 360-degree concept, however, is not withoutproblems. In fact, its application is fraught withparadoxes that cannot be resolved through me-chanical, technical actions. In order to gain thebenefits that the process offers, and minimize theadverse consequences, organizations need to cre-ate an atmosphere of trust, openness, and sharing.Feedback is most meaningful when there is a gen-uine desire on both sides for a meaningful andauthentic exchange of perceptions.

Endnotes

' Handy, C. 1994. The age of paradox. Boston: Harvard Busi-ness School Press, 11-14.

Mbid., 34-36.^ Yammarino, F. Y., & Atwater, L. E. 1997. Do managers see

themselves as others see them? Organizationai Dynamics,Spring: 35-44; Antonioni, A. 1996. Designing an effective 360-degree appraisai feedback process. Organizational Dynamics,Autumn: 24-38.

^ Handy, C. 1989. The age of unreason. Boston: Harvard Busi-ness School Press. See particularly his description of the Sham-rock Organization in Chapter 4.

^ London, M., & Beatty, R. W. 1993. 360-degTee feedback as acompetitive advantage. Human Hesource Management, 32 (2 &3): 353-372.

^ Antonioni, op. cit., 28. For other empirical research on thisissue, see: London, M. & Smither, J. 1995. Can multisource feed-back change perceptions ol goal accomplishment, self-evalua-tions, and performance related outcomes? Personnel Psychol-ogy. 48: 803-839; Waldman, D. A., Atwater, L. E. & Antonioni, D.1998. Has 360-degree feedback gone amok? The Academy ofManagement Executive, 12(2): 86-94.

' Lepsinger, R., & Lucia. A. D. 1997. The art and science of360-degree feedback. San Francisco: Jossey-Bass, 17.

^ For variations on practice on this issue, see: Edwards, M. E.,& Ewens, A. J. 1996. Providing 360-degree feedback. Scottsdale,AZ: American Compensation Association; Lepsinger & Lucia,op. cit.; Tornow, W. W., & London, M. 1998. Maximizing (he valueof 360-degree feedback. San Francisco: Jossey-Bass.

^ An entire conference dedicated solely to 360-degree feed-back issues was organized in June 1997 by the InternationalQuality & Productivity Center, Audio cassettes of this confer-ence can be obtained from IQPC, 150 Clove Road, P. O. Box 401,Little Falls, NJ 07424-0401. Also see: Hegarty, W. H. 1974. Usingsubordinate ratings to elicit behavioral changes in supervisors.Journal of Applied Psychology. 1974, 59(6), 764-766; Hoffman, R.1995. Ten reasons you should be using 360-degree feedback.HfiMagazine, April: 82-85.

'"London & Beatty, op. cit., p. 355; also see, Schneier, C. E.,Shaw, D., 8E Beatty, R. W. 1991. Periormance measurement andmanagement: A new tool for strategy execution. Human Re-source Managemenf, 30: 279-301.

" Variations in 360-degree practices are discussed in: Anto-nioni, op. cit.; London & Beatty, op. cit.; Yammarino & Atwater,op. cit.

'̂ Antonioni, op. cit.'̂ London & Beatty, op. cit., 353.'" Kluger, A. N., & DeNisi, A. 1996. The effects of feedback

interventions on pertormance: A historical review, a meta-anal-ysis, and preliminary feedback theory. PsychoJogica/ Bulletin,119: 254-284.

'^Excellent reviews ol the 360-degree feedback issues andresearch can be found in: London & Beatty, op. cit.; Antonioni,op. cit.; Yammarino & Atwater, op. cit.; and Waldman, op. cit.

'̂ All four sources used as points of reference in this article(see 15) support placing the emphasis on development in 360-degree programs: Waldman, et. al., op. cit., 86; Antonioni, op. cit.26; London & Beatty, op. cit., 359; Yammarino & Atwater, op. cit.,36.

"Antonioni, op. cit., 26.'̂ Price Waterhouse, Periormance measurement: The rating

whirl, http://www.govexec.eom/features/0297s2.h, February,1997, 4-5.

'̂ Waldman. et al.. op. cit., 87.°̂ Antonioni. op. cit.; and Lepsinger 8t Lucia, op. cit.'̂ For additional ideas on conducting sharing and clarifying

meetings, see: Lepsinger & Lucia, op. cit.. 179-181.^̂ Waldman. et al.. op. cit., 87-89.^̂ For additional ideas on combining 360-degree with perfor-

150 Academy of Management Executive February

mance appraisal, see: Waldman, et. al., op. cil.; and Tomow &London, op. cit., 68-73.

^̂ For reviews of early research on multiple raters in perfor-mance appraisal see, Cascio, W. F. 1997. Applied psychology inhuman resource management [5th ed.). Englewood Cliffs, N]:Prentice Hall, 62-65.

" For an empirical development of feedback items from dif-ferent sources, see: Herold, D. M., & Parsons, C. K., 1985. Assess-ing the feedback environment in work organizations: Develop-ment of the job feedback survey. Journal of Applied Psychology,70(2): 290-305.

''^Campbell, D. J., & Lee, C. 1988. Self-appraisal in perior-mance evaluation: Development versus evaluation. Academy olManagement Review, 13(2): 302-314.

^' Data on relative competencies of managers and subordi-nates as raters can be found in: Bernardin, H. J., & Beatty, R. W.1987. Can subordinate appraisals enhance managerial produc-tivity? Sioan Management Review, Summer, 63-73; Greller,M. M. & Herold, D. M. 1975. Sources oi ieedback: A preliminaryinvestigation. Organizad'ona/ Behavior and Human Perfor-mance, 13: 244-256.

^̂ Cooper, W. H. 1981. The ubiquitous halo. PsychoJogicaJBulletin, 90, 218-244.

^^Miller, D, T.. & Ross, M. 1975. Self-serving biases in theattribution of causality: Fact or fiction? Ps/choiogica/ Bulletin,82: 213-225. Also see the discussion of fhe egocentric-bias theoryin, Harris M. M., & Schaubroeck, J. 1988. A meta-analysis ofself-supervisor, self-peer, and peer-supervisor ratings. Person-nel Psychology, 41: 55.

^° Campbell & Lee. op. cit., 306.^' Burns, T. 1969. Industrial man. Baltimore: Penguin Books,

232.^̂ Longenecker, C. O., Sims, H. P., & Gioia, D. A. 1987. Behind

the mask: The politics of employee appraisal. The Academy ofManagement Executive, 1: 183.

^̂ For a review of this literature, sae: Campbell & Lee, op. cit.;Longenecker, et. al., op. cit.; Gioia. D., & Longenecker. C. O. 1988.Delving into the dark side: The politics of executive appraisal.Organizafiona/ Dynamics, 22(3): 47-57.

" Lepsinger & Lucia, op. cif., 120-121.^Anfonioni, op. cit., 28. Also see, London & Beatfy. op. cit,,

366.^̂ Cascio, op. cit., 64-66.^̂ Ibid., 361.^^For more on criteria for evaluating feedback dafa, see:

London & Beatty, op. cit.. 364-366; Lepsinger 8E Lucia, op. cif.,62-66.

^̂ Lepsinger 8t Lucia, op. cit., 24-28. For additional ideas forimproving rating effectiveness through rafer training, see:Waldman. et al., op. cit., 91-92.

*''A technique known as frame-of-reference training has

proven promising in improving accuracy of appraisals. For areview of research on this technique, see: Cascio. op. cit., 75-76.

*' For a comparative review of research-based managementdevelopment scales, see: McCauley, C. D., Lombardo, M. M., &Usher, C. J. 1989. Diagnosing management development needs:An instrument based on how managers develop. Journal ofManagement. 15(3): 389-403.

*^ Information obtained directly from the company by tele-phone.

"Lepsinger & Lucia, op. cit., 180-181.•*•* London 8E Beatty, op. cit., 364.^̂ For variations of the approach recommended in this paper,

see: O'Neal. S. & Palladino, M. 1992. Revamp ineffective perfor-mance management. Personnei Journal, February; Yammarino& Atwater, op. cit.. p. 36.

^̂ McCormick. E. J., Jeanneret, P. R., & Mecham, R. C- 1969. Theposition analysis questionnaire. West Lafayette, Indiana: Occu-pational Research Center, Purdue University.

^̂ U. S. Department of Labor. 1977. Dictionary of occupationaltitles, 4th Edition. Washington, D. C: U. S. Government PrintingOffice.

*^ McCauley, et. al., op. cit." Fleishman, E. A. & Reilly, M. E. 1992. Handbook of human

abilities. Palo Alto, CA: Consulting Psychologists Press, Inc.^° Lepsinger & Lucia, op. cit.^' Antonioni, op. cit., 26-27.^̂ Price Waterhouse. op. cit., 6.^̂ For a discussion ol the role of interpersonal trust in the

feedback process, see: Ilgen, D. R., Fisher, C. D., & Taylor, M. S.1979. Consequences ol individual feedback on behavior in or-ganizafions. Journal of Applied Psychology, 64(4}: 340-371.

" For a discussion of the benefits and risks involved in thehiring oi consultants, see: O'Shea, J., & Madigan. C. 1997. Dan-gerous company.- The consulting powerhouses and the busi-nesses they save and ruin. Chicago: Times Business.

Jai Ghorpade is professor ofmanagement in the Coiiege ofBusiness Administration at SanDiego State University, He hasserved on the editorial board ofthe Academy of ManagementReview, and published articlesin the Academy of Managemen tJournal, Personnel Psychoiogy,and Journal of Applied Psychol-ogy. He has experience in theapplications of multisource per-formance feedback. Contact:Giiorpade@inaii.sdsu.edu.