the evolution of is job skills: a content analysis of is job … · 2016. 1. 22. · cluding...

Tracking IS Skills Over Time

The Evolution of ISJob Skills: A ContentAnalysis of IS JobAdvertisements From1970 to 1990

By: Peter A. ToddSchool of BusinessQueen’s UniversityKingston, Ontario K7L 3N6Canadatoddp@ post.queensu.ca

James D. McKeenSchool of BusinessQueen’s UniversityKingston, Ontario K7L [email protected]

R. Brent GallupeSchool of BusinessQueen’s UniversityKingston, Ontario K7L [email protected]

AbstractChanges in the knowledge and skill requirementsof information systems (IS) positions were ex-amined by analyzing the content of advertise-ments for IS professionals placed in four majornewspapers over the 20-year period 1970-1990.Three types of jobs were examined: program-mers, systems analysts, and IS managers. Theanalysis of the frequency of phrases in theseadvertisements suggests that job ads for pro-grammers have changed very little--technicalrequirements remain high, and business andsystems knowledge requirements remainrelatively low (although the frequency of mentionof business requirements has increasedsomewhat). IS management positions are also

relatively stable (as reflected in the makeup ofjob ads) from the standpoint that businessknowledge requirements have remained high,with technical and systems requirementsspecified less frequently. The greatest transitionin specified job requirements over this 20-yearperiod has occurred for systems analysts.Although this is perhaps not surprising, thenature of this transition is. Contrary to expecta-tions, the relative frequency and proportion ofstated technical knowledge.requirements in adshave increased dramatically, while the relativefrequency of business and systems knowledgerequirements has actually decreased slightly.

These results raise questions concerning the im-plicit understanding by academics and practi-tioners alike of the need for business knowledgeon the part of systems analysts end other IS pro-fessionals. Various interpretations of these find-ings are provided, and the implications for botheducation and recruitment are discussed.

Keywords: Information systems jobs, job skills,job advertisements, content analysis

ISRL Categories: EB, EH0106, EH0107,EH0101, EH0206, EH0208, IA01

IntroductionWhat are the requisite knowledge and skills forIS personnel? This is an often asked and impor-tant question. It is widely advocated that organi-zational, functional, and managerial skills arebecoming increasingly important for most infor-mation systems (IS) positions. Many studies havenoted that organizational and behavioral skills areimportant to systems analysts and IS managers(e.g., Cheney, et el., 1989; Leitheiser, 1992; Wat-son, et al., 1990). The IS curriculum, as sug-gested by the ACM and DPMA, places a strongemphasis on the importance of managerial/organizational knowledge and skills for IS pro-fessionals (DPMA, 1986; Nunamaker, et al.,1982). Similarly, trade journals contain articlesciting the increased need for IS professionals withsignificant managerial and interpersonal skills--cften bemoaning their absence (e.g., Bell andRichter, 1986; Davis, 1993).

Thus, the perception exists that a successful ISprofessional blends technical knowledge with a

MIS Quarterly/March 1995 1


sound understanding of the business while com-manding effective interpersonal skills. This paperuses job advertisements to provide a window onthe degree to which these skills are demandedin the marketplace and to look at, how those skillrequirements specified in ads have changed be-tween 1970 and 1990. Job ads serve asa primaryrecruitment vehicle (Walsh, et al.,’ 1975) and thusshould, we assert, reflect the nature of the skillsrequired of IS personnel in different positions.

Based on this assumption, this paper sets out todetermine the degree to which the perception ofchanging knowledge and skill requirements ismatched in job advertisements for three classesof IS professionals--programmers, systemsanalysts, and IS managers--over a 20-yearperiod from 1970 to 1990. The:intent is to in-crease our current understanding of the IS pro-fession in order to enhance the processes ofeducating, training, recruiting, hiring, and pro-moting IS professionals.

The paper is organized as follows. First, theprevious work on knowledge and skill require-ments for IS jobs is reviewed. Based on thisreview, our general expectations of howknowledge and skills required for IS positionshave evolved over time is then presented. Next,our approach to using job ads as a vehicle to ex-amine the nature of IS jobs is discussed. Final-ly, the results of the analysis and theirimplications for the IS profession and for ISeducation are discussed.

A Review of the IS Job SkillsLiteratureA number of studies have been conducted overthe past 20 years to examine .the skills andknowledge required of IS personnel. Thesestudies typically examine a variety of IS job posi-tions, ranging from programmers to managers,with the intent of identifying the type and mix ofskills needed for particular IS positions. Thesestudies examine skill requirements of program-mers, systems analysts, and IS managers overthe period 1970-1990.

ProgrammersVery few studies of skill and knowledge re-quirements of programmers were conducted in

the 1970s. The available literature consistedmostly of experienced practitioners commentingon what they thought the training background ofprogrammers should be. This debate identifiedthe need to sharpen technical skills of program-mers with in-house training programs (Anderson,1969), the need to upgrade private electronic dataprocessing schools to improve the technical ex-pertise of programmers (White, 1970), and theneed for improved communication skills for pro-grammers (Roark, 1976). Thus, the limitedliterature that exists from the 1970s primarilydiscussed the importance of technical skills forprogrammers.

In 1980, the requirement that a programmerpossess good specialist/technical skills (com-puter, hardware, software, and systems) was stillacknowledged, but a lack of generalist/mana-gerial (organizational, people, society) skills wasseen to inhibit the progression of programmersto managerial positions (Cheney and Lyons,1980). By the mid-1980s, a study of employers’requirements for entry-level IS positions (par-ticularly programmers) found that strongbusiness communication skills were more impor-tant than technical skills such as exposure tomicro-computer packages and application pro-gramming languages (Albin and Otto, 1987).

But in 1990, another study found that, for program-mers, the top skills were technical (applicationlanguages, systems analysis and design skills,database and operating systems), followed closelyby business knowledge and skills (Watson, et al.,1990). In summary, the literature shows a strongcontinuing emphasis on technical skills for pro-grammers, with business and communication skillsappearing to increase in importance over time.

Systems analystsSystems analyst skills also show changes overtime. In 1971, a survey of systems analysts foundthat technical and systems skills were highlyvalued, while managerial skills were consideredsecondary (Strout, 1971). By 1976, various commentators were calling for systems analyststo improve both technical and management skills(Roark, 1976; Seller, 1976). In 1980, a survey the skills required for various IS positions foundthat systems analysts required both goodsystems skills and strong interpersonal and com-munication skills (Cheney and Lyons, 1980).

2 MIS Quarterly~March 1995


By the mid-1980s, additional studies confirmedthe importance of both interpersonal/manage-ment skills and technical skills for the systemsanalyst (Harrison and Springer, 1985; Jenkins,1986). By the late 1980s and early 1990s, ISmanagers and systems analysts were saying thatspecialist/technical skills were decreasing in im-portance for systems analysts while generalist/managerial skills were increasing in importance(Cheney, et al., 1989; Green, 1989; Watson, etal., 1990). Recently, a survey of 107 MISmanagers found that interpersonal skills wereranked consistently higher for systems analyststhan were any other category of IS skills(Leitheiser, 1992).

In summary, the literature concerning skills re-quirements for systems analysts, while not dis-counting technical or system skills, appears toincreasingly stress the importance of com-munication and general business skills.

IS managersFor IS managers, skill requirements do not ap-pear to have changed much since 1970. In theearly 1970s, no surveys of skill and knowledgerequirements for IS managers were published.Commentators noted that IS managers should beprofessional managers and stressed the impor-tance of management and communication skillsover technical skills (Butt, 1969; Exton, 1970). 1976, a survey of IS managers reported thatmanagerial and interpersonal skills were most im-portant, while problems requiring technical skillscould be handled by subordinates (Joslin andBassher, 1976). In 1980, the picture was similar.A survey of IS executives found that IS managersshould have more generalist/managerial skillsthan specialist/technical skills (Benbasat, et al.,1980). This is confirmed by another study that ex-amined what practicing managers actually did.This study found that IS managers valued in-terpersonal and management skills more thantechnical skills (Ives and Olson, 1981). Anotherstudy, which asked IS managers what skills theyshould have, reported that both managerial andtechnical skill sets were needed (Cheney andLyons, 1980). A study that compared skill re-quirements of IS personnel (IS managers andsystems analysts) in 1978 with those still requiredin 1987 concluded that generalist/managerialskills were increasing in importance relative totechnical skills (Cheney and Lipp, 1987).

In summary, most of the findings over the yearsindicate that IS managers must have good in-terpersonal/managerial skills and that technicalskills, while important, are definitely secondary.

Research QuestionFollowing on prior research in this area, this studyexamines how the mix of skill requirements haschanged over the period 1970 to 1990, asreflected in job ads for programmers, analysts,and managers.

At a basic level, we argue that an IS professionalneeds to have a knowledge of informationtechnology, a knowledge of business and how tooperate in a business environment, and finally,a knowledge of systems and Systems approachesto problem solving. Thus, to provide a clearer pic-ture of skill requirements over time, we definedthe following three knowledge/skills categories(see Appendix for details): technical, business,and systems knowledge. The technicalknowledge~skills category includes specificknowledge and skills related to hardware andsoftware. This category is similar to the special-ist/technical category from previous studies butseparates systems knowledge and skills into adistinct category. Such a separation is similar tothe one provided in the ACM Curriculum Guide(Nunamaker, et al., 1982). The businessknowledge~skills category is similar to thegeneralist/managerial category of previousstudies and includes knowledge of industries andfunctional areas, management and organiza-tional skills, and interpersonal/communicationskills. While it is possible to make an argumentfor the separation of the interpersonal and com-munication skills into a distinct category, weargue that the incidence of such skills tends tocorrelate highly with the other skills within thebusiness category. The systems knowledgecategory captures problem-solving skills, in-cluding analytical and modelling skills as well asknowledge of development methodologies andsystems analysis/design tools and techniques.

For each job (i.e., programmers, analysts, andIS managers), we examine how job ads describethe requisite knowledge and skills for informationsystems personnel and use the ads to providea window on how those skills may have changedover time. In general, we argue that the IS field

MIS Quarterly~March 1995 3


as a whole has seen a relative shift from an ex-clusive emphasis on technical issues to arecognition of the importance of business-relatedissues. This shift seems apparent in the cur-riculum of IS programs, in the body of publishedIS research, and in a review of IS trade journals.Specifically, we see such a shift occurring in theIS job skills literature with increasing emphasisbeing placed on the importance of businessknowledge and skills. Our objective is to examinethe extent to which these shifts are, or are not,reflected in the content of job ads.

Therefore, our research question is:

Has the mix of job skill requirements, asspecified in job ads, changed over theperiod from 1970 to 1990 for (1) pro-grammers, (2) analysts, and (3) managers?

Data Collection and AnalysisTo address the research question, IS job adswere collected from two U.S. newspapers (TheWall Street Journal and The New York Times)and two Canadian newspapers (The Globe andMail and The Toronto Star) at five-yearintervals--1970, 1975, 1980, 1985, and 1990. Atotal of 1,634 ads were collected, which includ-ed ads for 581 programmers, 348 systemsanalysts, and 305 IS managers. The remaining400 jobs included EDP auditors, systems pro-grammers, database administrators, data entrystaff, and consultants, among others. Becausethe number of ads for these job types was small,no further analysis was performed. Instead, wefocused on the 1,234 programmer, analyst, andIS manager jobs. The distribution of these posi-tions across the five time periods is shown inTable 1. In general, the number of ads for latertime periods is somewhat greater, reflecting thegrowth in the field.

Figure I shows the average number of skills men-tioned in an ad. Information that was not of directinterest for this study, such as descriptions of theorganization, its culture, job benefits, and salary,was left out. Skill requirements specified in theads have been growing in number for all jobtypes. In 1970, the average ad mentioned slightlyover four distinct skills. By 1990, this had grownto approximately seven skill requirements. Also,there appears to be little difference in the number

Table 1. Sample Size by Job Title and Year(n = 1,234)

Year Programmers

1970 48

1975 80

1980 128

1985 154

1990 171

Total 581

Analysts Managers

27 29

50 30

81 82

103 92

87 72

348 305

of skill requirements mentioned by ads for dif-ferent job types.

A discussion of the role of job ads in the recruit-ment and hiring process, as well as a full descrip-tion of the method of collecting, classifying, andanalyzing the data, is included in the Appendixfor the benefit of readers wishing more detail.Table 2 shows how all phrases were classifiedinto one of three basic knowledge/skill categories:(1) technical (hardware and software knowledge),(2) business (functional, managerial, and socialknowledge/skills), and (3) systems (problem solv-ing and systems development knowledge/skills).The three general categories were further decom-posed into seven more-detailed categories. Boththese levels of analysis are presented in theresults section that follows.

ResultsThe findings are presented at two levels. F!rst,for each job, summary data are presented usingthe three categories described above (see alsoFigures 2 and 3). Second, a detailed look insidethe three categories is provided for each of thejobs (Tables 3, 4 and 5). Details on how thevalues shown in the figures were computed aredescribed in the Appendix.

Because the data are open to a variety of inter-pretations, they are summarized in three waysas: indicator data, raw data, and percentage data,with each highlighting a somewhat differentperspective.

¯ Indicator Data--shows the proportion of adsthat refer to each category at least once in thead. The indicator data shows the ad from theperspective of each skill included and takes

4 MIS Quarter/y/March 1995


01970 1975 1980 1985 1990

Years

Legend

Programmers

..... Analysts

’"" Managers

Figure 1.

into account the fact that the number of timesa job skill is referenced may not be indicativeof its overall importance to the job. In otherwords, the fact that a skill class is mentionedat all implies that it is important to the job, thefact that it is mentioned three, four, or fivemore times may not indicate any additionalimportance.

Raw Data--shows the average number ofphrases per skill category in the ad. Thisshows the actual composition of the ads, in-dicating the emphasis that is placed ondescribing different kinds of skills. In contrastto the indicator data, this measure may in-dicate the relative importance of a particularskill.

Percentage Data--shows the averagenumber of phrases per category expressed asa percentage of the total number of phrases

Length of Ads

in the ad. This best demonstrates the relativecomposition of the ads, facilitating Com-parisons between job types and over time. Thepercentage data is especially important sincethe ads have been increasing in length since1970.

In addition to the summary data, a window is pro-vided on the detailed composition of each of theskill categories. This gives the reader a look in-side the categories to better understand howtechnology, business, and systems skills arechanging. This is done by decomposing the threecategories into seven more-detailed categories(hardware, software, business, management,social, problem solving, and developmentmethods) described in Table 2. In turn, each ofthese seven detailed categories is broken downfurther into sub-categories. These detailed dataprovide a perspective from which to interpret the



Tabl~ 2, Classification of IS Knowledge/Skills

Class Category Description ACM Categories

TechnicalKnowledge

BusinessKnowledge

SystemsKnowledge

1. Hardware

2. Software

3. Business

4. Management

5. Social

6o ProblemSolving

DevelopmentMethodology

Mainframe, mini, and personal computers.Other devices such as storage devices,controllers, printers, and other peripheralsplus networks.

Application systems, operating systems,packaged products (such as database,graphics, word processing), networkingsoftware and languages.

Functional expertise (such as finance,marketing) and industry expertise (suchas retail, mining).General management skills includingleadership, project management, planning,controlling, training, and organization.

Interpersonal skills, communication skills,personal motivation and ability towork independently.

Creative solutions, quantitative skills~analytical modelling, logical capabilities,deductive/inductive reasoning, innovation.

Knowledge of systems developmentmethodologies, systems approach,

Computers

Organizations

People,Society

Models

implementation issues, operations andmaintenance issues, general developmentphases, documentation, and analysis/design tools/techniques.

Systems

data and to draw inferences about possible ex-planations for the results. Moreover, the detaileddata provide the reader with the opportunity todraw his/her own interpretations and inferencesabout the results. For each job type, the summarydata results (indicator, raw, and percentage) areprovided, followed by the detailed data.

Programmers--summary viewThe indicator data for programmers are shownin Figure 2a. References to technology appearin roughly 90 percent of the programmer ads; andthis has been relatively constant over time.References to systems knowledge have beenmore volatile and have showed a modest upwardtrend over the 20-year period. The percentageof ads including at least one reference to abusiness skill, however, increased considerably.

In 1970, only 30 percent of the ads mentioneda business skill, but by 1990, about 60 percentof the ads referred to at least one business skill.

The raw data for this group are shown in Figure3a, and the percentage data is shown in Figure3b. Figure 3a shows-that the frequency oftechnical knowledge phrases doubled from 2.2phrases per ad in 1970 to 4.3 phrases per ad in1990. In other words, while a 1970 job ad for aprogrammer indicated the need to know oneoperating system plus one programminglanguage (typically COBOL), an ad in 1990 in-dicated that a programmer was expected to haveskills in multiple operating systems and program-ming languages. Over this same period, theaverage number of references to businessknowledge almost tripled from 0.5 phrases perad in 1970 to 1.3 phrases per ad in 1990. The



2a PROGRAMMERS110

,®]PROGRAMMERS

1970 1975 1980 1985 1990

Technology 92 94 86 94 96

Business 28 38 41 58 60

System 54 56 76 62 68

2b SYSTEMS ANALYSTS

Technology

Business

System

SYSTEMS ANALYSTS(%)

1970 1975 1980 1985 1990

37 58 70 82 77

74 68 70 69 70

74 82 77 72 77

2c MANAGERS

MANAGERS(O/o)

1970 1975 1980 1985

Technology 52 53 43 48

Business 86 87 83 87

System 79 87 82 73

1990

65

92

80

Figure 2. Indicator Data(% Ads Specifying a Requirement at Least Once)



RAW DATA PERCENTAGE DATA

3b PROGRAMMERS

3c ANALYSTS "

ol

3d ANALYSTS

3e MANAGERS

’t

3f MANAGERS

Figure 3. Composition of Jobs by Skill Class



Table 3. Programmer Job Ads

N 48 80 128 154 171Year 1970 1975 1980 1985 1990

Knowledge/Skill Cagetories Number of Phrases

TechnicalHardware Mainfame 17 48 32 49 13

Mini 1 8 36 49 53Des~op 0 0 3 22 16Other 8 1 20 0 7

Software 2GL 7 13 19 16 163GL 20 45 53 53 944GL 0 1 0 18 36COBOL 20 42 60 73 77Database 2 16 34 63 117CASE 1 2 1 1 11Oper~ing systems 17 90 96 212 243Packages 2 0 4 10 26Other 10 0 0 0 17

Total Number of Technical Phrases 105 266 358 566 726

BusinessFunctional Industry specific

Function specificOther

Management General managementLeadership skillsOrganization skillsProject managementPlanningMonitor and controlTrainingOther

Social Communication skillsIndependent/motivatedInterpersonal skillsOther

Total Number of Business Phrases

0 8 10 26 168 11 12 14 80 0 0 8 18

1 0 2 3 31 2 10 6 131 0 0 1 60 0 1 3 30 0 0 0 01 0 0 0 21 0 0 1 03 1 4 4 9

0 4 14 29 512 10 17 22 235 2 4 25 271 16 11 22 34

24 54 85 164 213

Sy~emsProblem Solving Quantitative/logical

General problem solvingTechnical expertiseCreative/in novativeOther

Development AnalysisMethodology Design

ProgrammingImplementationOperations/maintenanceGeneral developmentGeneral technologyOther

Total Number of Systems Phrases

0 0 2 4 25 2 5 17 350 0 2 12 80 0 2 6 36 1 0 10 4

3 7 12 7 149 11 21 36 31

13 28 57 52 754 6 14 14 91 3 3 8 85 10 17 30 212 11 21 30 29

15 0 51 27 39

63 79 207 253 278



Table 4. Systems Analyst Job Ads

N 27 50 81 103 87Year 1970 1975 i980 1985 1990


TechnicalHardware Mainfame, 5 11 14 21 11

Mini : 1 3 20 25 12Desktop 0 0 0 17 9Other 0 1 7 6 15

Software 2GL 0 2 2 4 13GL 3 6 18 23 154GL 0 0 0 18 17COBOL 6 8 25 22 18Database 2 6 14 47 62CASE - 0 1 0 0 3Operating’systems 2 25 31 98 113Packages 1 1 9 10 29Other 1 0 24 26 2


BusinessFunctional Industry spec!fic


Management General managementLeadership skillsOrganization skillsProject managementPlanningMonitor arid controlTrainingOther

Social Communication skillsIndependent/motivatedInterperson, al skillsOther


10 106 95 5

1 12 51 01 31 00 00 00 3

2 103 72 33 13

37 69

21 17 1014 14 615 16 23

3 2 37 12 43 1 12 2 34 2 20 7 00 1 0

12 22 50

20 26 3611 11 511 15 129 31 18

132 179 173

SystemsProblem Solving Quantitative/logical

General problem solvingTechnical expertiseCreative/innovativeOther




23030

77972507

52

0112

12

10161414

417

725

123

0 2 16 9 243 4 34 1 04 28 9

8 20 2622 13 3418 14 1418 14 6

5 ¯ 3 125 25 3211 14 1156 67 48

180 214 209



Table 5. IS Manager Job Ads

N 29 30 82 92 72Year 1970 1975 1980 1985 1990


TechnicalHardware Mainfame 12 16 17 16 14

Mini 1 4 7 17 19Desktop 0 0 0 5 10Other 0 0 5 0 0

Software 2GL 1 0 2 3 13GL 7 8 12 12 64GL 2 1 0 5 4COBOL 3 5 8 13 10Database 0 2 12 21 30CASE 0 0 1 0 6Operating systems 3 2 14 27 46Packages 0 3 2 2 5Other 0 0 0 0 0


BusinessFunctional Industry specific


Management General managementLeadership skillsOrganization skillsProject managementPlanningMonitor and controlTrainingOther

Social Communication skillsIndependent/motivatedInterpersonal skillsOther


5 8 17 18 158 5 13 13 127 4 23 21 12

12 7 18 23 167 7 17 21 281 1 10 4 31 2 5 6 81 3 10 9 64 3 5 11 60 1 1 1 23 4 7 26 39

1 1 14 20 140 3 13 14 51 1 11 24 178 12 14 12 22

59 62 178 223 205

Sy~emsProblem Solving Quantitative/logical

General problem solvingTechnical expertiseCreative/innovativeOther




0 0 0 0 01 2 1 6 60 0 0 3 03 0 3 1 32 2 3 10 14

4 5 6 7 98 6 18 10 147 7 17 14 72 6 13 18 81 8 8 12 38 11 39 31 263 4 21 22 22

19 17 37 47 35

58 68 166 181 147



effect of this increase, however, is reducedbecause the absolute number of references issmall compared to the overall size of the ad. Thetypical ad for a programming position in 1990 in-cluded a single reference to an item in thebusiness knowledge category, a single referenceto an item in the systems knowledge category,and four references to items in the technicalknowledge category.

By adjusting for the increasing length of ads,Figure 3b demonstrates that little has changedin the composition of these ads over time. Theads indicate that technical knowledge remainsparamount (68 percent of the ads in 1970 and61 percent in 1990), with business knowledge (9percent in 1970 increasing to 19 percent by 1990)slightly lower than systems knowledge (23 per-cent in 1970 easing to 21 percent by 1990). Thus,while the business knowledge and skills specifiedin the ads have increased in absolute numbers,they have been overshadowed by the specifica-tion of technical requirements.

Programmers--detailed viewTable 3 shows the detailed data for programmerjob ads within each category. Technicalknowledge is subdivided into hardware and soft-ware and, as can be seen in the table, isdominated by software requirements. The growthin references to software is perhaps a reflectionof the rapidly burgeoning development in soft-ware technology. Actually, the software categoryaccounted for most of the increase in length ofthe ads over the 20-year period studied.

The shifts within the hardware category reflectthe emergence of non-mainframe computing plat-forms since 1970. The software category alsoreflects the technology shift toward distributedcomputing; the proliferation of operating systems;a progression through second, third, and fourth-generation languages; and the emergence ofnewer technologies such as database, CASEtools, and application packages. However,COBOL and other second and third-generationlanguages still made up a significant proportionof the technical requirements as of 1990.

Business knowledge is subdivided into threecategories: business (which includes industry andfunctional area knowledge), management (incor-porating management skills such as leadership,

planning, and controlling) and social/interper-sonal (including communication, motivation, andgeneral interpersonal skills). Of these, thegreatest growth was in the social category. Withinthis category, communication skills were not evenmentioned in our sample of 1970 ads for pro-grammers but grew to be the most frequentlymentioned phrase in this category by 1990 (seeTable 3).

Within the business knowledge subcategory,phrases are of two types: references to industry-specific knowledge (e.g., insurance, chemical,publishing, or transportation) and references tofunctional area knowledge (e.g., finance,accounting, or marketing). In 1970, if businessknowledge was mentioned, it was likely to be areference to a specific function. This trend wasreversed over the 20-year period. By 1990, if areference was made to business knowledge, itwas twice as likely to refer to industry experienceas it was to specific functional experience.

Under the management category, leadershipskills were the most likely requirements to befound in programmer job ads; and this did notchange in 20 years, although their incidence in-creased somewhat. However, of the sevencategories, management skills rank Seventh outof seven (in terms of frequency of mention) forprogrammers.

Systems knowledge is subdivided into problemsolving and development methodology. Refer-ences to both these subcategories increased infrequency since 1970 and, within these subLcategories, the most frequently mentioned skillrequirements were general problem solving anda knowledge of programming methodology.

Analysts--summary viewFigure 2b shows the indicator data for systemsanalysts. It suggests a significant change for theanalysts in terms of the proportion of adsreferencing technical skills over the 20-yearperiod. In 1970, about 35 percent of the adsreferenced the need for technical skills. By 1990,about 75 percent of the ads contained suchreferences. Moreover, as shown in Figure 2b, asof 1990, the need for technical skills is indicatedwith the same frequency in ads as is the needfor both business and system skills, which havetended to be more stable over time.



Figure 3c shows the ad profile for systemsanalysts (raw data). The pattern here shows ris-ing references to technical knowledge and skills,with only modest increases in the references forbusiness knowledge over the past 20 years.Thus, the analyst position, as reflected in the jobads, appears to have become more, not less,technical over time. The commonly discussed in-crease in emphasis on business and communica-tion skills was not as strongly reflected in the jobads as one might expect. These results aresomewhat perplexing and are certainly at oddswith those generally reported in the IS literature(see, for example, Cheney, et al., 1989;Leitheiser, 1992; Nelson, 1991; and Watson, etal., 1990). Figure 3d shows the percentage data.Between 1970 and 1990, the percentage ofphrases relating to technical knowledge in-creased from approximately 13 percent to 43 per-cent of the ad content. This proportional increasecame at the expense of both business knowledge(decreasing from 34 percent to 24 percent) andsystems knowledge (slipping from 53 percent to33 percent).

Analysts--detailed viewBy looking at the detailed data, it is possible todetermine exactly what skills employers aredemanding when advertising systems analystpositions and how that pattern has changed overthe years. These details are presented in Table 4.

Similar to the programmer ads, technicalknowledge requirements for systems analystswere dominated by software phrases as opposedto hardware phrases. There seem to be fewtrends here, other than the fact that software re-quirements, particularly as they relate tooperating systems and databases, were specifiedin more and more detail over time. As withprogrammers, ads for analysts are shifting awayfrom the older technologies and languages andmoving toward packages and newer technologies.

Examining the three subcategories within thebusiness knowledge category (i.e., business,management, and social), no dramatic changewas evident since 1970. Requirements forspecific business knowledge and skills aretypically industry-specific rather than function-specific. This was consistent from 1970 through1990. There has been growth within the sub-categories of management and social skills.

Specifically, within the ads, employers stressedthe need for communication skills to a greater ex-tent over time.

The systems knowledge category was broken in-to problem-solving and development method-ology. Problem-solving abilities were brokendown into four basic categories: general problemsolving, qualitative and logical abilities, technicalexpertise, and creativity. The data in generalshow that these attributes were mentioned in theads far less frequently than the other categories.Within the problem-solving category, generalproblem solving was most frequently mentioned;creativity was mentioned infrequently. Develop-ment methodology knowledge remained impor-tant over the 20 years studied. Categories ofgeneral development, analysis, and design arethe most frequently mentioned. "Generaltechnology" included generic references totechnologies such as knowledge of databasesand data communication concepts.

IS managers--summary viewIndicator data for the IS managers is shown inFigure 2c. Business skills were consistently men-tioned in about 85-90 percent of the ads overtime. Systems knowledge showed a similar leveland consistency of references over time. The in-cidence of references to technical knowledge in-creased somewhat over the past 20 years for ISmanagers. The raw data profile for IS managersis shown in Figure 3e. The increasing businessorientation of IS managers is evident in the data."Ads for IS managers in 1970 contained about tworeferences to business knowledge. This climbedthrough the 1980s to about three references.References to technical knowledge remained lowuntil 1985, and then increased between 1985 and1990. References to systems knowledge andskills remained constant at about two phrases perad since 1970.

Figure 3f demonstrates that, when adjusted forthe increasing length of these ads, little changedin their composition until 1980, when technologyil~creased at the expense of systems knowledge.In 1970, the ad profile for an IS manager was 20percent technical requirements, 40 percentbusiness requirements, and 40 percent systemsrequirements. By 1990, this profile had changedto approximately 30 percent technical, 40 percent



business, and 30 percent systems. Businessknowledge remained the most important categoryof knowledge mentioned since 1970 and did notchange in emphasis over time.

IS managers--detailed viewTable 5 shows the detail behind the IS managers’job ads. The first category, technical knowledge,is comprised of hardware and software. In thiscategory, there were some differences from theprogrammer and systems analyst ads--there wasnot the dramatic increase in references to soft-ware. References to both hardware and softwareincreased slowly. Again, the emphasis is onoperating systems and database software.

Under the second major category of businessknowledge, there was an increase in the statedrequirements relating to both management andsocial skills and a proportionate reduction in func-tional and industry-specific knowledge require-ments, The most important subcategory in termsof frequency of mention within the businesscategory is management. Within this subcategoryof management, the dominant references are togeneral management and leadership skills.References to leadership appeared to be increas-ing quite rapidly over time. In the social category,communication skills and general interpersonalskills were most often referenced. Over time, in-creasing references were made to communica-tion and interpersonal skills.

The two subcategories within :the systemsknowledge category are problem solving anddevelopment methodology. Relatively speaking,there were few references to problem solving forIS managers. References to knowledge ofdevelopment methodology were,, however, in-cluded within IS manager ads. Within thiscategory, the most frequently mentioned skillswere under the subcategories of general develop-ment and general technology.

Discussion of ResultsThis study has examined the skill!requirementsof programmers, systems analysts, andmanagers over the period 1970 to 1990 asspecified in 1,234 job ads found in fournewspapers. There are two key results from thisstudy.

First, not much changed. In absolute terms, therewas a significant increase in references totechnical requirements for both programmersand analysts and a modest increase formanagers (see Table 3). References in the adsto business and systems knowledge showedminimal change in absolute terms, though the in-crease in business requirements for managerswas apparent. When these results are examinedin relative terms, virtually all of the change for pro-grammers and IS managers can be accountedfor by "ad inflation"--that is, the average lengthof an ad has increased over the past 20.years.For programmers and IS managers, one canargue that references to technology, business,and systems was consistently apportioned overtime.

The second important result involves systemsanalysts. First, the indicator data show that thepercentage of analyst jobs referring to technicalskills is increasing over time. Second, in termsof the absolute number of references, there wasan increase in all three categories, though themost obvious increase occurred in the technicalrequirements. The average number of referencesto technical requirements more than tripled--from being the least referenced category in 1970to being the most referenced category in 1990.Even when the data is adjusted for the increas-ing lengths of ads, the technical references in-creased from about 13 percent of the ad in 1970to approximately 43 percent in 1990. Further,there were slight decreases in the proportion ofthe ads devoted to business and systems skills.

The changes in each job are discussed in moredetail in the next section.

ProgrammersTechnical requirements represent mostlyoperating systems skills, business requirementsmostly communications skills, and systems re-quirements mostly programming skills. The threelowest-ranking skills (in terms of number ofreferences in ads) are management skills,business skills, and problem-solving skills. Basedon this evidence, programmers (as characterizedby their job ads) appear to be specializedemployees who are required to understand infor-mation technology and general development pro-cess, and must be able to communicate withothers.

14 MIS Quarterly/March 1995


The increasing references to technical skillsseem reasonable, given that technical variety hasincreased due to the proliferation of softwarepackages and development platforms--from afew low-level languages in 1970 to a wide varie-ty of high-level application development en-vironments in the 1990s. As applicationgenerators make coding easier and moreautomatic, there may be less need for thedevelopment and enhancement of problem-solving skills. Furthermore, as jobs become morespecialized over time, it is less likely that pro-grammers will interact directly with users ondesign issues; rather, they will perform specializ-ed coding functions based on detailed specifica-tion and may indeed need to know little about thebusiness environment in which the systems areembedded. While it is fair to say that there hasbeen some increase in the demands for com-munication and other business skills among pro-grammers, the changes have not been dramatic.Further, there do not appear to be any univer-sally accepted business skills for programmers.

Approximately half of the programmer ads from1990 mention at least one business skill, an in-crease from 30 percent in 1970. However, thereis little consistency in the types of business skillsrequired. In 1990, though communication skillswas the most frequently referenced item in the

-business category, it was included for only 30percent of all programmer ads (see Table 3).Thus, it can hardly be considered a universal re-quirement. In fact, none of the 13 other detailedbusiness categories in Table 3 is mentioned inmore than 30 percent of the ads. This suggeststhat, while we can conclude that there has beensome increase in the need for businessknowledge by programmers, there is clearly noconsensus among organizations concerningwhat that knowledge should be. In short, the adssuggest that the work of programmers is becom-ing (both in technical and business terms) morespecialized.

Systems analystsThe requirements for systems analysts, asspecified in the job ads, appear to be very similarto those of programmers; that is, technical re-quirements relate mainly to operating systemsknowledge, business requirements relate to com-munications and related skills, and systems re-

quirements relate to development methodology(design) skills. The three lowest-ranking skills (asreflected by the number of phrases in ads) arehardware knowledge, industry and functionalarea skills, and problem-solving skills. Perhapsmost interesting is that the ad profile for ananalyst, which was quite different from the ad pro-file for a programmer in 1970, had become quiteSimilar by 1990. Technical skills appeared inabout 75 percent of the ads, in contrast to 40 per-cent in 1970. Furthermore, they were as likely tobe mentioned as were references to business,and they occurred with twice the frequency.Overall the systems analyst position, as de-scribed in job ads, appears to be getting more,rather than less, technical.

The data show a slight increase in the absolutenumber of references to business skills over time,though the percentage data shows a decline inthe proportion of ads devoted to business skills.Even factors such as communication skills, whichare of almost universally acknowledged impor-tance for systems analysts, appeared in only 40percent of the ads in 1990, and the need forstrong interpersonal skills appeared in less than15 percent of the ads. Specific business or func-’tional area knowledge appeared in less than 20percent of the ads (see Table 3). Thus, whilemore attention is being paid to the need forbusiness skills, there is, as with programmers,little consensus as to what that skill set shouldbe. Overall, the job ads indicate that the analyst’sjob requires a greater mix of skills than in thepast. However, it appears that increase is notcoming from increasing the business demandsplaced on analysts, but rather from increasing thelevel of technical knowledge required.

IS managersThe technical knowledge requirement for ISmanagers is predominately related to operatingsystems, the business knowledge requirementsare predominantly related to leadership, and thesystems knowledge requirements are predomi-nantly development methodology. The threelowest-ranking skills (as reflected by the numberof phrases in ads) are hardware knowledge,business skills, and problem-solving skills.

In absolute terms, there was an increasing de-mand for business skills reflected in the ads, butalso a surprising increase in the demands for

MIS Quarter/y/March 1995 15


tech ,nical skills. Business-related skills were con-sistently mentioned in approximately 90 percentof all ads over the past 20 years. Similarly,systems knowledge was consistently mentionedin about 80 percent of the ads, References totechnology increased from about 50 percent ofads in 1970 to about 65 percent of the ads in1990. Thus, both the number of technicalreferences and the proportion of ads they appearin has been increasing. At the same time, the adssuggest that technology is not universallyacknowledged to be important to the job of anIS manager.

As was the case for both programmers andanalysts, there is little convergence on thebusiness-related skill set that is specified in theads. In other words, the ads suggest no singlebusiness-related skill that is universally requiredof all managers. In 1990, the most frequentreference was to leadership, appearing in about40 percent of the ads, up from about 25 percentin 1970. Interpersonal and communication skillswere referred to in about 25 percent of the adsin 1990. Thus, there is no universally articulatedskill set for managers and little agreement acrossads on which business skills are important: Insummary, while the ads indicate that in 1990every manager required two to three businessskills, the same two or three skills were not be-ing demanded across jobs.

To summarize the findings across jobs, there issignificant data to support the argument that notmuch has changed in 20 years. The changes thathave occurred in the ads reflect increasing de-mand for technical skills in absolute terms acrossall three jobs. Further, while a case can be madefor an increase in the absolute demand forbusiness skills across all three jobs, the increasesare modest at best, and show little consistencyin terms of the business skills that are demand-ed, with no single skill being universally impor-tant. Thus, in spite of the tremendous changesthat have occurred in the IS field over the last 20years, in spite of the changes that have occurredin the nature of education and training, and inspite of the consistent assertion that businessand other "soft" skills are of increasing impor-tance to those working in the IS field, there is lit-tle evidence of significant change in the ads overthe past 20 years. Furthermore, those changesthat have occurred tend to suggest the need formore, not less, technical knowledge, particular-

ly for systems analysts. A number of possible ex-planations are explored in the .next section;however, in order to get a better understandingof the pervasiveness of the findings, reporteddifferences between industries are brieflyexamined.

Differences between companies, based on thenature of the business and their reliance on in-formation technology, were examined to see ifthey revealed differences in the skill setsdemanded. We felt that organizations that areinformation-intensive and more reliant on their ISmay have different ad profiles and may indeedrequest more business skills in a job ad. Toassess this, ads were grouped into two generalclasses--the service sector (including financialinstitutions, retail businesses, communicationscompanies, and other consumer services) andthe manufacturing sector (including resource ex-traction, and industrial and consumer manufac-turing). Because the service sector is moreinformation-intensive than the manufacturing sec-tor, a different emphasis may be placed onspecific IS knowledge and skills. Based on a splitsample of 778 service sector ads and 456manufacturing sector ads, we then reproducedthe analysis described above (see Figure 4 forthe raw data results by industry). The data in-dicate little by way of differences between set-.vice and manufacturing industries. With theexception of IS manager job ads (whose highvariation may be due to small sample sizes forthe manufacturing sector in some years), thevalues for the technical, business, and systemsskills tend to move together for the two industrygroups. Thus, we could find no apparent supportin the data for the argument that different skillsets may be required by different industrysectors.

Alternative ExplanationsGiven that these finding are somewhat unex-pected, and that they seem to be consistentacross industries, it is important to consider thealternative explanations for the results prior toconsidering their implications. One explanationfor these results would be that firms are not look-ing for people with more technical skills, butrather for people with the "right" mix of technicalskills--that is, familiarity with the particular blendof hardware, operating systems, and application



SERVICE INDUSTRY MANUFACTURING INDUSTRY

4a PROGRAMMER

4

4b PROGRAMMER

4c SYSTEMS ANALYST

=.4tt.$ I " " " " ’ " "

4d SYSTEMS ANALYST

4e MANAGER 4f MANAGER

=" .].’~ -- ’ ", , "

Figure 4. Composition of Jobs by Skill Class--Service and Manufacturing Industries (Raw Data)



development tools that the organization has inplace. Further, that mix is likely to differ for eachorganization and thus, must be clearly artic-ulated. In other words, in 1990,.more phrasesmay have been needed to accurately describethe technical environment in which the applicantwould work, whereas 20 years ago, it may havesufficed to specify a "COBOL shop." It is clearthat the proliferation of technology has led to awider variety of packages and more options fororganizations in terms of technica! infrastructure.In this sense, firms are becoming more distinctand specialized in their technical requirements,and the job ads simply reflect this. In short, the"right mix of skills" explanation infers thatorganizations believe that specific skills, inspecific technical platforms, have been and arecritical skill requirements. Thus, it might beargued that jobs are not requiring more technicalknowledge, but rather more specialized technicalknowledge.

A second possible explanation f(~r the apparentincrease in the demand for technical skill re-quirements reflects the notion that IS personnelmust become more technical because the pop-ulation of users they support are becoming moresophisticated. As "suppliers" of informationtechnology, IS personnel need to increase theirlevel of Sophistication simply to stay ahead of theusers. This argument appears to be plausible.Certainly, the level of technical proficiency amongusers has risen dramatically since 1970. IS staffmust be much more technologically oriented thanin the past, in order to maintain credibility (Davis,1993). Furthermore, information, requirementsdetermination is often being carried out bymembers of the functional areas rather than byIS personnel. As this work is offloaded, the ISstaff are left with the more technical developmen-tal tasks. Given this, it would not be unreasonableto expect some increase in the demand fortechnical skills among IS personnel.

A third potential explanation ’.is that moredemands for technical skills have resulted froma shift in the labor supply pool. This argumentis based on the notion that the composition ofpeople entering the IS labor pool has changedover the past 20 years and that ads have changedto reflect that. As IS programs within businessschools reach maturity, people are being trainedwith better business-related skills and com-paratively weaker technical skills (relative to those

coming from traditional computer science pro-grams). The supply side argument says that adswould include requests for the scarcer or moredifferentiable skills. Thus, today, if most peoplein the IS labor pool have the requisite businessand interpersonal skills, these skills do not needto be included in thead. Technical skills that aresomewhat lacking are included. In short, thesupply side argument suggests that IS jobs arenot changing, but the mix of the skill pool is, forc-ing organizations to focus more closely ontechnical skills in the hiring process.

While this argument undoubtedly has somemerit, we are not convinced that it is wholly sup-ported by our data. First, since we were not ex--amining strictly entry-level positions, the relativeimpact of education might be moderated. Sec-ond, if the supply side argument held, and weassumed an increasing business proficiencyamong the labor pool, then we should have seena significant decrease in references to businessskills in the ads themselves and a significantdecrease in the frequency of practitioner andacademic articles decrying the paucity ofbusiness skills among IS employees. Neither hasoccurred. In fact, both increased over the period1970-1990. This explanation, then, may be usefulin explaining changes in entry-level positions, butis less likely to help explain the increase intechnical skills across the entire sample, forwhich approximately 80 percent of the jobsspecified some amount of prior experience.

A fourth explanation may be that, while jobs arenot becoming more technical, it is the technicalaspect of the job that attracts candidates.Organizations may deliberately try to make theirtechnical environment appear sophisticated inthe belief that this is what prospective applicantsare looking for. Thus, rather than using thetechnical description to find people with the righttechnical skill mix, ads are laced with technologyto lure people who are excited by technology.This explanation may be plausible, but still in-dicates that the people they are attracting arepeople primarily interested in technology. Fur-ther, the attraction argument may make sensein a tight labor market, which may have existedin the 1970s but is less compelling in the leanerorganizational environment of the 1990s.

A fifth explanation may be that the ads do notreflect the desires of senior managers, but rather



the views of lower-level recruiters who may viewtechnical skills as being more important overall.There may be a gap between what senior mana-gers believe are the important skill sets for thefirm (as is reflected in the studies conductedpreviously) and what actually gets placed in a jobad. Thus, while IT managers might specify oneset of skills for a certain job, those who composeand place ads may suggest a different set of skillrequirements. Each individual may suggest therequirements from his/her unique perspectiveand may genuinely differ in what he/she thinksare the critical success factors for IS jobs.

A sixth possible explanation may be thattechnology is easy to incorporate into an ad butthat business and management skills are not. Itmight be argued that it would take severalphrases to adequately describe the technical en-vironment, but one needs to mention motivationor leadership only once without repetition. Whilethis explanation may have some merit, it wouldbe more easily supported if there was significantconsistency across organizations in specifyingcertain business skills. As noted above, this isnot the case. As an example, 12 business-relatedskills are listed in Table 4 for analysts. The mostfrequently mentioned of these skills, communica-tion, appears in only 40 percent of the ads. Thus,while each organization mentions one or twobusiness skills, the set is not consistent. This runscounter to the notion that there is little to sayabout business skills and is more likely indicativeof the fact that organizations are unclear aboutwhat to say. However, organizations may usemechanisms other than ads, such as interviews,to establish a job candidate’s business aptitude.Thus, companies may omit these qualificationsin the ad and look for them, instead, in the con-text of an interview and at other stages in the hir-ing and recruiting process.

It is clear that the job ads contain increasingreferences to technical skills, particularly forsystems analysts. This is likely reflective of theincreasingly sophisticated and diverse technicalenvironment that firms are working in today. Eachof the explanations offered above probably pro-vides some insight into the findings reported inthis paper, but it is unlikely that any one of themcompletely explains the findings. Further, thereare undoubtedly additional explanations that mayshed some light on this result. The fact remains,however, that in a time when both the research

and practitioner literatures are strongly emphasiz-ing business skills, iob ads are increasingly em-phasizing technical skills. This apparent paradoxis clearly worthy of additional investigation.

LimitationsBefore considering the implications of theseresults it is important to acknowledge the limita-tions of the study. This study uses a uniquemethodology to examine an important and often-asked question: What are the requisite knowl-edge and skills for information systems person-nel and how have those skills changed over time?The results of this study provide a challenge tothe conventional wisdom that business skills arebecoming increasingly important. Job ads, overthe period from 1970-1990, in fact, showed littlechange. While the results, as discussed above,are open to a variety of interpretations, it is im-portant to also consider the limitations of this ap-proach before reaching any specific conclusion.

It is reasonable to question whether newspaperads constitute an accurate reflection of the natureof actual IS jobs. While we would argue that adsshould reflect organizational skill requirements,it is also clear that organizations may employ adsfor other purposes, such as image enhancementand self-promotion. We see little reason, how-ever, why these purposes would conflict with thespecification of skill requirements within the ad.Thus, while there are a variety of purposes forwriting a job ad, it seems reasonable that (1) thereis no discernible benefit (of which we are aware)to an organization in deliberately misrepresent-ing the job being advertised; (2) because of therole ads play in presenting a portrait of the"ideal" candidate, they would tend to over-represent rather than under-represent this idealin the ad; and finally (3) even if certain skills arebest evaluated during interpersonal interviews,it is still in the organization’s interest to advisepotential applicants of the importance of this skillby including it in the ad. Nevertheless, just as wewould not expect that managers’ perceptions ofskill needs perfectly reflect actual skills demand-ed of employees, job ads may have similarlimitations.

Other limitations of this study relate to the natureof the sampling procedures. First, we sampledonly four newspapers and chose papers that



tend to be national in scope. While we know ofno specific reasons why there should .be regionaldifferences in the nature of skills required, itshould be cautioned that these ads are notnecessarily reflective of the population of all ads.Furthermore, since only a small proportion of jobsare actually advertised in newspapers, the poolof ads may not represent the entire populationof IS jobs. At the same time, the sample weemployed is significantly larger than those usedin previous studies of job skills and provides apicture of how jobs have changed over time.

One may question the precise Coding schemeemployed in this study. However, to the extentthat we separated specific technical skills frommore general systems skills, we did, if anything,bias the sample against the finding of a move-ment toward technical demands and away frombusiness demands. In addition, much of thedetailed coding data has been presented inTables 3 to 5, allowing readers to draw their owninferences.

Finally, it should be remembered that ads do notdescribe the actual candidate who was eventuallyhired. This candidate may not resemble the adin significant ways. An avenue of further researchin this area is to discover how closely the suc-cessful job candidates match the job descrip-tions. In spite of these limitations, implications forrecruitment and education can be derived fromthis work; these implications are presented in thenext section.

ImplicationsAssuming the evidence presented in this paperindicates that organizations are !ndeed lookingmore to technical skills, or at least to incumbentswith a more diverse set of technical skills, andthat the demand for business skills, as stated inthe job ads, has not really changed dramaticallyover the past 20 years, there are a number of im-plications for both management and educationthat should be considered.

Two views of these results may be taken. The firstis that we are experiencing an "education gap";the second is that we are experiencing a "recruit-ment gap." An education gap implies that educa-tion efforts may be misdirected to the extent thatthey focus on business skills in their curriculum.It would argue that technical skills are importantand that educational programs should emphasize

those skills. A recruitment gap would suggestthat, despite the recommendations of theacademic community and despite the writings inthe trade press of the importance of business-related skills, the recruitment process still has notresponded and is misdirecting its attention towardtechnical skills. This implies that business skillsare indeed critical and that organizations needto change the focus in their hiring practices toemphasize those skills. 1 In short, the remedy inone case requires a shifting of educationalpriorities and in the other the shifting of organiza-tional priorities. As always, the truth of the mat-ter probably lies somewhere in the middle.Nevertheless, it is instructive to look at theevidence in support of each side of this argument.

The education gapFirst, consider the educational gap. Shouldteaching priorities focus more on technology andrelatively less on business skills? Most would like-ly agree that MIS programs in universities haveshifted from being largely technologically focusedto being more business focused over the time-frame for this study. Has the shift been too ex-treme? The arguments in favor of the need formore technical skills can be supported by anumber of factors. As mentioned above, workersin general are becoming more technicallysophisticated, and thus, it is important that IS per-sonnel be trained to a level that keeps themahead of those they are trying to service. Thisis important to ensure that IS can both make acontribution and maintain credibility. Second,much of the work that requires business skills isbeing offloaded to end users or is being per-formed by joint project teams, where users fromthe functional areas bring the requisite businessknowledge to the project. Third, the pressuresfacing organizations today imply that they needworkers who will be immediately productive in theorganization and not require long periods of train-ing to become proficient in a new technical en-vironment. The absence of communication andinterpersonal skills or functional businessknowledge may be undesirable but perhapsviewed as more manageable for organizationsthan a lack of technical skills. As noted in a re-cent article, "For all the talk about soft skills, nocompanies are racing to hire humanities or socialscience graduates as entry level IS employees"



(Davis, 1993, p. 30). In short, the arguments fora technical emphasis are that knowledge oftechnology is a necessary condition and con-stitutes the distinctive competence for ISemployees and thus, should be given special em-phasis in the hiring process.

This position has clear implications for theteaching and training of IS professionals. Takingthe data at face value, technical prowess appearsto be the paramount concern. Prospective ap-plicants faced with ads filled with technicalphrases receive a very clear message--tech-nology is important. Those among us whoeducate students for IS positions might be welladvised to arm our students with technology sothat they will compete favorably in such amarketplace. Knowledge of business, problem-solving skills, and general management skillsmay be relegated to the "nice to have" category.These sorts of skills may help organizationsdecide between two aspiring applicants whootherwise appear equivalent. Programs unwillingto train students in technical skills may find theirstudents at a loss in the job market and certain-ly in the entry-level job market. Proof of this maybe revealed in a study of hiring practices over thelast few years. Do organizations look to thebusiness schools for IS professionals or to com-puting science programs? Has this patternchanged over time? Answers to questions likethese would provide additional insight into thisarea of study.

The recruitment gapNow consider the notion of a recruitment gap.This suggests that an emphasis on technicalskills is misguided and that there should be morefocus on business and interpersonal skills for ISemployees in job ads and in hiring. Certainly, thisposition is supported by the bulk of the academicand practitioner literature that discusses requisiteIS knowledge and skills (see Nelson, 1991, fora summary). Given this, it is surprising thatorganizations devote so much effort to describ-ing their technical needs. Further, given theirspecialized nature and the rapid changes intechnology, specific skill sets may quicklybecome obsolete. Thus, it would behove theorganization to focus on general technical skillsand, more importantly, on the ability to learn andadapt to new technologies rather than to focuson the specifics of any one technology. In short,

the emphasis on specific sets of technical skillsas a necessary condition for hiring an employeemay be misguided. Furthermore, given the diver-sity of technical platforms, it seems unlikely thatorganizations will find exact matches in terms oftechnical skills, and, therefore, to employ thoseas an initial screening criteria might unwisely limitthe pool of qualified applicants.

In addition, it is clear that IS is becoming morebusiness oriented. IS managers are forced tofocus on bottom-line responsibility, and increas-ingly the IS group is a significant contributor tothe development and implementation of organiza-tional strategy. Thus, it seems unlikely that theneed for business and systems skills is waning.There are frequent expressions by academicsand practitioners alike for the need for IS profes-sionals to have these skills. Smith and McKeen(1992) point out the common malady of "usersnot being able to talk to IS personnel," "IS per-sonnel being in their own world," "IS personnelnot understanding anything about the business,"and "IS personnel without effective inter-personalskills." The "hard demand for soft skills" amongIS personnel has also been recognized in thepractitioner literature (Davis, 1993). The need forbusiness skills is at the heart of the CACM andDPMA curriculum guidelines. Further, surveys ofIS professional and users have indicated theneed for improvements in business and organiza-tional knowledge and skills among IS personnel(Nelson, 1991).

Given this evidence, it seems reasonable toargue that it is the recruitment process itself thatmay be in need of change. It may well be thatsenior members of the IS organization valuebusiness skills but that those views are nottranslated to the lower levels of the organizationwhen it comes to recruiting practice. This is, inpart, supported by the observation that manage-ment positions show the most movement towarda business emphasis. It is for these positions thatwe might expect to see senior management takea more active role in the recruiting process. Tak-ing this view, we would argue that organizationalrecruitment practices may require some adjust-ment, and there is, atthe very least, a mismatchbetween what organizations claim to be impor-tant when asked about key skills and what thoseorganizations do in developing recruitingstrategies.

MIS Quarterly~March 1995


It is, of course, also possible that organizationsare looking for business qualities during the in-terviewing process since they are omitting theserequirements at the first screenir~g stage. On theother hand, because hiring decisions tend to bemade at relatively low levels in organizations andbecause the majority of the people doing the hir-ing are apt to be technically oriented, it is possi-ble that the phenomenon of ~"people hiringpeople just like themselves" is at work. Certain-ly, a highly technical person, who, after a fewyears on the job has acquired a sufficientknowledge of the practice of bu,siness from on-the-job experience, might feel that othertechnically oriented people could do the same.It may simply be that IS staff believe it is easier(and perhaps better) to hire technically orientedpeople and teach them business ,skills rather thanto follow conventional wisdom and do thereverse.

To summarize, the results from this study are,to some extent, at odds with some of the prevail-ing thinking concerning the evolution of skills forIS personnel and particularly the changing natureof the systems analyst position. The resultsreported in this paper suggest that job profiles,as specified in ads, have not changed dramatical-ly over the past 20 years and that furthermore,the changes that have occurred suggest an in-creased emphasis on the technidal aspects of ISjobs within ads. Clearly a realignment of ourpreviously held impressions of the profession andthe data reported in this study is in order if weare to understand this phenomenon fully. Theresults suggest that we need to rethink botheducational and recruitment strategies to ensurethat they are in alignment. We hope that theseresults, which we anticipate will arouse somecontroversy, will be provocative and will foster ad-ditional debate on the nature of both educationand recruitment strategies.

AcknowledgementsThis work has been supported b~/the ResearchProgram of the School of Business at Queen’sUniversity. We are grateful to Alix Acs, Ann Fitz-gerald, Pamela Hearty, and Carrie Wheeler fortheir assistance with the collection and analysisof the data.


Endnote’ Of course, a less extreme set of positions would be that bothsets of skills are important but that organizations believe itis preferable to have individuals come in with strong technicalskills and provide training to enhance business skills. ISacademics would likely argue for the opposite course.

ReferencesAlbin, M. and Otto, R.W. "The CIS Curriculum:

What Employers Want from CIS and GeneralBusiness Majors," Journal of Computer In-formation Systems (27:1), January 1987, pp.15-19. :

Anderson, J.J. "Developing an In-HouseSystems Training Program," Data Manage-ment (7:6), July 1969, pp. 26-31.

Bell, J.D. and Richter, W.W. "Needed: BetterCommunications for Data Processors," Per-sonnel (63:5), May 1986, pp. 20-26.

Benbasat, I., Dexter, A.S., and Mantha, R.W."Impact of Organizational Maturity on Infor-mation Skill Needs," MIS Quarterly (4:1),March 1980, pp. 21-34.

Butt, J.M. "Third Generation Management,"Data Management (7:9), September 1969,pp. 38-40.

Cheney, P.H. and Lyons, N.R. "InformationSystems Skill Requirements: A Survey," MISQuarterly (4:1), March 1980, pp. 35-43.

Cheney, P.H. and Lipp, A. "Information SystemsPersonnel Skill Requirements: 1978 and1987," University of Georgia Working Paper#28, Athens, GA, 1987.

Cheney, P.H., Hale, D.P., and Kasper, G.M. "In-formation Systems Professionals: Skills forthe 1990s," in Proceedings of the 22nd An-nual Hawaii International Conference onSystems Sciences, J.F. Nunamaker, Jr. andR. H. Sprague, Jr. (eds.), IEEE ComputerSociety Press, Honolulu, HI, January 1989,pp. 331-336.

Davis, D.B. "The Hard Demand for Soft Skills,"Datamation (39:2), January 1993, pp. 28-32.

Dawson, K.M. and Dawson, S.N. Job Search:The Total System, John Wiley, New York,1988.

Data Processing Management Association.DPMA Model Curriculum for UndergraduateComputer Information Systems Education,Park Ridge, IL, 1986.


Exton, W. "The Information System Staff: MajorObstacles to Its Effectiveness, and a Solu-tion," Journal of Systems Management (21:7),July 1970, pp. 32-36.

Green, G.I. "Perceived Importance of SystemsAnalysts’ Job Skills, Roles, and Non-Salary In-centives," MIS Quarterly (13:2), June 1989,pp. 115-133.

Harrison, W. and Springer, D. "A SoftwareSystems Management MBA Option," DataBase (16:2), Spring 1985, pp. 19-23.

Ives, B. and Olson, M.H. "Manager or Tech-nician? The Nature of the Information SystemsManager’s Job," MIS Quarterly (5:4),December 1981, pp. 49-63.

Jenkins,G.H. "Education Requirements for theEntryLevel Business Systems Analyst," Jour-nal of Systems Management (36:3), August1986, pp. 30-33.

Joslin, E.O. and Bassher, R.A. "SystemManagers Speak Out on Performance Direc-tions," Journal of Systems Management(27:2), February 1976, pp. 18-21.

Leitheiser, R.L. "MIS Skills for the 1990s: ASurvey of MIS Managers’ Perceptions," Jour-nal of Management Information Systems (9:1),Summer 1992, pp. 69-91.

’ Nelson, R. "Educational Needs as Perceived byIS and End User Personnel: A Survey ofKnowledge and Skill Requirements," MISQuarterly (15:4), December 1991, pp. 503-525.

Nunamaker, J.F., Couger, J.D., and Davis, G.B."Information Systems Curriculum Recom-mendations for the 80’s: Undergraduate andGraduate Programs--A Report of the ACMCurriculum Committee on InformationSystems," Communications of the ACM(25:11), November 1982, pp. 781-805.

Roark, M. L. "Information Systems Education:What Industry Thinks," Data Management(14:6), June 1976, pp. 24-28.

Seller, R.B. "What Makes a Successful SystemPerson," Journal of Systems Management(27:3), March 1976, pp. 6-11.

Smith, H.A. and McKeen, J.D. "Computerizationand Management: A Study of Conflict andChange," Information & Management (22:4),1992, pp. 53-64.

Strout, E. "The Activities and Education ofSystems Analysts," Journal of SystemsManagement (22:1), January 1971, pp. 37-40.

Walsh, J., Johnson, M., and Sugarman, M. HelpWanted: Case Studies of Classified Ads,

Olympus Publishing Company, Salt Lake City,.UT, 1975.

Watson, H.J., Young, D., Miranda, S.,Robichaux, B., and Seerley, R. "RequisiteSkills for New MIS Hires," Data Base (21:1),Spring 1990, pp. 20-29.

White, T.C. "The 70’s: People," Datamation(16:2), July 15, 1970, pp. 40-46.

About the AuthorsPeter A. Todd is an associate professor ofmanagement information systems and chair ofthe Research Program in the School of Businessat Queen’s University. He received his Ph.D. inmanagement information systems from theUniverity of British Columbia and has previouslyserved on the faculty of the University of Houston.He has published papers in a variety of journalsincluding MIS Quarterly and ISR. His primaryresearch interests include human-computer in-teraction, decision support systems, behavioraldecision making and the adoption and diffusionof information technology.

James D. McKeen is an associate professor atthe School of Business, Queen’s University atKingston, Canada. He received his Ph.D. in MISfrom the University of Minnesota. For the past 24years he has worked in and researched issuesin effective information technology managementincluding: user participation, project selection,strategic information systems, and the businessvalue of information technology. Most recently,he has established Queen’s ManagementForums to help organizations deal with new ap-proaches to work. Jim is currently a feature writerfor Ernst & Young’s Information TechnologyForum and serves as the director of the Queen’sResearch Consortium for IT Management. He isalso the MIS editor for the Canadian Journal ofAdministrative Sciences. He has published ar-ticles on related topics in the MIS Quarterly, In-formation & Management, Communications ofthe ACM, OMEGA, Canadian Journal of Ad-ministrative Sciences, Strategic InformationTechnology Management: Perspectives onOrganizational Growth and Competitive Advan-tage, and Business Process Engineering: AManagerial Perspective. He is currently co-authoring a book on IT management issues.



R. Brent Gallupe is a professor of informationsystems and director of the Queen’s ExecutiveDecision Centre at the School of Business,Queen’s University at Kingston, Canada. His cur-rent research interests are in computer supportfor groups and teams, the evaluation of informa-

tion systems in public sector organizations, andthe history of information systems. His work hasbeen published in such journals as ManagementScience, Academy of Management Journal, Jour-nal of Applied Psychology, and Sloan Manage-ment Review.

AppendixFull Description of Methodology Used

A number of methodological app~oaches have been used to assess IS job skill requirements. The mostcommon approach has been to survey experienced practitioners (usually IS managers) and academics.Questionnaires have been the major data gathering tool (Albin and Otto, 1987; Benbasat, et al., 1980;Cheney and Lyons, 1980; Leitheiser, 1992). Cheney and Lyons also used personal interviews to aug-ment the questionnaires. Jenkins (1986) used the Delphi technique with questionnaires to capture thedata in his study. Finally, Watson, et al. (1990) used group support software to capture the views face-to-face groups of IS managers and professionals to determine current IS skill needs.

All of these approaches rely on asking current practitioners and academics about the importance ofdifferent IS skills. Thus, these approaches are based on collective, subjective perceptions of currentpractice. They may reflect what managers and others believe to be the desirable skill set. On the otherhand, they may reflect what respondents have read in the trade journals or curriculum guidelines ofprofessional associations, or simply what they expect the researchers want to hear. Given this relianceon subjective data, it would seem useful to examine the skills question using descriptive data to deter-mine the degree to which these managerial perceptions reflect the reality of organizational action(Leitheiser, 1992, p.86). This study provides one such data set, using job advertisements, and further-more, does so while examining Changes over time.

This approach has several strengths that complement the methods employed in existing studies. First,published advertisements provide a description of what organizations actually say about their needs.Thisinformation is produced to meet the needs of the organization and not to meet the needs of researchers.Thus, it should reflect what an organization perceives to be important characteristics of job applicants.Our contention is that, if organizations advertise IS positions in major newspapers, they would be carefulto describe the skills they actually require. Further, job descriptions are likely to reflect the ideals ofwhat organizations look for in a candidate, since organizations are motivated to attract (and ultimatelyhire) personnel with the highest possible qualifications and best fit with the organization. Second, becausewe are using secondary data, we are not restricted in terms of sample size, as is often the case withinterview or survey-based studies. This provides a better opportunity to examine a wide range of jobs.In fact, it permits us to employ a sample that is roughly an order of magnitude larger than those usedin prior research. Third, the use of secondary data facilitates an analysis over time. Previous studiesin this area have been largely cross-sectional, and changes over time have been captured by comparisonsof different studies. Fourth, content analysis, little used in IS research but widely employed in otherdisciplines, allows us to capture information about the changing nature of IS jobs.

There are limitations to this research approach. First, the data may not be in a consistent form, andextra effort is required to make the data comparable. Second, decisions must be made by coders asto the meaning of phrases and their categorization. This means that coding reliability must be assessed.Despite this limitation, we believe this approach provides a unique window on the issue of IS knowledgeand skill requirements and should complement the previous work in the area.

IS Job AdvertisementsVery little has been published on the analysis of IS job advertisements or on the use of job advertisementsin general as research data. A seminal study in analyzing job advertisements was published by Walsh,

24 MIS Quarter/y/March 1995


et al., in 1975. They reported the results of an indepth analysis of job advertisements from the two majornewspapers in two large Western U.S. cities. They also included follow-up data from newspapers in10 other U.S. cities. The researchers coded all the information in the job advertisements found in theSunday editions of the two newspapers for the period 1968-1972.

The major findings of the study by Walsh, et al., 1975, are the following:

1. "As a recruitment tool, they [job ads] are of considerable value to a comparatively small but impor-tant group of employers" (p. 19).

2. "As a job seeking tool, job ads are used extensively and with a considerable degree of success" (p. 19).

The researchers found that it was large organizations that used job advertisements the most to recruitemployees. They also found that the greatest success in using job ads for recruiting were for profes-sional, managerial, and technical workers. Overall, they argued that "it is possible to analyze job adssystematically and to assess their content and value to employers and job seekers" (p. 89).

A different view of the importance of job ads is found in Dawson and Dawson (1988). They argue thatonly 10 to 20 percent of jobs are actually found through job ads and search firms. They do, however,state that "our informal surveys indicate that about 80 percent of job seekers concentrate their job searchon job ads" (p. 82).

The data for this study come from IS job advertisements found in newspapers. Therefore, it is criticalthat we understand exactly what these ads represent and how and why they are used. Job ads are usedby organizations to aid them in finding qualified applicants to fill the positions they have available. Theyare a first screening device meant to simultaneously encourage qualified applicants to apply anddiscourage unqualified applicants. As such, they must strike a delicate balance. If job ads are not writ-ten effectively, then many potential applicants will be missed, or potentially unqualified applicants willbe encouraged, thereby wasting organizational resources.

Job ads include a variety of information and can be employed for a variety of purposes (Walsh, et al.,1975). Typically, an ad can be divided into three parts. One part of an ad typically describes the employerand the nature of the business, and conveys some sense of the corporate environment or culture. Thisinformation may be present to sell the potential applicant on the organization and to promote the organiza-tion to the wider community. A second portion of most ads discusses what the job offers the employeein terms of compensation and benefits. It is intended to attract applicants at the right level, in termsof salary expectations, and to further reinforce an applicant’s view of the organization. The final partof an ad describes the nature of the job and the characteristics of the prospective ideal candidate. Itis this portion of the ad that contains the information of interest for this study. In this section, an adwill typically describe the "successful" or the "ideal" applicant. The person eventually hired may notmeet all the requirements as set out in the ad. In this sense, the ads themselves represent an organiza-tion’s ideal of a good programmer, analyst, or IS manager (i.e., what knowledge and skills the organiza-tion truly values, not the knowledge and skills contained in the organization).

Coding SchemeIn order to extract the desired information and not be overwhelmed by incidental information, a specificcoding scheme was developed to categorize information and to exclude information that was not rele-vant to the specific issues under investigation.

We started with the basic premise that all information systems jobs are likely to require some knowledgeof information technology, some knowledge of business, and some set of problem-solving skills. Severalspecific categories of knowledge and skills were drawn from the literature examining systems analystskill needs (Benbasat, et al., 1980; Cheney and Lipp, 1987; Cheney and Lyons, 1980; DPMA, 1986;Nunamaker, et al., 1982; and others) and work by Nelson (1991) and others looking more generally the types of IS knowledge and skills required by employees within an organization. From these wedeveloped a set of seven skill categories: hardware, software, business, management, social, problem



solving, and development methodology. The categories are described in Table 2. They follow the ACMcurriculum model developed by Nunamaker, et al. (1982) with some minor modifications. Specifically,two ACM categories were subdivided: "computers" into "hardware" and "software," and "organiza-tions" into "business" (which focuses on industry and functional area knowledge) and "management"(which includes such skills as organizing, planning, controlling, and the like). Two ACM categories werecombined: "people" and "society" into a single category called "social." This was done because wecould not determine significant differences between the two categories as they pertain to job skills. Inaddition, the ACM category called "systems" was enlarged to include all knowledge pertaining to develop-ment methodologies (as described in Table 2).

In order to simplify presentation and analysis, these seven categories are grouped into three broaderclasses as follows:

¯ Technical Knowledge --* (1) Hardware and (2) Software

¯ Business Knowledge --, (3) Business, (4) Management, and (5)

¯ Systems Knowledge --, (6) Problem Solving and (7) Development Methodology

Prior to collecting a complete s~mple, research assistants analyzed a pilot sample of 200 ads and at-tempted to classify each word or phrase mentioned in the ads into one of the seven categories describedin Table 2. From this, an index was built that included the specific phrases used in the ads and thecoding categories to which they belonged. This index was reviewed by the authors, and corrections weremade. The final index was then used as the coding guide for the primary sample. This provided a referencedictionary for the classification of terms. To check the coding accuracy, the actual words and phraseswere recorded onto a coding sheet under the appropriate category. Counts of the number of words orphrases within each ad (classified by coding category) make up the basis of data reported in this paper.

Phrases were most often comprised of groups of words but in some cases Were only single words. Forexample, the phrase "administrative supervision of project teams" would be classified as a manage-ment skill; the phrase "excellent’communication skills" would be classified as a social skill; the phrase"IMS/DB" would be classified as a software skill; and the phrase "experience in retailing" would beclassified as a required business skill.

A detailed coding of phrases within categories was also conducted. Tables 3,4 and 5 show the detailedresults of this analysis. This codir~g identified the major classes of phrases within a category. For exam-ple, references to software frequently mentioned specifics such as COBOL, or a particular operatingsystem such as VM. These would be recorded as COBOL and operating systems respectively in thetables. We attempted to capture the major themes within each category through this analysis. For somecategories, there are large numbers of "other phrases" that did not seem to fit into the identified categories,or occurred in such numbers as ~o require categorization. For example, under social skills there werecategories for communication, in~lependence/motivation and interpersonal skills. Phrases that were notrelated to the specific subcategories, but were felt to reflect social skills, were placed in the "other"category. Among these were phrases such as perceptive, mature, hard-working, concientious, aggressive,reliable, and bright, among others.

The SampleThe sample was drawn from four newspapers. Two were from the U.S. (The Wall Street Journal andThe New York Times) and two were from Canada (The Globe and Mail and the Toronto Star). In eachcase, these are among the largest circulation newspapers in the country, and they tend to draw careeradvertising that is national in scope. Thus, the jobs are not strictly limited to cities within the geographicalarea in which the newspaper is published.

Samples of IS job ads were drawn from the four newspapers at five-year intervals--1970, 1975, 1980,1985 and 1990. The choice of five-year intervals was made to have a window that was broad enoughto allow for a reasonable chance t~) observe differences and to keep the sample size within each windowrelatively large. Ads were chosen from each month of the year to avoid the possibility of seasonal or



cyclical effects in the data. The search for ads was arbitrarily started on the third Saturday in the month,and any computer-related career ads were extracted. This avoided limiting the selection to specific jobtitles, which may not reflect the true nature of all IS jobs. Although we were interested specifically inprogrammers, analysts, and IS managers, we also collected data on consultants, auditors, systems pro-grammers, database administrators, and data entry staff, among others. Typically, for any year, the sampleof such ads is relatively small and is not included in the analysis reported here.

Our target sample was to collect from 1500-2000 ads. Again, this target was relatively arbitrary, but itwas felt that it would provide sufficient numbers of each type of ad (analyst, programmer, and manager)over each of the five time periods. Given this target, selection from a given newspaper, for a given month,continued until between five and 10 ads were extracted (i.e., 5-10 ads * 12 months * 4 papers * 5 timeperiods = 1200-2400 ads). This range was provided since experience from our pilot indicated that, attimes, 10 or more ads could easily be found in a single Saturday issue, but in some cases, particularlyfor the earlier time periods, several days would need to be searched.

Our target sample was met for each five-year period except for 1970, where only 135 ads, in total, werecollected due to the small number of ads available. For all other time periods we collected between fiveand 10 ads per newspaper per month. In total, 1,634 usable ads were collected that contained sufficientinformation on which to base an analysis. Some ads were rejected because they were non-computer-related jobs (typically engineering analysts and engineering project managers). Still others could notbe used because they contained little or no information about the job, or did not include a job title. Intotal, there were 75 such ads. The final sample included ads for 581 programmers, 348 systems analysts,and 305 IS managers. The remaining jobs included EDP auditors, systems programmers, database ad-ministrators, data entry staff, and consultants, among others. Because the number of ads for these jobtypes was small, no further analysis was performed. Instead, we focused on the 1,234 programmer,analyst, and IS manager jobs. The distribution of these positions across the five time periods is shownin Table 1. In general, the number of ads for later time periods is somewhat greater, reflecting the growthin the field.

Computation and Presentation of ResultsThis section presents a detailed description of how the values appearing in Figures 2 and 3 werecalculated.

The indicator data (Figure 2) shows the proportion of ads that refer to a given skill class (technical,business, or systems) in any given year. For each ad, a value of "1" was assigned if the skill categorywas mentioned one or more times, and a value of "0" was assigned otherwise. The information in theFigure was then computed for each job at each of the five-year intervals as:

Total number of ads referring to the categoryTotal number of ads

The raw data (in Figure 3) presents the average number of phrases per ad that refer to each of the threeskill categories. The average is calculated for each job type and for each year as follows:

Total number of phrases referring to each categoryTotal number of ads

The percentage data (also in Figure 3) is computed by calculating the proportion of each ad that is madeup of phrases in each of the three categories. This is done for each ad by dividing the number of phrasesin a given category by the total number of phrases in the ad. For example, to calculatethe percentageof technical phrases the formula is as follows:

Number of technical phrases in the adTotal number of phrases in the ad

The percentage shown in Figure 3 results from calculating the arithmetic mean of the values computedfor each ad (for each year and each job type).


the evolution of is job skills: a content analysis of is job … · 2016. 1. 22. · cluding...

Documents