Association for Information SystemsAIS Electronic Library (AISeL)

ICIS 1984 Proceedings International Conference on Information Systems(ICIS)

1984

Implementation Failure and System DeveloperValues: Assumptions, 7[tuisms and EmpiricalEvidenceKuldeep KumarUniversity of Waterloo

Richard J. WelkeMeMaster University

Follow this and additional works at: http://aisel.aisnet.org/icis1984

This material is brought to you by the International Conference on Information Systems (ICIS) at AIS Electronic Library (AISeL). It has been acceptedfor inclusion in ICIS 1984 Proceedings by an authorized administrator of AIS Electronic Library (AISeL). For more information, please contactelibrary@aisnet.org.

Recommended CitationKumar, Kuldeep and Welke, Richard J., "Implementation Failure and System Developer Values: Assumptions, 7[tuisms and EmpiricalEvidence" (1984). ICIS 1984 Proceedings. 12.http://aisel.aisnet.org/icis1984/12

Implementation Failure and System Developer Values:Assumptions, 7[tuisms and Empirical Evidence

Kuldeep Kumar, University of Waterlooand

Richard 1 Welke, MeMaster University

ABSTRACTIn the information systems literature the incidence of implementation failures hasincreasingly been attributed to excessive attention to technical and economic issues,and an absence of concern about the social, political, and psychological (individual)aspects of the system being developed. On an intuitive level this has been explainedby assuming a techno-economic value orientation of the system developer This paperpresents empirical evidence in support of the assumption of the dominance oftechnical and economic values in system developers.

The basis of this evidence is a field study of developer values. These values weremeasured by adapting a value measurement methodology developed by England(1967). This methodology determines the behavioral relevance of values by classifyingthem from operative (most likely to govern behavior) to non-relevant (values having noimpact on behavior).

The study results show that technical and economic values are the most operative ofsystem developer values. In the social, political, psychological domain, systemicvalues, and the values relating to the organization and functioning of the develop-ment project were found to be operative. However, the study found that the developersconsidered user job satisfaction related values mostly non-relevant

Introduction implementation failures by suggesting that the design-ers of computer-based information systems subscribe

A survey of information systems literature (Ackoff, to overly rational technical and economic design ideals.1967; Argyris, 1971, 1980; Swanson, 1974; Hedberg At the same time they suggest an absence of attention .

and Mumford, 1975; Lucas, 1975; Kling, 1977; Bostrom paid to the social, political, and psychological issues insystems development For example, Hawgood, Land,and Hienen, 1977a; Hawgood, Land, and Mumford, and Mumford (1978, p. 40) state that"manypastfailures1978; DeMaio and Bartezzagi, 1979; Welke, 1979, of computer-based information systems can be directlyDeMaio, 1980; Bostrom, 1980; and Zmud, 1983) sug- attributed to-a lack of knowledge of human needs andgests that implementation failures are widespread and

serious. The consequences of such failures are reflected motivation on the part of technically oriented systems

in the nonacceptance of the system by users, the jeop- analysts and designers." Bostrom and Hienen (1977a)ardizing of technical and economic investments in the include the sytem developer's limited goal orientation,system, high maintenance costs, and the opportunity optimizing the technical system and limited frame-costs ofunrealizedbenefits (Zmud, 1983). Furthermore, works, nonsystemic view with limited focus on decision

such failures also tend to build upon one another making and data processing as some of the causes ofthrough their effect on the user attitudes towards the

MIS problems and failures. Bmer (1981) suggests that

information systems developers. Lucas (1973, 1975) has major design decisions are usually made by technicalspecialists who tend to be guided by machine efficiencyshown thal in the long run, these attitudes influence the considerations. DeMaio (1980) summarizes this reason-

success or failure of future system development efforts. mg as:The advocates of socio-technical and participative "-the primary cause of problems and failures ofapproaches to system development (Hawgood, Land computer based information systems is theand Mumford, 1978; Bostrom and Hienen, 1977 b; inadequacy of the conceptual frame of referenceDeMaio, 1980;.'Illrner, 1981) explain the incidence of of the information system analysts/designers. In

1

particular such a conceptual frame of reference At the operational/empirical level there are two previ-can be analysed through the-"non-systemic" ous studies with the stated intent of measuring systemapproach to design, as only variables relevant to developer values. Hedberg and Mumford ( 1975) meas-technical- economic subsystem are considered, ured designer values in terms of their Theory X vs.while those relevant to social subsystem are Theory Y (McGregor, 1960) view of the system users.omitted;-the resulting focus (is) on a limited This study, which reported the designer's pre- disposi-objective, namely the optimization of the techni- tion towards a Theory X view of system users, hascal- economic subsystem." implications for developer preferences for technical

efficiency and control oriented information systemOther authors have attributed this orientation of system solutions. Hedberg and Mumford also found that thedevelopers to a technological imperative (Davis, 1971), system designers perceive themselves as having a ratherthe new utopian attitude (Boguslaw, 1965), and anoverly limited role in terms of the system design contributionsmechanistic orientation (Zmud, 1983). Kling (1977) to the organization. "They see their principal workstates that "the prevailing non:ns of computer system activities as increasing efficiency through stream-liningdesign are machine-oriented" procedures and providing better information. They do

not appear to appreciate the potential of computerThe advocates of socio-technical and participative sys- technology for improving the overall quality of workingtemdevelopmentapproachesgoontosuggestthatthese life" (Hedberg and Mumford, 1975, p. 50). Bostrom andapproaches be used to compensate for the limited Hienen (1977a) report similar results from a 1971 U.S.techno-economic perspective of the system developers. study (Thylor 1971).On the other hand, Hedberg (1980) suggests that thevery same limited value orientation of the system Anderson (1978) measured the value orientations ofdevelopers has been a major obstacle in the adoption of computer science students using a variation of Rokeach'sthese approaches in practical systems development (1973) measurement of terminal values (desirable end-

states). These terminal values consist of values such asUnderlying the preceding discussion is the as- family security, a world of beauty, world at peace, salva-sumption that the developers of the information tion, mature love, scientific knowledge, etc.; a list whichsystem primari& subscribe to technical and eco- does not have direct implications for the system designnomic values and consider the social, politicaland processpsychological design ideals nonrelevant to the sys-tem design process. In addition to these studies with the stated aim of

measuring developer values, there have been someUsually this assumption is stated as a truism, though studies measuring developer preferences in terms ofsometimes it is justified in terms of the technical (com- their objectives, criteria, priorities, etc These prefer-puter-oriented) background, the rational training, and/ ences can be interpreted as the concept of the desirable,orthe efficiency oriented reward structure of the system ie., values (Kluckhohn, 1951). Smith (1977) measureddevelopers (Hedberg, 1980). developer rankings of ten system control objectives

such as materiality, timeliness, security, useability,Thepurposeofthispaperis to empirically validate retrievability, etc., and compared them to rankings pro-(or disprove) this assumption. vided by system users. Hallam and Scriven (1976)

surveyed MIS managers for their EDP objectives.Schussel (1974) measured 200 DP and user executives

System Developer Values- on the level of importance they attached to fourteen DPperformance criteria, such as meeting deadlines,

A Review of the Literature accuracy and completeness, quick response to userrequests, budgetperformance, and control Allowayand

The issue of values has surfaced occasionally in the Nolte (1979) surveyed DP executives and systemsliterature of management science, systems theory, and developers in five firms in an attempt to measure theinformation systems. At the theoretical/analyticallevel, importance they attached to sixteen analyst skills. Thethe role of values in systems has been recognized by priority attached to such sldlls as user orientation andChurchman (1961, 19684 1968b), Kling (1977, 1978), behavioral sensitivity could be interpreted as theMattessich (1974, 1978), Klein, et aL, (1981), Klein importance developers attach to these surrogate values.(1981), Sage (1977), Berg, Chen, and Zissis (1976).Though most of these works discuss the role ofvalues in WiththeexceptionofHedbergandMumford(1975),thesystems development, they do not isolate the values in preceding studies were not designed to empirically veri-question explicitly. and, therefore, do not attempt to .fytherelative level of importance attached toa variety ofempirically measure the subscription to these values. technical, economic, and socio-political-psychological

2

system development values (design ideals). Therefore, A literature survey was used to identify various objec-the list of values considered in these studies tend to be tives and concerns which arise in the course of systemsrather sparse, and do not include the numerous value development. These concerns identified the aspects ofdimensions which drive systems development Most the computer-based information system which werestudies tend to be localized in the technical and economic relevant in the context of systems development Theseitems, but their categories are so broad that much of the aspects were associated with attributes, thereby gener-finer discrimination is lost ating a list of end values. The development project

analogs of the object system aspects and their associ-In summary, ourreview ofliterature suggests thatexcept ated attributes were identified to determine a possiblefor some peliminary evidence found by Hedberg and setof means values. The precedinganalysis resulted inaMumford (1975), the assertions of techno-economic large value list which was reviewed for clarity, complete-dominance and the advocates of socio-technical and ness, and orthogonality by a panel of system expertsparticipative systems development approaches, have, at (system analysts, experienced system users, managersbest, a very weak empirical justification for their as- and directors of systems, consultants, and academics).sumptions. The remainder of this paper describes a A final list of 86 value concepts was used in developingstudy conducted to more adequatelytestthis assumption an information systems development personal value

questionnaire (ISD-PVQ) for measuring system de-veloper values.

Research Problems and Method The second methodological problem concerned therelationship of professed values to system developmentbehavion Jick (1981) and England (1967) note that

RESEARCH HYPOTHESIS ANDMETHODOLOGICAL PROBLEMS people do not always act on all the values they verbally

indicate as important Hedberg (1980) reports similar

A field survey was conducted to empirically substantiatefindings from his 1975 study of system developer values

(or disprove) the assumption that system developers (Hedberg and Mumford, 1975). This problem was

primarily subscribe to technical and economic values, addressed in England's (1967) "Theoretical Model of

and find the social, political, and psychological values the Relationship of Values to Behaviour." This modeladdresses the problem by determining the "behavioral

nonrelevant in the context of the development of com- mtentionality" ofvalues. The modelwas operationalizedputer- based information systems. Stated in the null by England as a managerial Personal Value Question-form the research hypothesis was: naire (PVQ). England's model and his PVQ were used

as a basis for developing our instrument for measuringHO: There is no difference between the relative the value profiles of system developer,The modelis brief-levels of importance attached to technical, eco- ly described below.nomic, and socio-political-psychological valuesby the developers of computer-based informa-tion systems.

Testing this hypothesis posed several methodological ENGLAND'S THEORETICAL MODEL OF THEproblems. First was the problem of enumerating a value RELATIONSHIP OF VALUES TO BEHAVIORlist which is both relevant to the system developmentprocess, and complete and comprehensive from the three England's model recognizes four behavioral categoriesperspectives of technical economic, and socio-political of values. The total value space consists of all potentialpsychological values. An analysis of the definition of value values. The potential values for a specific group orprovided a likely starting point Kluckhohn (1951) individual are made up of two classes of values; (1) Non-defines values as: relevant values (values having little or no impact on

behavior), and (2) conceived values (values which may" . . .a conception, explicit or implicit, distinctive influence behavior). Conceived values are further parti-of an individual or characteristic ofa group, ofthe tioned into (i) operative values (those that have a verydesirable which inBuences the selection from high probability of translation from intended to actualavailable means and ends of action" behavior), (ii) adopted values (those values which are less

a part of the personality structure of the individual, butThis definition was used to develop a framework for may affect behaviorthrough situational factors), and (iii)system developmentvalues whichclassified them either intended values (values which the individual states asas end values (values that influence the selection of the being important to him, but have only a moderatemeans or the "development approach" employed probability of being translated into behavior because of(Kumar, 1984). situational reasons). The modelis presented in Figure 1.

3

BehaviorPotential Values Channeling

Nonrelevant ConceivedValues Values Alternative Generationfor a - Testing, Decision Mak- EnvironmentalSpecific Operative \ ing, Problem Solving Influences &Individual Constraints

or a ValuesGroup      

Adopted LimitedRange of

Values Behaviorm

Intended   Selecting, Filtering &- Interpreting Incoming

Values Sensory Data

Perceptual Screening

Figure 1

Theoretical Model of the Relationship of Values to Behavior. (Adaptedfrom England, Agarwal, and 'Iterise, 1971).

ENGLAND'S PERSONAL VALUE 2. The Reason Mode-As the focus of the PVQ isQUESTIONNAIRE to make operational the behavioral effectof values, it

is necessary to make operational the theoretical dis-The above model was used in devising a personal value tinction between the intentionality ofvalues and theirquestionnaire (PVQ) which has been used by England translation into behavior (operative from among con-and several others to measure the value systems of ceived values). 'Ib the extent that it is possible tovarious managerial groups (managers-England, 1967; determine a consistent rationale as to why an indi-England, Agarwal et aL, 1970; England, Agarwal, and vidual thinks certain concepts are important orDhingra, 1974; union leaders-England, Agarwal, and unimportant one has a reasonable basis for distin-Derise, 1971; small business owners-Lindecamp, guishing operative from conceived values. The1981; educational administrators-Sjogren, England, secondary, or reason mode measures the primaryand Meltzer, 1969; and US. naval officers-England, rationale (success, right, and pleasant) attached toAgarwal, et aL, 1970). the value concept

The development of the PVQ is based on the rationale The overall responses to all the value concepts are usedthat the meaning an individual attaches to a carefully to determine the dominant reason mode (primary valuespecified set ofconcepts provides a useful description of orientation in England' terminology) for the individualtheir personal value system which, in turn, is related to respondent The dominant reason mode is the reasontheir behavior in systematic ways (Osgood et aL, 1957). (success, right, and pleasant) most frequently attached

to important value concepts.England's PVQ determines the behavioral strength ofeach value concept by measuring it along two dimen- A combination of the importance and reason modes issions (modes) of meaning: thought to be a better behavioral predictor than the

importance mode alone. For example, if a manager's1. The Importance Mode-Since the general dominant reason mode is success (Le. when he says

value of an object of an idea is thought to be largely a something is important, he is more frequently apt to seefunction of its degree of importance, the primary it as successful as opposed to right or pleasant), hismode of valuation utilized is an importance scale. behavior would be predicted best by viewing itas a joint

4

function of those concepts he though as important and England. In addition to the standard PVQ questions, the

successful (operative values). On the other hand, those questionnaire also included items relating to the demo-value concepts which are neither important nor fit the graphic attributes of the respondents. In order to main-

respondent's dominant reason mode will be his non- tain respondent interest in the face of a rather lengthyrelevant values (ie. values which are not expected to questionnaire, principles of design from marketinginfluence his behavior). In between are the behaviorally research (Dillman, 1978) were used to format theless relevant intended values (concepts which are re- questions. Some sample questions are presented ingarded as important, but do not fit the person's domi- Appendix A.nant reason mode) and the situationally inducedadopted values (values which fit the dominant reason The questionnaire was pre-tested on a representativemode of the individual, butare regarded tobe of average group. In addition, a test-retest was peiformed with aor low importance). sample of thirteen accounting and business students.

The test-retest reliability co-efficients for the primatymode and the secondary mode were 0.89 and 0.84

INDIVIDUAL AND GROUP VALUE PROFILES respectively, which are comparable to the results re-ported by England, Olsen, and Agarwal (1971) for the

An individual's value profile can be constructed by PVQs given to educational administrators and navalanalyzing the value questions in the PVQ according to officersthe above procedure in order to determine the indi-vidual's operative, adopted, intended, and nonrelevantvalues. Methodology Application andSuchaprofilecanbeconstructedforeachrespondent.It Resultsis also possible to derive an overall profile for therespondent group by aggregating the individual value CONDUCT OF THE SURVEYprofiles. This aggregationmaybeused fortabulatingtheproportion of the group for whom a particular value The field survey was conducted in thirteen Canadian

concept is operative, adopted, intended, and nonrele- business and government organizations. The organi-vant, and for determining the modall value category for zations sampled included federal, provincial, and city

the group. The overall value profile will then show a government departments, electric and nuclear power

similar categorization for each of the concepts in the utilities, manufacturing, retail, insurance, and universi-

PVQ. ties. The organizational levels of respondents includedvice-president of systems directors of MIS, MIS man-

While this data is complete, it is somewhat difficult to agers through to programmer-analysts.interpret because of its voluminous nature (number ofvalue concepts x 4 value categories). England, Dhingra, Tb obtain the sample we contacted the highest rankingand Agarwal (1974) have developed a summary index to information systems executives in various organizations.

portray value patterns and their behavioral relevence. Of approximately twenty organizations which initially

This index is called the Behavioral Relevance Score of agreed to participate, seven dropped out after thethe value concept, and is the percentage of the total second contactgroup for whom the concept is an operative value. Thisscore can vary between 0 and 100 for any given concept In the remaining thirteen organizations, the contactA high score for a concept indicates a high behavioral information systems executive was requested to ran-

relevance of the concept for the respondent group and, domly select information system developers to partici-as such, indicates a group value which is very likely to pate in the survey. The information systems executivegovern behavior. was also requested to aIrange a one hour meeting for

the researcher with the selected respondents

THE INFORMATION SYSTEMS The ISD-PVQ was administered to the respondentDEVELOPMENT-PERSONAL VALUE group at this meeting. The meeting was opened with anQUESTIONNAIRE explanation of the purpose of the study (surveying

system developer values), followed by instructions onThe Information Systems Development-Personal completing the questionnaire. The respondents wereValue Questionnaire (ISD-PVQ) was developed by sub- encouraged to respond according to their personalstituting the information systems development listof 86 preferences by suggesting that there were no right orvalue concepts into the PVQ framework developed by wrong answers and that the individual responses were

confidentiaL The researcher stayed in the room with the1 Where '·mode" has the usual meaningof the class orthe category with the larg-

est frequency of occurrence. respondents to answer any clalifying questions.

5

RESULTS communication structures, etc.). At the nonrelevant endwere most of the job satisfaction related values such as

The final sample contained 132 system developers from the status of the user's job in the organization, thethirteen business and government organizations. 0f the alignment of user salaries to his/herjob description, job132 respondents, 34 were found to have a mixed domi- security for the user, and the variety of tasks in the usernant reason mode (i.e. when they said that a value job. Also at the nonrelevant end were values related toconceptwasimportant, theywereequallylikelytoattach the analyst's job satisfaction. However, the concept ofthe rationale of success, or right, or pleasanttoit). These user job design and the resulting job satisfaction wererespondents were excluded from further analysis be- found to be marginally operative (with 36 percent of thecause the intersection data required to classify value respondents classifying it as operative). In addition, theconcepts can not be reliably calculated for persons alignment of the mode of information display to thehaving a mixed dominant reason mode. cognitive/decision style of the users was found to be non-

relevantFor the remaining respondents, each person's valueprofile was determined by classifying each of the valueconcepts for that person into one of the operative, Behavioral Relevance Scoreadopted, intended, and nonrelevant categories. The Value Profile.individual value profiles were then aggregated intogroup value profiles. Two methods of aggregation, were The behavioral relevance score for each of the valueused to produce the group value profiles; the modal concepts was determined by calculating the percentagevalue category method, and the behavioral relevance of the group for whom the value concept is operative.score. Figure 3 presentsthebehavioralrelevance scores forthe

technical, economic and the socio-political-psychologi-cal values for the sample. The figure has been subdi-vided by horizontallines to indicate the quartile rankingModal Category Value Profile of the value concepts

Figure 2 presents the modal category value profile for With respect to both technical and economic values,the technical, economic and socio-political psycho- most of the value concepts were found to be clustered inlogical values of the system developers inthe sample.'Ib the high (top two quartiles) behavioral relevance scoregenerate this group profile, the individual profiles weretabulated to determine the modal category (Le. the

range. In both these classes of values, only a few value

category with the highest frequency of occurance) for concepts (such as values relating to computer hardwareand system software in the technical value class, and

each of the 86 value concepts in the ISD-PVQ. Each of monitoring and control of clerical and operating work inthe values was classified over all four behavioral cate- the economic value class) were found to have low behav-gories of values (Le., operative, adopted, intended, and ioral relevance scores. This indicates that most of thenonrelevant). However, the modal categories in all technical and economic values have a high likelihood ofcases, were polarized into the two extreme categories ofoperative and nonrelevant None of the value concepts influencing the system design decisions.

were predominantly adopted or intended At the high end of the socio-political-psychologicalvalues are those related to the conduct of the system

Most of the technical values were found to be operative. development project (such as user participation andHowever, values relating to computer hardware and formal assignment of responsibility) and systemicsoftware, the use of latest technology, and latest system values (such as the status of the user's job within thedevelopment methodologies were found to be non- organization, job security for the user, variety of tasksrelevant. and learning, and growth in user jobs) are clustered in

the bottom quartile. The values related to the analyst'sIn the domain of economic values all value concepts,except for the monitoring and control of clerical and own job satisfaction are also rated as having low behav-

ioral relevance. These scores imply that the job satisfac-operating work were found to be operative. tion related values have a low likelihood of influencingsystem development behavior.In the domain ofsocio-political-psychological values two

different value polarities emerged. At the operative endwere values related to the conduct of the project (suchas user participation in system design, formal assign- Conclusions and Implicationsment of responsibility forthe project frequency of userreviews, etc.), and systemic values (such as the primary The objective of this paper was to substantiate orclientof the system, organizational goals and objectives, disprove the assumption that system developers pri-

6

Technical Values Economic Values Socio-Political-Psychological Values

Operative System Reliability Effect of System on Primary Client.f OrganizationUser Understanding of Overall System Design

Values Input/Output ProcessingTimeliness of Information Planning and Control of Development Project Uaer Manager Participation in Design DecisionsAccuracy & Consistency of Data: Analyst Errors in Design

Documentation Useability by Project learn and by Users System Responsiveness to PeopleSystem FlexibilitySystem Controls User Undent of Dev. Plan  Formal Resp. Assgmnt on Project

Deve|opment Project on ScheduleData Flows (Input Forms, Reports Messages, Etc) Organization's Goals and ObjectivesMaintainability of the System procedures Development Project on BudgetCompatability with Interfacing Systems

Frequency of User ReviewsSystem Throughput Currency or Recency of information Elapsed Time for the Development Project Support for Organization's Goals & ObjectivesRelevance of Information to Decision or Task Level of Analysis Expertise/Skills RequiredConsistency of Work between different AnalystsAdequacy and Comp. of Data: System Documentation Operating Costs of the System

Asaignment & Formalization of User Resp. in SystemData Stores (Files and Databases) System Development CostsPrompt Response to DeveL Requests: Program ModifiabilityFlexible Devel Standards: Computer Support for Decisions User Clerks Participation in the SystemEase of Documentation Prep. and Maint: User Decisions User Manpower Required for DeveL Project Communication between Organizational UnitsComputer ProgramsManual ProceduresHuman Errors in Data Processing Communciation StructureSystem Response Time Flow and Consumption of Organization's Resources

Systems Dept Manpower Required for Dev. ProjecLDate Privacy; User Job De8ign/ Satisfaction User's sense

User Manpower Required for Operating the System of Contribution.Security of Update/Retrievat Computer Support for DeveL Control of Organization'B Resources Organizational Structure

Nonrelevant Hardware/Software Latest Developmental Methodologies Routine & Repetitive User Tasks

Values Alignment of Information Display to Cognitive StyleLearning New Skills during System Development

User Job's Health/Safety; Analyst's Autonomy on Project

Job [nduced Mental Stress on UserLear:u'Growth in User Jobs; Analyst Social-Contact on Job

Sophisiticated of Hardware/Software User-Social Conflict on Job; Variety of Analysis'IksktiMonitoring and Control of Clerical Work

Usermisk Variety, User Autonomy; Analyst Job RoutinessProportion of Challenging vs. Simple .lobs-Analysts

Proportion of Challenge vs. Simple Tasks- UsersUser Job Security: Others in Society affected by System

Centralization of Authority & Decision MakingCentralization of Hardware/Software User Salary-Job Alignment, Statu s of User Job

User Understanding of Technical SystemFigure 2

Modal Category Value Profile

Technical Values Economic Values Socio-Political-Psychological Values

74. System Reliahility 74. 74.

73. 73. 73.72. 72. 72.71. 71. 71.70. inpuUOutput Procesang 70. 70.69. Timeliness of Information 69. Planning and Control of Development Project 69. 'Ibp68. Accuracy & Consis. of Data; Analyst Enon in Design 68. 68. Effect of System on Primary Client of Organization67. 67. 67. User Understanding of Overall System Design Quartile66. 66. 66.65. Documentation Useability-by Proj. 7Aam;-by Users 65. 65. User Manager Participation in Design Decisions64. System Flexibility 64, 64.6,3. SyRtem Controls 63. 63.62. 62. Development Project on Schedule 62.61. Data Flows (Input forms, Reports, messages etc) 61. 61.60. Maintainability of the System Procedures 60. Development Project on Budget 60.

59. Compatability with Intelfacing Systems 59. 59. System Responsiveness to People58. System Throughput; Cum)ncy or Recency of Inf(1 58. Elapsed Time for the Development Project 58.

57. Relevance of Information to Decision or Task 57. Level of Analysis Expertise/Skills Required 57. User Underst of Dev. Pla  Formal Resp Assgn on Proj.56. Consistency of Work between Diff. Analysts 56. 56. 2nd55. Adequacy and Comp. of Data System Documentation 55. Operating Costs of the System 55. Organization's Goals and Objectives54. 54. 54. Quartile53. Data Stores (Files and Databases) 53 53.52. Prompt Response to DeveL RequestK Program Mod. 52. 62. Frequency of User Reviews51. Flexible DeveL Std&; Computer Support for Decisions 51. 51. Support for Organization's Goals & ObjectiveR50. Ease of Documentation preF and Main; User Decia 50. 50.49. Computer ProgramB 49. 49.

Behavioral 48. Manual Procedures 48. System Development Costs 48.

00 Relevance47. Human Errors in Data Processing 47. 47.

Score 46.46. System Response Time 46. Assignment & Formalization of User Resp. in System45. 45. 45.44. 44. User Manpower Required for DeveL Project 44.43. 43. 43.42. 42. 42. User Clerks Participation in the System41. 41. 41. Communication between Organizational Units 3rd40. Security of Update/Retrieval; Comp. Supf for DeveL 40. 40. Quartile39. 39. Flow and Consumption of Organization's Resources39.38. 38. Systems Dept Manpower Required for Dev. Project38, Communication Structure37. 37. User Manpower Required for Operating the System37.36. 36. Control of Organization's Resources 36. Data Privacy; User,lob Design/Satisfaction: User's35. 35. 35. -- Senseof Contribution34. 34. 34. Organizational Structure33. 33. 33.32. 32. 32. Routine & Repetitive User Tasks31. 31. 31.30. Hdwe./Softwe.: Latest Develop. Methodologies 30. 30. Alignment of Info. Displays to Cognitive Slyle

29. 29. 29.28. 28. 28. Learning New Ski!]a during System Development27. 27. 27. Bottom26. 26. 26. User Job's Health/Safety; Analyst'R Autonomy on Pro. Quartile25. 25. 25.24. 24. 24.23. 23. 23..Job Induced Mental Stress on User22. 22. 22.21. Sophistication of Hardware/Software. 21. 2].

20. 20. Monitoring and Control of Clerical Work 20.

19. 19 19. [zarn/Gmwth in User Johz Anal. Social-Cont. 0,1 .IIi)18. 18. 18. User-Social Contact on J,h; Variety of Anaysis TaHks

Use

rTas

kV

.Ils

erAu

tono

my:

Anal

yst·

1ohRo

ut.

engin

;mp

le.fohs

naly»t

.i

3.C

entra

lizat

ion

ofHar

dware/S

oftware

.en

gevi

Sipi

e

Fig

ure

3

Beha

vior

alR

elev

ance

Scor

eVa

lue

P

, marily subscribe to technical and economic values, andfind the socio-political-psychological values nonrele-vant in the context of information systems development.

: The results from the field study give strong support tothe assumption of the dominance of technical and eco-nomic values. Most technical and economic values are

14{ 1 exception of values related to computer technology,

clustered in the high behavioral score ranges. With the

* development methodologies, and the monitoring ande  3 1 controlofclericaltasks,allthesevaluesareintheopera-tive behavioral categorx This implies that the technical

'8 and economic value concerns are highly likely to influ-3 - 0 2 N ence and govern system design choices and behavior.

*2 59£ 5 1  0.2 0 0 : E The nonrelevant values in the technical and economicc = c CO 274 3 9€ 4 1 value classes suggest two interesting implications. First

& 2 12 1that system developers seem to have outgrown their

2.2. . earlier fascination with the latest and most sophisti-/Koiddiggw#*Wivvit cated technology (both hardware, software, and metho-

dologies), and second, that there seems to be a move4 away from the monitoling and contml of people-oriented

Theory X view of systems.

In the socio-political-psychological domain, values arefound relevant as long as they contribute directly to thesystem development project (such as frequency of userreviews participation by users, formal and definiteassignment of project responsibilities), or are related tosystemic goals and concerns (such as the primary clientof the system, organization's goals and objectives, com-munication structure).

On the other hand, job satisfaction and quality of work-ElfidLE<*t#*diWWdE ing life values are considered non relevant to the design

of information systems. These values are clustered inthe lowest behavioral score ranges. This implies thatthey are likely to have very little or no impact on thedesign decisions made by the system developers.

Information systems development is a complex processwith numerous competing objectives. Usually the re-sources (bothtime and money) available forthe systemsanalysis phase are limited. This means that only the highpriority objectives or value concerns will be accommo-dated in the design of the new system. Job satisfactionand quality of working life value concerns, with lowbehavioral relevance (in the bottom quartile), are there-fore likely to be ignored by the system developersconcerned with bringing the system development pro-ject on schedule (behavioral relevance score- 62), andwithin budget (behavioral relevance score-60).

Efid-atigist@gggLE

However, the advocates of socio-technical and partici-pative development (Hawgood, Land, and Mumford,1978; DeMaio, 1980; Bostrom and Hienen, 1977) havesuggested that the lack of attention to the job satisfac-

9

tion concerns is one of the primary causes of the imple- REFERENCESmentation failures of computer-based information sys-tems. Sackman (1983) states that "computer profes- Ackoff, L. "Management Misinformation Systems,"sionalsoftensubvertthedevelopmentprocesswiththeir Management Science, Volume 14, Number 4, 1967,

narrow sectarion concerns-." Therefore measures are pp. B147-B156.needed to compensate for this lack of attention. Alloway, R.M. and Nolte, J.T. "What They Don't Know

Can Hurt Alarming Results from a Survey of SkillsLand, Mumford, and Hawgood (1980), DeMaio (1980), Priorities and Promotion Criteria for Systemsand Welke (1979) suggest that this lack of attention to Analysts," CenterforInfonnation Systems Research,job satisfaction issues can be compensated by the intro- Working Paper, Sloan School of Management,duction and adoption of systems development metho- Massachusetts Institue pf 'Ibchnology, Boston,dologies with strong socio-technical and participative Massachusetts, 1979.components. However, Hedberg (1980) suggests that Anderson, R.E. "Value Orientations of Computerthis very lack of attention is a major obstacle in the Science Students," Communications of ACM,adoption of such methodologies. Sackman (1983), and Volume 21, Number 3, March 1978, pp. 219-225.Hoyer (1980), have indicated that the system analyst's Argyris, C. "Management Information Systems, Thelimited value concerns are major problems in adoption Challenge to Rationality and Emotionality,"of participative development approaches. Management Science Volume 17, Number 6,

February 1971, pp. B-275.This means that we may need to go to the root cause of Argyris, C. "Some Inner Contradictions inManagementthe problem and attempt to modify the underlying value Information Systems," in Information Systemsstructures of the system developers. Courbon and Enuironmen< H.C. Lucas, F.F. Land, T.J. Lincoln,Bourgeois (1980) suggest that we need "a new breed of and K Supper, (eds.) North-Holland, Publishingdesigners, who will be nurturing agents" for the socio- Company, Amsterdam, The Netherlands, 1980.technical process. We suggest two options for develop- Berg, M., Chen, K, and Zissis, G., "A Value-Orienteding this new breed. First, we should train the aspiring Policy Generation Methodology for Technologysytems developers not only in the technical and eco- Assessment," 7tchnological Forecasting and Socialnomic aspects, but also in the human and job design Change Volume 8, Number 4, 1976, pp. 401-420.aspects. Nygaard (1983), through the suggestion for Boguslaw, R. The New Utopions, Prentice-Hallinformatics as an academic discipline, and Land, Englewood Cliffs, New Jersey, 1965.Mumford, and Hawgood (1980) through their proposal Bostrom, R.R and Hienen, J. "MIS Problems andin "hining the Systems Analysts of the 19804" support Failures: A Socio-Technical Perspective; PART I:this approach The Causes," MLS Quarteriz Volume 2, Number 3,

September 1977(a), pp. 17-32.Second, we can change the reward structure for the Bostrom, R and Hienen, J. "MIS ProblemsandFailures:system developers in such a manner that the conscious A Socio-'Idchnical Perspective; PART II: Theconsideration and maximization of job satisfaction and Application of Socio-Technical Theory," MIShuman concerns is rewarded Hedberg (1980) states: Quarterly, Volume 2, Number 4, 1977(b), pp. 11-28.

Bostrom, FLP."A Socio-Technical Perspective on MIS"Designers design to please those who control Implementation, a paper presented at ORSA/

TIMS National Conference, Colorado Springs,the rewards. They live in a world where technical Colorado, November, 1980.constraints, costbudgets, anddeadlinesarereal Churchman, C.W. Prediction and Optimal Decisionand demanding, but wherehumanneeds, demo-cratic organization structures, and user partici- Philosophical Issues ofa Science of Values, Prentice-

pation are little but window dressing. Should we Hall, Englewood Cliffs, New Jersey, 1961.Churchman, C.W. Challenge to Reason, McGmw-Hillmoralize overthat they first attend to problems

of costs, timing, and tehcnical functioning-. No,New York, New York 1968(a).

Churchman, C.W. Systems Approach, Dell Publishingthey act according to the rules of the game they Co., New York„ New York, 1968(b).are in. It is these rules that must be changed Courbon, J.C. and Bourgeois, M. "The InformationSystem Designer as a Nuturing Agent of the Socio-

It is our belief that a suitable combination of the above Technical Process," in The Infonnation Systemsthree strategies (Le. socio-technical and participative Environmen4 H C Lucas, F.F. Land, T.J. Lincolnmethodologies, system analyst training, and a modified and K Supper (eds.), North- Holland Publishingreward structure) will be instrumental in modifying the Company, Amsterdam, The Netherlands, 1980.system developers' value structure, such that adequate Davis, LE. "Job Satisfaction Research: The Postattention is given to the human and job satisfaction Industrial View," Industrial Relations, May 1971,issues. pp. 176-193.

10

DeMaio, A., Bartezzaghi, E. and Zanarini, G. "A New Enuironments, H.C. Lucas, F.F. Land, T.J.System Analysis Method Based on the STS (Socio- Lincoln, and K Supper, (eds.), North-HollandTechnical Systems) Approach" in Formal Models Publishing Company, Amsterdam, The Nether-and P>·actical 7bols for Infonnation System Design, lands, 1980.H.J. Schneider, (ed,) North-Holland Publishing Jick, T.D. "Managers and Quality of Working Life: ACompany, Amsterdam, The Netherland, 1979. Clash of Values?" in Management Under Di#erent

DeMaio, A. "Socio-Technical Methods for Infonnation Value Systems, G. Dlugos, and K Weiermeir, (eds.)Systems Design," in The Information Systems Walter De Gruyter, Hawthorne, California, 1981.Enuironmeni H.C. Lucas, RE Land, T.J. Lincoln Klein, H.K, Meadows, LS.G. and Welke, RJ. "Improvingand K Supper, (eds.) North-Holland Publishing the Quality of Working Life through Better DesignCompany, Amsterdam, The Netherlands, 1980. of Information Systems," in Proceedings of the

Dillman, D. Mail and 7klephone Surveys-The 7btal International Conference on Applied System ResearchDesignMethod John-Wiley& Sons, NewYork,New & Cybematics, PergamonPress Inc., Elmsford, NewYork,1978. York,1980.

England, G.W. "Personal Value Systems of American Klein, H.K Design Ideals and their CriticalManagers," 7'he Academy of Management JournaL Reconstruction, ISRAM DP-8105-2.1, McMasterVolume 10, Number 1, 1967, pp. 53-68. University, Hamilton, Ontario, Canada, 1981.

England, G.W., Agarwal, N., et aL, Personal Values and Kling, R. "The Organizational Context of User-Militaiy AdministratioG Technical Report No. Centered Software Designs," MIS Quarterly,AD711358, University of Minnesota, Minneapolis, December 1977, Volume 1, Number 4, pp. 41 -52.Minnesota„ August 1970. Kling, R. "Value Conflicts and Social Choice in

England, G.W, Agarwal, N. and Dhingra O.R The Electronic Funds'Dansfer System Developments,"Manager and Man, A Cross-Cultural Study of Communications oftheACM August 1978, VolumePersonal Values, Comparative Administration 21, Number 8, pp. 642-657.Research Institue, Kent State University Press: Kluckhohn, C. "Vdue and Value Orientations in theKent, Ohio, 1974. Theory of Action: AnExploration in Definition and

England, G.W., Agarwal, N. and Ferise, R.E. "Union Classification," in 7bward a General Theory ofLeaders and Managers: A Comparison of Value Action, T. Parsons, andE.A Shils, (eds.) HarperandSystems," Industrial Relations Volume 10, Row, New York, New Yorl£ 1951.Number 2, May 1971, pp. 211-226. Kumar, K Participant Valws in Systems Developmen4

England, G.W., Olsen, K and Agarwal, N. A Manual of Unpublished Doctoral Dissertation McMasterDevelopment and Research for the Personal Value University, Hamilton, Ontario, Canada, 1984.Questionnaire, The CenterforComparative Studies Land, F.F., Mumford, E. and Hawgood, J. "Faining thein Technological Development and Social Change, Systems Analyst of the 1980's; Four AnalyticalUniversity of Minnesota Minneapolis, Minnesota Procedures to Assist the Design Process," in The

Hallam, E and Scriven, D.T "EDP Objectives and Information Systems Enuironmen4 H.C. Lucas, F.F.Evaluation Process," Data Managemen4 Volume Land, T.J. Lincoln and K Supper, (eds.), North-14, Number 5, May 1967, pp. 40. Holland Publishing Company, Amsterdam, The

Hawgood,L.,FrankE andMumford, E."AParticipative Netherlands, 1980.Approach to Forward Planning and Systems Lindecamp, D.R Personal Values andBusiness SuccessChange," in G Bracchi, and P.C. Lockermann, (ed£), of SmaU Independent Retailers, UnpublishedInfonnation Systems Methodology, Proceedings of Doctoral Dissertation, Texas A&M University,the 2nd ConferenceoftheEuropeanCooperationin College Station, Texas, 1981.Informatics, Springer-Verlag, Venice, Italy, 1978. Lucas, H.C. "User Reactions and the Management of

Hedberg, B. andMumford,E."TheDesignofComputer Information Services," Management Informatics,Systems: Man's Vision of Man as an Integral Partof Volume 2, Number 4, 1973, pp. 165-172.the System Design Process," in Human Choice and Lucas, FIC. Why Infomzation Systems FaiL ColumbiaComputers, E. Mumford, and H. Sackman, (eds.), University Press: New Yorl£ New York, 1975,North-Holland Publishing Company, Amsterdam Mattessich, R. "The Incorporation and Reduction ofThe Netherlands, 1975. Value Judgements in Systems," Management

Hedberg, B. "Using Computerized Information Systems Science, Volume 21, September 1974 (a), pp. 1-9.to Design Better Organizations and Jobs," in 77ze Mattessich, R. Instrumental Reasoning and SystemsHuman Side of Information Processing, Bjorn- Methodology, D. Riedel Publishing Co.: Dordrech.Anderson, Niels, (ed.) North-Holland Publishing Holland, (1978).Company, Amsterdam, The Netherlands, 1980. McGregor, D. 77:e Human Side ofEnterprise, McGraw-

Hoyer, R. "User Participation-Why is Development so Hill, New York, New York, 1960.Slow? The Dynamics of the Development of End Nygaard, K "Participation in Systems Development-User Control," in The Information Systems The Tasks Ahead," in Systems Design, FoL With,

11

and By the Users, U. Briefs, C. Ciborra, and L. Journal of System Management, Volume 28,Schnieder, (eds.) North Holland Publishing Number 7, July 1977, pp. 16-22.Company, Amsterdam, The Netherlands, 1983. Swanson, B. "Management Information Systems:

Osgood, E.C., Suci, G.C. and Tannenbaum, P.C. 77:e Appreciation and Involvement" ManagementMeasurement of Meaning, University of Illinois Science, Volume 21, Number 2, October 1974, pp.Press, Urbana Illinois, 1957. 178-188.

Sackman, H. "Problems and Promise of Participative Taylor, J.C. 7kchnology and Planned OrganizationalInformation System Design," in Systems Desigrd Change, CRUSK: Institute for Social ResearchFo# With, and By the Users, U. Briefs, C. Ciborra, The University of Michigan, Ann Arbor, Michigan,and L. Schneider, (eds.) North Holland Publishing 1971.Comany, Amsterdam, The Netherlands 1983. Tirner, J.A. "Computer and Clerical Jobs: The Missed

Sage, R MethodologyforLarge ScateSystems, McGraw- Opportunity for Work Redesign," Working PaperHill„ New York New York, 1977. CRIS #24, GBA #81-32(CR), CenterforResearch

Schussel, G. "Scoring DP Performance," Infosystems, on Information Systems, New York University, NewVolume 21, Number 9, September 1974, pp. 59-62. York, New York 1981.

Sjogren, D., England, G.W. and Meltzer, R Devebpment Welke, R.1 "User Oriented Approaches to MIS," CAof an Instntment for Assessing the PersonalValuesof Magazine, Volume 112, Number8, August 1979, pp.Educational Administrotors, Final Report Project 62-68.No. 8-H-016, Colorado State University, Fort Zmud, R.W. "CBIS Failure and Success," in InformationCollins, Colorado, 1969. Systems in Organization, Scott Foresman, and

Smith, C.P. "Resolving User/System Differences," Company, Glenview, Illinois, 1983.

12

Appendix A

Information Systems Development-Personal ValueQuestionnaire

Sample Questionnaire Items

Section II: Having decided which aspects of the system to examine and/or develop, the nextquestion deals with the direction in which development should take place (the "norms" ofdevelopment). This section measures the level of importance you attach to each of thesenorms of criteria. For each of the items listed below, please indicate:

1. The level of importance you attach to it, and2. The primary meaning (Le. success, right or pleasant) you attach to the item.

IMPORTANCE RATING MEANINGV. LO MED HI V. SUCC RGHT PLSNLO HI123451 2 3

1. Timeliness of informationsupplied by the system

4. Security of update/retreival I ] [ ] [ ] [ ] [ ] [ ] I ] 1 ]access to information

16. Operating costs of the Il Il Il Il Il Il Il Ilsystem

40. Job security for the users

45. Development project on Il fl Il Il Il Il Il Ilschedule

13