User Involvement

The Effect of UserInvolvement onSystem Success: AContingencyApproach

By: Peter TaitTouche Ross Pty. Ltd.Queen StreetBrisbane, 4000 Australia

Iris VesseyGraduate School of BusinessUniversity of PittsburghPittsburgh, Pennsylvania 15260

AbstractDespite the fact that commercial computer sys-tems have been in existence for almost threedecades, many systems in the process of beingimplemented may be classed as failures. One ofthe factors frequently cited as important to suc-cessful system development is involving usersin the design and implementation process. Thispaper reports the results of a field study, con-ducted on data from forty-two systems, that in-vestigates the role of user involvement and fac-tors affecting the employment of user involve-ment on the success of system development.Path analysis was used to investigate both thedirect effects of the contingent variables on sys-tem success and the effect of user involvementas a mediating variable between the contingentvariables and system success. The results showthat high system complexity and constraints onthe resources available for system developmentare associated with less successful systems.

Keywords: User involvement, system success,user information satisfaction, userattitudes, system impact, systemcomplexity, resource complaints

ACM Categories: D2.9, H1.0, K6.1, K6.3

IntroductionDespite the fact that the commercial use of com-puters is now three decades old, many of thesystems currently being implemented may beclassed as failures. Certain systems imple-mentations, even in the 1980s, run well overbudget, some are discontinued, while others per-form at levels far below those expected (McFar-lan, 1981 ). Other systems require major, expen-sive modifications after implementation beforethey are acceptable to users. The reasons for thefailure of computer-based information systems(CBIS) are not well understood. Research needed to understand why so many systems fail,and to identify and~or develop tools and tech-niques to aid the successful development andimplementation of CBIS.

Many factors are believed to affect the successof the development and implementation of CBIS.Zmud (1981), for example, suggests that factorssuch as the organization, the environment, thetask, personal and interpersonal characteristics,as well as MIS staff characteristics and policiescan influence the success of system imple-mentation. More recently, however, failures havebeen attributed to individual and organizationalreaction to CBIS implementation (Maish, 1979).

User involvement in CBIS design and imple-mentation is frequently cited as one possiblemethod of overcoming implementation failures.This research addresses the role of user involve-ment and factors affecting the employment ofuser involvement on the success of systems im-plementation. It draws on the literature in organi-zational change theory, participative decisionmaking, and information systems to develop atheoretical basis for the study. Further, it usescasual modeling to develop a path model of thearea under investigation.

Prior Research on theEffect of User Involvementon System SuccessAll research into user involvement has been pos-ited in the belief that user involvement has apositive influence on the successful introductionof a CBIS into an organization. Although somestudies have found positive relationships be-tween the effectiveness of user involvement andsystem success, the results of many studieshave been inconclusive. Several studies have

MIS Quarterly/March 1988 91

User Involvement

assessed the impact of a number of variables,one of them user involvement, on the success ofthe system (Alter, 1978; Alter and Ginzberg,1978; Gallagher, 1974; Guthrie, 1974; Lucas,1975, 1976; Maish, 1979; Olson and Ives, 1981;Powers and Dickson, 1973; Schewe, 1976;Swanson, 1974; and VanLommel and DeBra-bander, 1975). The Alter, Gallagher, Guthrie,and Swanson studies found positive rela-tionships between user involvement and,systemsuccess; the other studies produced inconclu-sive results. Ginzberg (1979) states that "there little conclusive evidence of the value of partici-pative design." In their comprehensive review of22 studies on the influence of user involvementon system success, Ives and O~son (1984) reportthat eight studies found a positive relationshipbetween user involvement and system success,seven produced mixed results,’ while the re-maining seven studies produced negative ornonsignificant results. They concluded that:

1. Research on user involvement is rarely basedon strong theory.

2. Empirical research has not convincingly dem-onstrated the benefits of user involvement.

3. The majority of studies on user involvementhave been methodologically flawed to the ex-tent that few conclusions can be made aboutuser involvement’s relationship to systemsuccess (p. 587).

The lack of empirical evidence in favor of theeffect of user involvement on system successmay be the result of either point 1 or point 3.Further studies with strong theoretical founda-tions that build on the knowledge from prior re-search are essential to understanding the roleuser involvement plays in the successful imple-mentation of CBIS.

Theory and HypothesesThe results of previous studies on user involve-ment have been so mixed as to render appropri-ate a study of selected factors perceived to medi-ate the influence of user involvement on systemsuccess. This study provides a framework fororganizing knowledge about those factors.

Where there was more than one measure of the vari-able, the results were in opposite ~/irections.

Rather than attempting to investigate all factorsaffecting user ir~volvement and its impact on sys-tem success, the model provides a structurewithin which to examine constructs central to theinfluence of user involvement on system suc-cess. Once the validity of this structure is estab-lished by using global constructs, those con-structs can be’ further disaggregated to deter-mine the factor~ that are driving the effects, untilwe reach a level of understanding rather thanprediction (Dub’in, 1978).

This section I~resents the conceptual modeltested in this re, search. It justifies the inclusion ofglobal constructs in the model and presents thetheoretical grounds for their inclusion. Next it ar-ticulates the theoretical basis for the inclusion ofspecific factors into the model and the rela-tionships of those factors to user involvementand system success.

Conceptuai model investigatedFigure 1 presents the.model tested in the study.It describes two sets of relationships: the con-tingencies affecting the extent of user involve-ment in system design; and the effect of thosecontingencies .and of user involvement on thesuccess of the ’.system. The contingencies identi-fied for further investigation in this study are:user system variables, technical system vari-ables, and development process variables.

The variables influencing user involvement andsystem success investigated in the model are asfollows:

1. User system variables:(a) User attitudes(b) Impact of the system on the organization

2. Technical system variable:System complexity

3. Development process variable:Resource constraints

The model in Figure 1, illustrates the propositionstested in this research:

Proposition 1 : The extent of user involvement in~successful system design and.implementation is affected by’user system, technical system,and development process vari-ables.

92 MIS Quarterly~March 1988

User Involvement

USER SYSTEM

[ System Impact I[ User Attitudes ]

TECHNICAL SYSTEM

System Complexity

DEVELOPMENT PROCESSResource Constraints

USERINVOLVEMENT SYSTEM SUCCESS

Figure 1. Contingency Model of User Involvement and SuccessfulSystem Design and Implementation

Proposition 2: The success of a system isaffected by the extent of user in-volvement in its design and im-plementation, user system, tech-nical system, and developmentprocess variables.

Justification for the use of acontingency model

Several researchers have identified the need totake a contingent approach to user involvementin the design and implementation of CBIS. Forexample, Swanson (1974) concluded that therewas not a simple, direct relationship betweenuser involvement and system success. He sug-gested the need for a revised model of user in-volvement and system success that included en-vironmental factors. DeBrabander and Edstrom(1977) state that in certain cases user involve-ment may not result in successful implementa-tion; they advocate also taking into consideration

the context in which the system is developed topredict the impact of user involvement.2 Edstrom(1977), Lucas (1973), Mann and Watson (1984),Schonberger (1980), Swanson (1974), (1981), and Ives and Olson (1984) all presentmodels of factors affecting the contribution ofuser involvement to system success.

Contingency theory acknowledges that certainvariables may affect the outcome of a particularprocess. Contingency theory itself has no con-tent; it is merely a framework for organizingknowledge in a given area. Therefore, if we are

One of the few studies to take a contingent approachto the effect of user involvement on system successthat did not produce positive results. Edstrom ex-amined the type of system being developed, andhypothesized that user involvement was more effec-tive in structured systems. The hypothesis tested wascontrary to generally held beliefs that user .involve-ment will have a positive impact on system successwhen the system is unstructured. His hypothesis wasnot supported by empirical results.

MIS Quarterly/March 1988 93

User Involvement

to draw on contingency theory as a basis fordeveloping theories regarding the influence ofuser involvement on system success, we mustdraw on well-established contingency theoriesfrom other disciplinary areas as we~l as from priorresearch in information systems.

Theoretical basis for the modelFigure 2 presents the theoretical links betweenthe types of contingent variables and the extentof user involvement and successful system de-velopment and implementation. The followingsections outline the roles played by organization-al change theory, participative decision making,and the information systems literature in the de-velopment of the conceptual model of the influ-ence of user involvement on systems success.

the organization (Lewin, 1951; Schein, 1964).Generally, three broad stages are identified: (a)unfreezing, creating a climate for change; (b)changing; (c) refreezing, institutionalizing thechanges. First Kolb and Frohman (1970) usedthe Lewin-Schein model to represent the con-sulting process. ’Then Ginzberg (1981 b) used theKolb-Frohman model to develop a descriptivemodel for information system implementation: anappropriate climate for change is a contingencyfor user involvement in CBIS implementation.Hence, organizational change theory suggeststhat user system variables, such as user atti-tudes and the impact of the system on the orga-nization, be taken into account when im-plementing a CBIS; i.e., organizational changetheory provides a theoretical basis for the rela-tionship between the user system variables andboth user involvement and system success.

Contribution of Organizational ChangeTheory

The implementation of a new information systemcan be considered an organizational change(Carroll, 1982; Ginzberg, 1978, 1979, 1981a;Hopelain, 1982; Zmud and Cox, 1979). Plannedorganizational change theory recognizes the ex-istence of resistance to change in an organiza-tion and attempts to overcome this resistance byestablishing a suitable ’climate’ for change within

Contribution of Participative DecisionMaking ResearchUser involvement in system design and imple-mentation can be viewed as a special case ofparticipative decision making (PDM) with usersand system designers as the participants ratherthan employees and superiors. PDM researchsuggests that the effectiveness of a contingentapproach to decision making depends on. a num-ber of contextual factors, level of knowledge of

Independent DependentVariables Variables

Organizational Change Theory

User system

Participative Decision Making

User systemTechnical systemDevelopment process

Information Systems

Technical systemDevelopment process

System SuccessUser involvement

User involvementUser involvementUser involvement

System successSystem success

Figure 2. Theoretical Links Between Types of Contingent Variablesand User Involvement and System Success

94 MIS Quarterly~March 1988

User Involvement

Table 1. Hypotheses Tested in This Research

Relationship Hypothesis Direction

User involvementSystem success

User attitudesSystem successUser involvementUser involvement & system success

System impactSystem successUser involvementUser involvement & system success

System complexitySystem successUser involvementUser involvement & system success

Resource constraintsSystem successUser involvementUser involvement & system success

H1 +

H2S +H21 -H2C -

H3S -H31 +H3C +

H4S -H41 +H4C +

H5SH51H5C

employees, motivation of employees, organiza-tional factors, task attributes, group characteris-tics, and leader attributes. 3 Therefore, PDM isrelevant to all the relationships in the model thatconcern the adoption of user involvement inCBIS development (see Figure 2). PDM alsocontributes to our understanding of the effect ofuser involvement on system success both direct-ly and indirectly, via organizational changetheory. One of the major mechanisms for achiev-ing the climate for change (unfreezing) is to en-courage employees to participate in the change(Middlemist and Hitt, 1981).

Contribution of Information SystemsResearch

Since neither organizational change theory norparticipative decision making addresses comput-er-related issues, neither has any relevance tothe relationships between the technical systemand the development process variables, andsystem success. For theory in this area we mustrely on the results of prior studies in informationsystems.

For a comprehensive review of PDM research, seeLocke and Schweiger (1979).

Model variables: definition andrelationshipsThe following subsections define the variablesincluded in the model -- the influence of userinvolvement on system success, and the rela-tionships among them. Hypotheses are pre-sented for the relationships in the model (seeTable 1 ).’

Successful System Development andImplementation

There are many ways of viewing the success ofsystem development. Typically a system is per-ceived as successful when system usage in-creases, when perceptions of system quality aremore favorable, or when users’ satisfaction withthe information they receive increases. Such

4Hypotheses are identified by three characters:i. the letter, H;ii. a digit denoting the user and technical system

variables and the development process variable;and

iii. an S, I, or C depending on whether the hypothesisassesses the effect on system success (S), userinvolvement (I), or the effect on system successvia user involvement (C).

MIS Quarterly/March 1988 95

User Involvement

measures are surrogates for improved decision-making performance, a factor difficult to assessdirectly.

Greatest attention has been directed toward sys-tem use (Baroudi, Olson and Ives, 1986; Lucas,1975; Robey, 1979; Schewe, 1976) and user in-formation satisfaction (Bailey and Pearson,1983; Edstrom, 1977; Ives, Olson and Baroudi,1983; Pearson, 1977; Treacy, 1985)2 as mea-sures of system success. System use is a be-havior, while user information satisfaction is anattitude toward a new system. From a practition-er’s viewpoint, Cerullo (1980) reports that MISprofessionals consider user attitudes as the sin-gle most important success factor. From a theo-retical perspective, AIIport (1935) suggests thatan attitude is a state of readiness that exertsinfluence over one’s actions. Support is found forthis hypothesis in tests of Fishbein and Ajzen’sTheory of Reasoned Action (Ajzen and Fishbein,1980, 1972). Baroudi, Olson, and Ives (1986)also found support for the influence of user in-formation satisfaction on system use as opposedto the influence of system use on user informa-tion satisfaction. Further, assessing the satisfac-tion of the user with the information provided bythe system is probably the most commonapproach to measuring system success. It isalso the measure that has received most of therecent attention in MIS literature. User informa-tion satisfaction is the preferred indicator of sys-tem success and is the measure used in thisstudy to define successful system developrfientand implementation.

User Involvement

User involvement is defined as the "participationin the development by a member or members ofthe target user group" (Olson and Ives, 1981).Some researchers distinguish between the typeand extent of user involvement. The type of par-ticipation may be consultative, representative, orconsensus where the extent of participation in-creases from consultative to consensus (Mum-ford, 1979). According to Ives and Olson (1984),extent of user involvement can be categorized asfollows:

1. No involvement. Users are unwilling or notinvited to participate.

Note that users’ satisfaction with an information sys-tem is different from their satisfaction with their jobs.

2. Symbolic involvement. User input is re-quested but :ignored.

3. Involvement by advice. Advice is solicitedthrough interviews or questionnaires.

4. Involvement by weak control. Users have"sign-off" responsibility at each stage of thesystem development process.

5. Involvement by doing. A user is a designteam member, or is the official "liaison" withthe information systems development group.

6. Involvement .by strong control. Users may paydirectly for new development out of their ownbudgets, or the user’s overall organizationalperformance evaluation depends on the out-come of the development effort. (See alsoSchonberger, 1980!)

Extent of user involvement is a more generalconcept than type of participation and is used inthis research as a measure of user participationin the CBIS development and implementationprocess.

This research examines the often-cited positiverelationship between user involvement in thesystem development process and system suc-cess (Powers and Dickson, 1973; Guthrie, 1974;Carroll, 1982): Research into organizationalchange theory (,Elizur and Guttman, 1976), PDMresearch (Locke and Schweiger, 1979), and re-search in information systems (DeBrabanderand Edstrom, 1977; Ires and Olson, 1984) allsuggest that contextual factors determine the in-fluence of user involvement on system success.It is expected, theref6re, that the effect of theextent of user involvement on system successwill be mediated by the factors affecting user in-volvement. User system, technical system, anddevelopment pi’ocess variables are also takeninto account in assessing the effect of user in-volvement on system success (see Figure 1).These contingent relationships will be examinedin the sections relating to the contingentvariables. ~

H1 : As the extent of user involvement in systemdevelopment increases, the likelihood ofsystem su.ccess increases.

User System Variables

This section defines user attitudes and systemimpact and their relationships to user involve-ment and system success.

96 MIS Quarterly~March 1988

User Involvement

User Attitudes. The psychology literatur.e con-tains many definitions of attitudes. This studyused Rokeach’s (1968) definition of an attitudeas "an organization of interrelated beliefs arounda common focus;" the "common focus" of theattitudes studied in this research is the imple-mentation of a new computer system. Rokeachfurther describes attitudes as having three com-ponents: cognitive, affective, and behavioral orinstrumental. Elizur and Guttman (1976) usedthese three components in their study of thestructure of attitudes toward technologicalchange. They investigated the responses oforganization members to the introduction of anew computer system as follows:

There is usually a greater or lesser feeling ofbeing linked to, satisfied with, or anxiousabout a new device such as a computer.These are affecfive responses. One hasopinions about its advantages or disadvan-tages, usefulness, and necessity, and aboutthe knowledge and information required tooperate it. This is a cognitive response andone has taken or may in the future take ac-tion for or against the object, thus adoptingan instrumental response (emphasis added)(p. 611).

This study used Rokeach’s definition and classi-fication of attitudes, as elaborated by Elizur andGuttman, to define the attitude of users towardsthe implementation of a new CBIS.

Organizational change theory suggests that be-fore a change (in this case the introduction of CBIS) can be successfully implemented, theappropriate "climate" must exist. If the attitudesof users towards a new CBIS are unfavorable,then it’is likely that they will not accept the newsystem, leading to an increased risk of systemfailure. Users may even take action to sabotageor delay the design and implementation of thesystem (Keen, 1981; Markus, 1983). Guthrie(1974), in research based on organizationalchange theory, found empirical evidence that us-ers must "perceive a felt need" for the new sys-tem before it can be successfully implemented.This study hypothesized a positive relationshipbetween user attitudes and system success,based on this research.

H2S: As the favorableness of user attitudestowards the new system increases, thelikelihood of system success increases.

Organizational change theory suggests that un-freezing the organizational climate is essentialbefore a change (such as the implementation ofa new CBIS) can be implemented. Unfreezing isthe breaking of ties with the old system and mov-ing to a new system. User participation in systemdesign and implementation will aid in unfreezingand is, according to Ginzberg (1978, 1979,1981a), a method for achieving the appropriateclimate for the change. In an organization whereusers have few fears about a new system, theirattitudes will be more favorable toward the newsystem and the amount of unfreezing to achievethe appropriate climate for the change will beless (Middlemist and Hitt, 1981). The degree user involvement required to achieve theappropriate climate for the implementation of aCBIS will be contingent upon user attitudes to-ward the introduction of a new system. Thisstudy then hypothesized an inverse relationshipbetween the favorableness of user attitudes andthe degree of user involvement required in thenew CBIS development.

H21:As the favorableness of user attitudes to-ward the new system decreases, the extentof user involvement increases.

There is an indirect relationship between userattitudes on system success and user involve-ment. User involvement exerts a positive influ-ence on the effect that user attitudes have onsystem success.

H2C: As the favorableness of user attitudes to-ward the new system decreases, the ex-tent of user involvement in successful sys-tem implementation increases.

Impact of the System on the Organization.The impact of the system on the organizationrefers to the operational/internal impact ratherthan the strategic/external impact on the orga-nization. Technology is not socially neutral; itusually introduces some changes to the normsthat exist in an organization (Wolek, 1975). Forexample, it may lead to extensive changes in theorganization structure (Guthrie, 1974); it may al-ter departmental procedures and affect jobs andpeople (Carroll, 1982). Different systems willhave differing levels of impact on the organiza-tion within which they are implemented. Wolek(1975) proposed the following criteria for deter-mining whether a new system will change theorganization norms:

1. Changes in the personnel with whom oneinteracts.

MIS Quarterly/March 1988 97

User Involvement

2. Changes in the criteria by which managersevaluate their and/or others’ skills andperformance.

3. Changes in the criteria used to determine therelevant status of different persons in thesystem.

4. Changes in the importance of different inputs(thus the power of related persons) to thesolution of problems.

McFarlan (1981) similarly assesses the impact the system on the organization in terms of therisk of system implementation, which he definesas the extent of organizational structure and pro-cedure changes brought about by the newsystem.

This study used Wolek’s and McFarlan’s defini-tions of system impact to assess the effects ofthe implementation of a new system on the inter-nal user organization.

Organizational change theory suggests that sys-tem success is affected by the changes causedby a system. Attempts to introduce change intoan organization in equilibrium may create forcesthat oppose the change (Hopelain, 1982; Lewin,1951). Keen (1981) and Markus (1983) r~portthat users may actively resist the implementationof a new system. If a proposed CBIS causes alarge impact on the organization, the risk of re-sistance to, and failure of, the CBIS will increase.Information systems literature also suggests thatthe probability of successful system implementa-tion decreases as the impact of the system in-creases (see Carroll, 1982; Ginzberg, 1981a;Powers and Dickson, 1973; Zand and Sorensen,1975). Accordingly, this study hypothesized aninverse relationship between the impact of a sys-tem on the organization and the success of thesystem.

H3S: As the impact of the system increases, thelikelihood of system success decreases.

Organizational change theory suggests thatorganizations can prepare for a change, such asthe implementation of a CBIS, through the ap-plication of PDM, or, in this case, user involve-ment. User involvement in system developmentis especially important if the change is significant(Middlemist and Hitt, 1981). Hence, this studyhypothesized a positive relationship between theimpact of a system on the organization and thedegree of user involvement.

H31: As the impact of the system increases, theextent of user involvement increases.

There is an indirect relationship between systemimpact on system success and user involve-ment. User involvement has a positive influenceon the effect that system impact has on systemsuccess.

H3C: As the impact of the system increases, theextent of user involvement in successfulsystem implementation increases.

Technical System Variable: SystemComplexity

Wolek (1975) describes complexity as "a lack structure for thinking about a problem." Simon(1981) defines a complex system as "one madeup of a large number of parts that interact in anonsimple way" (p. 195). He suggests handlingcomplexity by decomposing the system hierar-chically into subsystems that are functionallycohesive and minimally connected. Mann andWatson (1984) discuss the similar concept task interdependence as a contingency for userinvolvement in system development~ ~n practicalterms, McFarlan (1981) defines as a complexCBIS, one in which the desired input, proces-sing, and output requirements among the interre-lated parts are not easily defined. We can con-sider a complex CBIS as being difficult to de-velop because of the large number of interactingparts (i.e., the number and nature of the interac-tions among subsystems that comprise the sys-tem) and lack of a structure or model to repre-sent them. System complexity is defined in thisresearch as the perceived complexity associatedwith the analysis and design of a system.

In complex systems, problems may arise in theanalysis and specification of the desired system,increasing the risk of unsuccessful developmentand implementation. Because of this, wehypothesize an inverse relationship betweensystem complexity and system success.

H4S: As the complexity of the system increases,the likelihood of system successdecreases.

Participative decision making suggests thatmore user participation is required as the taskbecomes more complex. Shaw and Blum (1966),Morse and Lorsch (1970), and Vroom and Yetton(1973) all assert that highly complex, unstruc-tured tasks require participative decision making

98 MIS Quarterly/March 1988

User Involvement

because of the increased knowledge and flexibil-ity required in decision making. This studyhypothesized a direct relationship between thedegree of user involvement needed to success-fully implement a system and system complexity,following participative decision-making theory.

H41: As the complexity of the system increases,the extent of user involvement increases.

There is an indirect relationship between systemcomplexity on system success and user involve-ment. User involvement positively influences theeffect that system complexity has on systemsuccess.

H4C: As the complexity of the system in-creases, the extent of user involvement insuccessful system implementationincreases.

Development Process Variable: ResourceConstraints

Ein-Dor and Segev (1978) identify two sources resource constraints on CBIS development andimplementation: constraints internal to the orga-nization such as time and funds available tocomplete the project, and those external to theorganization such as the availability of trainedmanpower, hardware, and software. Since theexternal resource constraints globally impact alldevelopment projects, only those internal re-source constraints that impact an organization’sprojects differentially are considered here. Re-source constraints are defined in this study asthe internal, short-run restriction of time and fi-nances available for CBIS development.

Ein-Dor and Segev (1978) propose that resourceavailability problems frequently contribute to thefailure of systems. If resources are insufficient,system designers may not adequately follow nor-mal development procedures, thus increasingthe risk of system failure. This study, therefore,hypothesized an inverse relationship betweenresource constraints and system success.

H5S: As the resource constraints on the de-velopment of the system increase, thelikelihood of system success decreases.

PDM research suggests that user involvement isa development procedure that increases theconsumption of both time and financial re-sources (Locke and Schweiger, 1979). This find-ing is substantiated by Boland (1978) in user in-

volvement literature. If there are internal re-source constraints, then the extent of user in-volvement for the system development processmay be decreased to meet the budget. Hence,this study hypothesized an inverse relationshipbetween resource constraints and the degree ofuser involvement in CBIS development andimplementation.

H51: As the resource constraints on system de-velopment increase, the extent of user in-volvement decreases.

There is an indirect relationship between re-source constraints on system success and userinvolvement. User involvement exacerbates thenegative effect that resource constraints have onsystem success.

H5C: As the resource constraints on system de-velopment increase, the extent of user in-volvement in successful system imple-mentation decreases.

Research MethodologyThis study used a survey-based field study ofmultiple organizations and their recently im-plemented systems to investigate the hypoth-eses just presented. This research strategy fol-lows that used in the majority of previous studies(Alter, 1978; Edstrom, 1977; Maish, 1979; Olsonand Ives, 1981).

The sample organizationsThe data for the study was collected from thirtyAustralian firms with recently implemented cus-tom-built information systems. Twenty of thefirms were initially contacted via their hardwarevendors. The other ten firms were contacted viathe software houses that developed their sys-tems. All organizations had their IS departmentsin the southeast Queensland region. The orga-nizations represented diverse sections of thebusiness community. They included brewers,building societies, builders, credit unions, meatprocessors, millers, retailers, and wholesalers.They ranged widely in size. All firms were fromthe private sector.

Users of the systems were surveyed to measurethe success of the system, the extent of userinvolvement in the system design, the impact ofthe system on the organization, and the attitudes "of the users. Designers of the systems were sur-

MIS Quarterly/March 1988 99

User Involvement

veyed to obtain data on the technical complexityof the system and the resource constraints onthe development of the system. One user of thesystem output and one of the system designersresponded to the survey for each of the systems.

Measurement of the modelvariablesThe following subsections outline the source ofdata used in the design of the questionnaire.

System Success

The instrument used to measure user satisfac-tion, the surrogate for system success, wasPearson’s questionnaire (Bailey and Pearson,1983; Pearson, 1977) modified and shortened asrecommended by Ives, Olson, and Baroudi(1983).

Extent of User Involvement

The instrument used to assess the extent of userinvolvement in CBIS implementation requiredthe users to rate the nature of their involvementas outlined in the section on user involvement(Keen, 1981)2

User Attitudes

The instrument developed by Elizur and Gutt-man (1976) was used in this study to measurethe attitudes of users to the system beingimplemented.

Impact of the System on the Organization

Users assessed system impact by rating the ex-tent and the risk of the design and implementa-tion of the system (McFarlan, 1981; Wolek,1975). Users also made an overall assessmentof the impact of the system on the organization.

User involvement, as measured in this study, is acategorical variable. Implicit in its use is the assump-tion that the differences among categories are insome sense equal, i.e., we assumed measurementon an interval scale. The reader should note thatthere are problems associated with the commonly-used Liked scales. These problems include scaleunits and origins (anchoring), derivation of a score adding heterogeneous item scores, and a problemsimilar to that with the t3ser involvement measure,and the use of parametric statistics (Galletta andLederer, 1987).

System Complexity’

Designers assessed system complexity by ratingthe difficulty of! determining the information re-quirements of the system, the complexity of theprocessing, and the overall complexity of thesystem design (McFarlan, 1981).

Resource Constraints

Designers were requested to rate the extent towhich the development of a system was con-strained by boih time and financial restrictions(Ein-Dor and Segev, 1978).

Administration o’f the field studyFirst, a pilot study was conducted on four sys-tems, with four system users and one systemdesigner completing the survey instruments. Us-ers had difficulty differentiating between two ofthe questions, one concerning accuracy and theother precision, that measured user informationsatisfaction, the surrogate for system success.As a result, the precision question was removedfrom the questi,onnaire.

The primary study surveyed 59 systems. Foreach participating firm, pairs of questionnaires(one for the user and one for the designer) and cover letter were sent to a manager in the dataprocessing department. Follow-up phone calls tofirms with incomplete responses were made af-ter two weeks. The participants were requestedto ensure responses within an additional week.

Data AnalysisComplete responses were obtained for 42 sys-tems, a response rate of 71 percent.

Scale refiability checkThe reliability of the responses to all instrumentswas assessed primarily by means of the Cron-bach alpha reliability coefficient (Cronbach,1951). The reliabilities of certain scales werefurther assessed by analyzing the correlationcoefficients among instrument items. Table 2(a)presents a summary of the reliability results foreach of the instruments used, while Table 2(b)presents the interitem reliabilities for the user in-formation satisfaction instrument. The reliabilityof the overall instruments ranged from .70 to .97;these figures are comparable to those reported

100 MIS Quarterly/March 1988

User Involvement

by Ives and Olson (1984) for similar instruments.The reliability of the figures for the two-item userinformation satisfaction instrument ranged from.78 to .98. The attitude of the EDP staff (.78) andthe time required for system development (.82)were considerably lower than the other alphascores in Table 2b. The reliability figures aremuch more variable than those reported by Ives,Olson, and Baroudi (1983). All instruments,however, meet the level of .70 suggested byNunnally (1978) as satisfactory for exploratoryresearch.

Table 2(a). Reliability Coefficients for theInstruments Used in This Research

Variable Alpha

System success .972

User attitudes .847

System impact .792

System complexity .704

Resource constraints .773

Table 2(b). Interitem Reliability Coefficientsfor the User InformationSatisfaction Instrument

Scales Alpha

Relationship with EDP staff .94Training provided users .89Users’ understanding of systems .92Users’ participation .97Attitude of EDP staff .78Reliability of output .96Relevance of output .95Accuracy of output .97Communication with EDP staff .94Time required for

system development .82Completeness of output .98

Several of the scales were investigated further.The instrument used to measure system impactcontained four basic questions and one overallitem. The reliability reported in Table 2(a) wasderived from the first four items. The Pearsonproduct moment correlation indicates that the re-sponse to the basic questions and the overallitem are closely related (r = .657, p = .001) andprovides some measure of construct validity.

In addition to the calculation of Cronbach’salpha, Elizur and Guttman (1976) suggest thatthe reliability of the user attitude scale is accept-able when the three attitude scales have positiveor zero correlations. Table 3 presents the matrixof intercorrelations for the three dimensions. Allare positive and satisfy this reliability criterion.

Table 3. Matrix of IntercorrelationsAmong the Three Dimensions of the

User Attitudes Instrument

Items

1 2 3

1 1.0000

2 0.3371 1.0000

3 0.6105 0.5742 1.000

Results of analysis of the surveydataTable 4 presents the means, standard devia-tions, and scale directions for the variables in-vestigated in this research. The means of thequestionnaire items were input as variables inthe model. The data was analyzed and tested inthis research (Figure 1).

The method used to test the conceptual modelwas path analysis. It is a multiple regressiontechnique particularly suited to investigating se-quential models such as that proposed in thisstudy (see Asher, 1976; Duncan, 1975; Pedha-zur, 1982; Nie, et al., 1975). One of the majorstrengths of path analysis is its ability to distin-guish among the different effects of one variableon another. For our purposes, path analysis per-mits the researcher to determine (1) the directeffects of one variable on another and (2) theindirect effects of the first variable on the second,through one or more intervening variables.

Path analysis of the model involved conductingtwo multiple regression analyses that tested theeffects of model variables on the internal vari-ables, system success and user involvement.Figure 3 presents the model and the path coeffi-cients obtained for the relationships hypothe-sized. Table 5 summarizes the results of hypoth-esis testing.

MIS Quarterly~March 1988 101

User Involvement

Table 4. Descriptive Statistics of the Variables Studied

Standard ScaleVariable Mean Deviation Direction*

System success 5.72

User involvement 3.31

User attitudes 6.52

System impact 3.51

System complexity 4.22

Resource constraints 5.10

1.16 1-7

1.67 7-1

0.64 1-7

1.46 1-7

1.45 7-1

1.34 7-1

*The scales are recorded from low to high values.

Table 5. Results of Testing the Study Hypotheses

Relationship Hypothesis Direction Beta*

User involvementSystem success

User attitudesSystem successUser involvementUser involvement & system success

System impactSystem successUser involvementUser involvement & system success

System complexitySystem successUser involvementUser involvement & system success

Resource constraintsSystem successUser involvementUser involvement & system success.

H1 + .195

H2S + .026H21 - .014H2C - .003

H3SH31H3C

H4SH41H4C

- - .049+ .027+ .005

- - .094+ .402**+ .078

H5S - -.494"*H51 - -.204H5C - -.040

*The signs of the coefficients reflect the hypotheses tested in this study rather than the scale directions.**Significant at the 0.05 level.

The first regression was conducted with systemsuccess as the internal variable, and user in-volvement, user attitudes, system impact, sys-tem complexity, and resource constraints as theexternal variables. This analysis permitted test-ing the success (S) hypotheses. The second re-gression used user involvement as the internalvariable, and user attitudes, system impact, sys-tem complexity, and resource constraints as theexternal variables. This regression analysis per-mitted testing the involvement (I) hypotheses.Table 6 presents the direct and indirect effects of

the model variables on system success.7 Thecontingency (C) hypotheses relate to the indirecteffects, through user involvement, of the contin-gent variables .on system success.

The model explained 25.03 percent of theobserved variation in system success and 5.80percent of that in user involvement. As expected,

7 The indirect effects due to the model variables areobtained by multiply!ng the contributing pathcoefficients.

102 MIS Quarterly/March 1988

User Involvement

USER SYSTEM

[ Systemlmpact 1-’-

[ User Attitudes ] .__

TECHNICAL SYSTEM

System Complexity

DEVELOPMENT PROCESSResource Constraints

.014,027

.4O2

- .204

- .049

.026

USER INVOLVEMENT.195 SYSTEM SUCCESS

Figure 3. Path Coefficients for the Contingency Model of User Involvement andSuccessful System Design and Implementation

the results for the ef:fects of the contingent vari-ables on system success via user involvementfollow the same pattern as for user involvementalone.

DiscussionWhat are the implications of the results for man-agers who must assign users to system develop-ment teams? The following sections draw someconclusions based on our findings of the effects

of the contingent variables on system success,and on the effect of user involvement as amediating variable between the contingent vari-ables and system success.

Effect of user involvement onsystem developmentUser involvement has a positive effect on systemsuccess, though the effect is not significant (beta= .195); as a consequence, hypothesis H1 is notsupported.

Table 6. Direct and Indirect Effects of Variables on System Success

Variable Model effects

Direct Indirect Total

User involvement

User attitudes ’

System impact

System complexity

Resource constraints

.195 -- .195

.026 .003 .029

-.049 .005 -.044

-.094 .078 -.016

-.494 -.040 -.534

MIS Quarterly~March 1988 103

User Involvement

Effects of user system variables onsystem developmentThe user system variables studied in this re-search are user attitudes and impact of the CBtSon the user organization. The effects of eachvariable will be considered in turn.

Effects of User Attitudes on SystemDevelopment

The association between user attitudes and sys-tem success is small and positive (beta = .026)so hypothesis H2S is not supported. It should benoted here that user attitudes in this study werevery high (mean -- 6.52 on a 7-point scale, sd .674) suggesting a ceiling effect. Similarly, theassociation between user attitudes and user in-volvement was small, positive, and not signifi-cant (beta = .014). Hence, hypothesis H21 is notsupported. As would be expected from the aboveresults, the indirect effect of user attitudes onsystem success through user involvement ispositive, though almost negligible (beta ff .003).The total effect of user attitudes on system suc-cess is .029.

Effects of System Impact on SystemDevelopment

There is a small negative association of systemimpact on system success (beta = -.049). Theresult is in the direction hypothesized, but is notsignificant. As expected, user involvement has apositive effect on system impact. Again however,the effect is small (beta = .027) and hypothesisH31 is not supported. Since the association be-tween system impact and user involvement issmall, so also is the indirect effect of user in-volvement as the mediating variable betweensystem impact and system success (beta .005). The total effect of system impact on thesuccess of the system is still negative (beta -.044).

Effects of the technical systemvariable on system developmentSystem complexity has a negative effect on sys-tem success, as predicted (beta = -.094). Thisresult is in the direction hypothesized, but is notsignificant. Hence, hypothesis H4S is not sup-ported. The association of system complexitywith user involvement is positive and significant

(beta = .402) and hypothesis H41 is supported.This result suggests that managers react to sys-tems that are iechnically complex by involvingusers in their development. The beta coefficientfor the indirecti effect of system complexity onsuccess via user involvement is .078. Again theresult is not s~gn~ficant. The resultant total effectof system corflplexity on system success is-.016.

Effects of the developmentprocess variable on systemdevelopmentResource cons’traints on system developmenthave a significant, negative effect on successfulCBIS design and implementation, as expected(beta = -.494)i. This result is significant; hypoth-esis H5S is supportedi If finances and time avail-able are limited then the consequences for thesuccessful impiementation of the system are se-vere. Resource constraints lead to a substantial,though not significant, decrease in the involve-ment of users in the development process (beta= -.204); hypothesis H51 is not supported. Theindirect effect of resource constraints on systemsuccess is exacerbated by user participation ashypothesized (beta ="-.040), though the effectis not significant. This results in a total negativeeffect of resource constraints on successfulCBIS implementation of -.534. This effect is farlarger than any of the other effects because ofthe large primary effect.

Limitations of the StudyThe major limitation of this research is the timingof the survey With regard to the completion dateof system development. Data for the study werecollected afterlthe systems were implemented.Subjects’ responses may have been influencedby the ultimate:success or failure of the systems.The timing of data collection may have had asecondary infldence on the measurement of userattitudes. The relationship between user involve-ment and user attitudes may be circular in na-ture, i.e., if user attitudes are unfavorable, thereshould be a high degree of user involvement,which may in ,turn, lead to an improvement inuser attitudes.i We need to conduct longitudinalstudies to distinguish ~mong the effects that mayOccur.

104 MIS Quarterly~March 1988

User Involvement

A further limitation of the study is the use of in-struments that are not sufficiently validated --the perennial problem of a new discipline. Rigor-ous design and testing of research instrumentsshould precede studies such as these. Althoughwe attempted to use tested instruments through-out (e.g., system success and user attitudes),this was not always possible. A major limitationof the measurement scales employed in this re-search is the use of a single-item scale to mea-sure user involvement. A great deal of attentionneeds to be paid to the development of valid,reliable test instruments that can serve as thebasis for further investigations. Only when wereach some consensus on what we should bemeasuring and how it should be measured, willwe be able to make progress in synthesizing theresults of relevant research studies.

ConclusionsThe primary objective of our research was to in-vestigate the potential for adopting a contingen-cy approach to involving users in the system de-velopment process. The investigation took theform of an empirical study that assessed certaincontingencies for the. relationship between userinvolvement in CBIS implementation and the ulti-mate success of the CBIS. This section presentsthe conclusions of our research, as well as pre-senting some directions for future research.

Contingency theory has proved a usefulapproach for studying the effect of user involve-ment on system success. System complexityand resource constraints were found to havestrong effects on system success, either directlyor indirectly through their influence on user in-volvement. It is interesting to note that these aretechnical system and development process vari-ables, variables over which users have the leastcontrol. The information systems departmenthas greatest knowledge of the complexities ofthe system to be implemented. However, wefound that users are involved to the greatest ex-tent in implementing complex systems, i.e., man-agement reacts to system complexity when de-termining the degree of user involvement to em-ploy in a system development effort. It is interest-ing to speculate whether the cost of less complexsystem development may be reduced by de-creasing the degree of user involvement, whilestill achieving the same level of system success.

The results of this study indicate the importanceof ensuring that adequate resources are avail-

able for the development of a system. If re-sources are insufficient, system designers maynot follow normal development procedures. Thisstudy suggests that the risk of system failure in-creases as available resources are constrained.Two types of resource constraints were ad-dressed in this research -- time and finances.Subsequent research could address the relativeimportance of these two types of constraints tothe success of the development process. Macrocost estimating models, such as those advancedby Boehm (1983), Parr (1980), and Putnam(1978), suggest that one time factor that is criti-cal to system development is the implementationtime, rather than the total personnel time com-mitted to system development. Hence, the typeof research reported here could be extended totest this concept empirically, or alternatively, andthe cost estimating models could be used nor-matively to ensure that project elapsed time isnot one of the resource constraints that lead toCBIS implementation failure.

Financial constraints are imposed by financialmanagers. Financial managers should be madeaware of the extent to which a system may bejeopardized if they do not devote adequate re-sources to the development effort. This studysuggests further, that they should devote someof those resources to involving users in the de-sign and implementation process.

The relevance of user system variables must stillbe assessed. User attitudes in this study wererated so high that it is not possible to determinethe extent of the consequences on system suc-cess of negative user attitudes. Longitudinalstudies should be conducted to determine userattitudes at the time of development.

AcknowledgementThe authors are indebted to Maureen Lahiff forassistance with the statistics used in thisresearch.

ReferencesAjzen, I. and Fishbein, M Understanding Atti-

tudes and Predicting Social Behavior, Pren-tice-Hall, Engelwood Cliffs, N J, 1980.

AIIport, G.W. "Attitudes," in A Handbook of So-cial Psychology, C. Murchison (ed.), ClarkUniversity Press, Worcester, MA, 1935, pp.798-844.

MIS Quarterly~March 1988 105

User Involvement

Alter, S, "Development Patterns for DecisionSupport Systems," MIS Quarterly, 2:3,September 1978, pp. 33-42.

Alter, S. and Ginzberg, M.J. "Managing Uncer-tainty in MIS Implementation," Sloan Manage-ment Review, 20:1, 1978, pp. 23-31.

Asher, H.B. Causal Modeling, Sage Publica-tions, London, 1976.

Bailey, J.E. and Pearson, S.W. "Development ofa Tool for Measuring and Analyzing ComputerUser Satisfaction," Management Science,29:5, 1983, pp. 530-545.

Baroudi, J., Olson, M., and Ives, B. "An EmpiricalStudy of the Impact of User Involvement onSystem Usage and Information Satisfaction,"Communications of the ACM, 29:3, 1986, pp.232-238.

Boehm, B.W. Software Engineering Economics,Prentice-Hall, New York, NY, 1983.

Boland, R.J., Jr. "The Process and Product ofSystem Design," Management Science, 24:9,1978, pp. 887-898.

Carroll, A.B. "Behavioral Aspects of DevelopingComputer-Based Information Systems," Busi-ness Horizons, 25, 1982, pp. 42-51.

Cerullo, M.J. "Information Systems SuccessFactors," Journal of Systems Management,31:12, 1980, pp. 10-19.

Cronbach, L.J. "Coefficient Alpha and the Inter-nal Structure of Tests," Psychometrika, 16:3,1951, pp. 297-334.

DeBrabander, B. and Edstrom, A. "SuccessfulInformation System Development Projects,"Management Science, 24:2, 1977, pp. 191-199.

Dubin, R. Theory Building, The Free Press, Lon-don, 1978.

Duncan, O.D. Introduction to Structural EquationModels, Academic Press, London, 1975.

Edstrom, A. "User Influence and the Success ofMIS Projects: A Contingency Approach," Hu-man Relations, 30:7, 1977, pp. 589-607.

Ein-Dor, P. and Segev, E. "Organizational Con-text and the Success of Management Informa-tion Systems," Management Science, 24:10,1978, pp. 1064-1077.

Elizur, D. and Guttman, L. "The Structure of Atti-tudes Toward Work and TechnologicalChange Within an Organization," Administra-tive Science Quarterly, 21:4, 1976, pp. 611-622.

Fishbein, M. and Ajzen, I. "Attitudes and Opin-ions," Annual Review of Psychology, 23,1972, pp. 487-544.

Gallagher, C.A. "Perceptions of the Value of aManagement Information System," Academy

of Management Journal, 17:1, 1974, pp. 46-55.

Galletta, D.F. and Lederer, A.L. "Some Cautionson the Mea(surement of User InformationSatisfaction," unpublished paper, University ofPittsburgh, 1987.

Ginzberg, M.JI "Steps Toward More EffectiveImplementation of MS and MIS," Interfaces,8:3, 1978, pp. 57-63.

Ginzberg, M.J. "A Study of the ImplementationProcess," TIMS Studies in the ManagementSciences, 13, 1979, pp. 85-102.

Ginzberg, M.J. "Key Recurrent Issues in the MISImplementation Process," MIS Quarterly, 5:2,1981a, pp. 47-59.

Ginzberg, M.J. "A Prescriptive Model of SystemImplementation," Systems, Objectives, andSolutions, 1:1, 1981b, pp. 33-46.

Guthrie, A. "Attitudes of User-Managers To-wards Management Information Systems,"Management Informatics, 3:5, 1974, pp. 221-232.

Hopelain, D.G. "Assessing the Climate forChange: A Method for Managing Change in aSystem Implementation," Systems, Objec-tives, and Solutions, 2, 1982, pp. 55-65.

Hull, H.C. and Nie, N.H. SPSS: Update 7-9,McGraw-Hill~ New York, NY, 1981.

Ives, B. and Olson, M.H. "User Involvement andMIS Success: A Review of Research," Man-agement Science, 30:5, 1984, pp. 586-603.

Ives, B., Olson, M.H., and Baroudi, J.J. "TheMeasurement of User Information Satisfac-tion," Communications of the ACM, 26:10,1963, pp. 785-793.

Keen, P.G.W. "Information Systems and Organi-zational Change," Communications of theACM, 24:1, .1981, pp. 24-33.

Kolb, D.A. and[ Frohman, A.L. "An Organization-al Development Approach to Consulting,"Sloan Management Review, 12:1, 1970, pp.51-65.

Lewin, K. Field Theory in Social Science, Harper& Row, New York, NY, 1951.

Locke, E.A. and Schweiger, D.M. "Participationin Decision-Making: One More Look," Re-search in Organizational Behavior, 1, 1979,pp. 265-339;

Lucas, H.C., Jr. "A Descriptive Model of Informa-tion Systems in the Context of the Organiza-tion," Data Base, 5:2, 1973, pp. 27-36.

Lucas, H.C., Jr. Why Information Systems Fail,Columbia University Press, New York, NY,1975.

Lucas, H.C., J~. "Performance and the Use of anInformation ’System," Management Science,

106 MIS Quarterly/March 1988

User Involvement

21:8, 1975, pp. 908-919.Lucas, H.C., Jr. "The Implementation of Comput-

er-Based Models," National Association of,Accountants, New York, NY, 1976.

Maish, A.J. "User’s Behavior Toward His MIS,"MIS Quarterly, 3:1, March 1979, pp. 39-52.

Mann, R.I. and Watson, H.J. "A ContingencyModel for User Involvement in DSS Develop-ment," MIS Quarterly, 8:1, March 1984, pp.27-38.

Markus, M.L. "Power, Politics, and MIS Imple-mentation," Communications of the ACM,26:6, 1983, pp. 430-444.

McFarlan, F.W. "Portfolio Approach to Informa-tion Systems," Harvard Business Review, 59,September-October 1981, pp. 142-159.

Middlemist, R.D. and Hitt, H.A. OrganizationalBehavior, Science Research Associates Inc.,New York, NY, 1981.

Morse, J.J. and Lorsch, J.W. "Beyond TheoryY," Harvard Business Review, 48, May-June1970, pp. 61-68.

Mumford, E., "Consensus System Design: AnEvaluation of this Approach," in Design andImplementation of Computer-Based Informa-tion Systems, N. Szyperski and E. Grochla(eds.), Sijthoff and Noordhoff, Groningen, Hol-land 1979.

Nie, N.H., Hull, C.H., Jenkins, J.G., Stennbren-ner, K., and Bent, D.H. Statistical Package forthe Social Sciences (2nd ed.), McGraw-Hill,New York, NY, 1975.

Nunnally, J.C. Psychometric Theory, McGraw-Hill, New York, NY, 1978.

Olson, M.H. and Ives, B. "User Involvement inSystem Design: An Empirical Test of Alterna-tive Approaches," Information & Manage-ment, 4, 1981, pp. 183-195.

Parr, F.N. "An Alternative to the Rayleigh CurveModel for Software Development Effort," IEEETransactions on Software Engineering, SE-6:3, 1980, pp. 291-296.

Pearson, S. "Measurement of Computer UserSatisfaction," unpublished doctoral disserta-tion, Arizona State University, 1977.

Pedhazur, E.J. Multiple Regression in Behavior-al Research: Explanation and Prediction, Holt,Rinehart and Winston, New York, NY, 1982.

Powers, R.F. and Dickson, G.W. "MIS ProjectManagement: Myths, Opinions, and Reality,"California Management Review, 15:3, 1973,pp. 147-156.

Putnam, L.H. "A General Empirical Solution tothe Macro Software Sizing and Estimating

Problem," IEEE Transactions on Software En-gineering, SE-4:4, 1978, pp. 345-361.

Robey, D. "User Attitudes and Management In-formation System Use," Academy of Manage-ment Journal 22:3, 1979, pp. 527-538.

Rokeach, M. "The Nature of Attitudes," Interna-tional Encyclopedia of the Social Sciences,MacMillan~ New York, NY, 1968.

Schein, E.H. "The Mechanisms of Change," inInterpersonal Dynamics, W.Bennis, E. Schein,F. Steel, and D. Berlew (eds.), Dorsey, Home-wood, IL, 1964, pp. 362-378.

Schewe, D.C. "The Management InformationSystem User: An Explanatory BehavioralAnalysis," Academy of Management Journal,19:4, 1976, pp. 577-590.

Schonberger, R.J. "MIS Design: A ContingencyApproach," MIS Quarterly, 4:1, March 1980,pp. 13-20.

Shaw, M.E. and Blum, J.M. "Effects of Lead-ership Style Upon Group Performance as aFunction of Task Structure," Journal of Per-sonality and Social Psychology, 3, 1966, pp.238-242.

Simon, H.A. The Sciences of the Artificial (2nded.), The M.I.T. Press, Cambridge, MA, 1981.

Swanson, E.B. "Management Information Sys-tems: Appreciation and Involvement," Man-agement Science, 21:2, 1974, pp. 178-188.

Treacy, M.E. "An Empirical Examination of aCausal Model of User Information Satisfac-tion," CISR Working Paper, M.I.T., 1985.

VanLommel, E. and DeBrabander, B. "TheOrganization of Electronic Data ProcessingActivities and Computer Use," The Journal ofBusiness, 48:3, 1975, pp. 391-410.

Vroom, V.H. and Yetton, P.W. Leadership andDecision-Making, University of PittsburghPress, Pittsburgh, PA, 1973.

Wolek, F.W. "Implementation and the Process ofAdopting Managerial Technology," Interfaces,5:5, 1975, pp. 38-46.

Zand, E.E. and Sorensen, R.E. "Theory ofChange and the Effective Use of ManagementScience," Administrative Science Quarterly,20, 1975, pp. 532-545.

Zmud, R.W. and Cox, J.F. "The ImplementationProcess: A Change Approach," MIS Quarter-ly, 3:2, June 1979, pp. 35-43.

Zmud, R.W. "An Exploratory Study of User In-volvement Role Sets," working paper, Schoolof Business Administration, University of NorthCarolina, Chapel Hill, NC, 1981.

MIS Quarterly~March 1988 107

User Involvement

About the AuthorsPeter Tait is currently a consultant in the Ad-vanced Technology Group of Touche Ross &Company, Brisbane, Australia. He received hisBachelor of Commerce and Masters of Informa-tion Systems degrees from the University ofQueensland, Brisbane, His current fields of in-terest are the development and implementationof information systems.

Iris Ves~ey is kssociate Professor in informationsystems at the Katz Graduate School of Busi-ness, Universit# of Pittsburgh. She received thePh.D. degree f~om the University of Queenslandin 1984. Her research interests focus on humanfactors in systems development. They includethe types of debomposition essential to handlingcomplexity in Systems design, the nature of theexpertise requilred to design and maintain com-puter-based systems, and the presentation of in-formation as a function of task type.

11)8 MIS Quarterly~March 1988