OMEGA Int. J. of Mgmt Sci., Vol. 13, No. 5, pp. 393~106,1985
Printed in Great Britain. All rights reserved
0305-0483/85$3.00+0.00 Copyright© 1985PergamonPress Ltd
Systems Analysts and the Politics of Organizational Control MICHAEL NEWMAN DAVID ROSENBERG University of Manchester, UK (Received November 1984; in revised form March 1985) The introduction of major information systems into organizations can be fraught with difficulties. The negative effects of such systems can, for many important organizational members, be too great a price to pay for the supposed benefits. The advent of computerised information systems has served not merely to deskill the clerical 'craft' but can also be used as a strategy by top management to reduce the management control gap. However, by undercutting the organizational 'rules of the game', information systems innovation can produce severe strains and stresses in the political system. The proposals in the literature to combat the effects of innovation are examined. A broader, political perspective is used to reveal some of the inadequacies of these proposals and supporting data, gathered from interviews in several Canadian organizations, are presented to illustrate this. The paper ends with a discussion of the implications of this broader approach.
INTRODUCTION
be inaccessible to management actors, and a culture which is alien to the organization, then IT IS A BANAL truism that organizations are troubles begin to inevitably accumulate. Estabboth strong and surprisingly fragile and never lishing a career path so that such a strata fixes more so when faced by the necessity to meet new its loyalties to the 'dominant coalition' of the challenges and pressures from the environment. organization without alienating other layers of It is a paradox that the willingness to attempt management is by no means an easy task as it to gain some measure of control over the envi- is likely that such a strata may well be pulled ronment which will reduce the level of external into 'cosmopolitan' rather than 'localist' loyalthreat can by itself bring about other internal ties with consequences for their role retensions in the life of the organization [8]. Thus, lationships [21]. The dialectic of the search for the suspicion that some organizational actors work autonomy and managerial control is unimay believe the cure is for them (if not the versal in those organizations which have to organization) more troublesome than the origi- integrate such new technology and its halfnal complaint can hold some empirical truth. captive 'priests'. The solutions vary considThe new technology of control and the various erably as the balance internal to the orlevels of technicians who service it may well be ganization, its service outputs, the nature of the regarded with a certain hidden ambiguity by blend of the internal cultures differs from orthose who, on the surface, welcome it and, when ganization to organization. However, few studthe technology is information systems which ies probe these complexities other than from the threaten to transform management from an art perspective of an abstract rationality model of to a positivist science of control and prediction, behaviour which merely notes deviances and the ripples in the managerial sea can become labels such behaviour as pathological. An imwaves. Superimpose on this a group of tech- portant development in organizations is the nicians who are relatively well educated, possess increased differentiation within the managerial a 'secret' language of their trade which may well ranks which has resulted from the growth of 393
394
Newman, Rosenberg--The Politics of Organizational Control
functional specialization. Management is neither a monolithic entity nor a single occupational role, but a range of separate specialisms, which deal with different aspects of the managerial function. Such differentiation has led to a distinction between line and specialist managerial roles, to the proliferation of different departments concerned with aspects of work control, and has created problems of control within the managerial group itself. This has stimulated the improved quality of the information with which managers work. Cost accounting systems and financial controls produce rapid and centralised information about the organization, which, prior to their development, was simply unavailable or too complex to be easily collected and assimilated. With the advent of computerized financial and production data the most senior levels of management can now monitor what is happening among the lower levels of the organization "to the extent that such data translate social behaviour into units of financial and production accounting" [24]. The dysfunctional effects of introducing comprehensive information systems into organizations are widely known if poorly understood [see for example, 1, 2, 29, 31]. These dysfunctions can emerge in organizations as aggression or sabotage. Thus operating personnel may deliberately distort or destroy input data such as time cards or production control information [1]. Marcus [31] in her description of the design of a financial information system for a large private company describes how the corporate accountants provided the momentum for introducing a new system as a way of tightening up control of an organization which was highly divisionalised. Markus documents how the divisions fought co-operation with the new system attacking it for its design, technical adequacy ~Throughout this paper the expression, "Management Information System" (or MIS) will refer to any computer-based information system. "Information system" will often be used as an abbreviation. The MIS is the product of the "MIS group" which some authors refer to as the EDP department (EDP standing for electronic data processing). The MIS group consists of "Systems Analyst", "Programmers", "operators" and management and other staff. The systems analyst is the person who designs and implements computer-based information systems, and it is his task we will be focusing upon. In this paper he will generally be known as the "system designer" or, more simply, "designer" or "analyst".
and feasibility. The process dragged on for years, and attempts were made by the divisions to sabotage the system. More frequently, employees may, where possible, avoid or by-pass the new system. Instead of using reports generated by the new system, a manager might develop and use his own informal information sources [36]. In either case the new system can be classed as a political failure by top management. The McKinsey report [32] indicated that this phenomenon was widespread even in 1969. Many organizations are finding the implementation process not merely complex but that it amplifies and reproduces strains in the internal political system.
Emergence of systems analysts It is important to look at the spectacular rise in importance of the MIS l group in its very short history as a means to understanding its current position within organizations. The majority of the first computers sold commercially were located in the accounting departments of business organizations. The kind of applications they handled were generally simple, mechanistic book-keeping jobs such as general ledger, accounts payable/receivable, and payroll. The people who programmed and ran the machines were usually highly intelligent mathematics majors. (For example, see [20, 37]). There were no compilers, so all programmes were written in machine code, or later, in assembler language. Naturally, the programmers were key personnel, highly paid, highly valued with considerable autonomy from close supervision at work. The office and methods experts had to comfort themselves with the responsibility for the manual part of the system.
The deskilling of programmers Gradually, the computer department became an entity in itself, and, with the rise of 'higher' level computer languages and more reliable computers, the power structure and role relations changed. A new group emerged called 'systems analysts' who took on the task of designing and implementing the information system. The programmers became subordinate to the analysts [12] and worked under greater direct managerial control. Hughes has commented how high status occupations have a tendency to pass on low status 'dirty work' to other occupations. This is the counter tendency to
395
Omega, Vol. 13, No. 5 Table 1. Firm (number of interviewees
Alpha (4)
Beta (4)
Category
Private Licensed by Government
Public utility
imperialist claims by rising occupations [25]. The discarding of 'impure' work activities by systems analysts in their rise to a separate and highly insulated occupation has meant that the control of the 'dirty work' of programmers is both seen as necessary but marginal to the core role attribute of system designers. Programmers then, in turn, become subject to the attempt to exercise greater control by managerial groups which downgraded their formerly high occupational status [28]. To that extent, while deskilling of programmers is relative rather than absolute the external control dimension and the attempt to pattern and regulate programmers means a more concentrated attempt to impose bureaucratic authority of the Weberian type. Whereas in the 1940s and early 1950s the data processing industries displayed craft like characteristics as installations were fairly small and the tabulating craft worker operated all the machines, the new division of labour fragmented the 'craft' elements (see Greenbaum's radical analysis [20]). A fairly highly rigid hierarchy developed with startlingly different levels of pay and a marked concentration of knowledge and control in a minute proportion of the hierarchy. In most MIS groups today, the communication between the analysts and the programmers is largely through the programme specification by which the analyst details the particular problem to be solved by the programmer. In this emergent and unequal relationship we will confine our attention to the role authority of the system designer for the remainder of this paper. SYSTEMS DESIGN AS A POLITICAL PROCESS It is difficult to read the literature on systems design/implementation without drawing the conclusion that the process is highly political with the various actors struggling for resources 2The number of employees was greater than 2000 in each case.
Gamma (4)
Private wholesale
Delta (4)
Epsilon (1)
Private primary resource
Public, crown corporation
and status [5, p. 139]. The struggle is aggravated by the contrasting values between the designers, who emphasize a technicist approach, and the various strata of management who have to adapt to innovation. Whilst the effects of the struggle may be mitigated by the universal advocation of user-involvement in design and implementation, in practice the benefits of participation are only marginal. Finally, the conflicts between designers and managers are brought sharply into focus by the highly rationalistic top-down approach often used in largescale information system innovation. Whereas some of the effects are minimalised in smallscale systems involving individual managers and departments, the disruption caused by large systems can be too much for the political system to cope with. To gauge the significance of the above issues, a series of interviews was conducted in five large2 corporations in Vancouver, Canada with systems analysts and managers. The types of companies and the number of interviews are given in Table 1. While a sample of five is clearly insufficient for generalising to a larger population of companies and systems analysts, the similarities among the comments were considered to merit reporting as an exploratory study in this area. Top management involvement in the MIS group has been considered to be important (see [23] for example). Only at Gamma was top management instrumental in introducing a systems' design aid. Such an aid, when used properly, makes the analyst involve the user in the design/implementation process. From the interviews it seemed that the main reason that the aid was promoted was not to encourage user involvement but to increase top management's control over projects, particularly in the areas of documentation and budget estimation, two areas that were deficient in that particular firm. If there was little evidence of top management involvement with systems analysts, there was a plethora of examples of how analysts perceive
396
Newman, Rosenberg--The Politics of Organizational Control
their work and the people they work with. In addition, their backgrounds were frequently technical. A technicist orientation
We interviewed an analyst at all firms except Epsilon. The backgrounds of the four were significant and interesting. At Alpha the analyst we spoke to had a thorough training in business before getting involved with systems' work. The people at Beta and Delta had both gone straight into systems' work from University. The man at Gamma had had a technical background gained by 18 years of work in one of the armed forces which had involved some programming. It was clear that most of the analysts could do some programming but that was not their major concern. We got a strong sense, however, that several of the analysts interviewed were exceptional in their backgrounds. Most of their colleagues had become analysts through the programming route. For instance, the analyst at delta said, in answer to the question: Interviewer: "Were you once a programmer that progressed to systems analyst?" Analyst: "No. I went into it immediately, I never did programme". Interviewer: "Is that rare here?" Analyst: "Yes it is. I think everyone in the department has programmed before".
The same situation was found at Alpha where the analyst saw himself very much in contrast to his colleagues: "They're always saying to me, 'What the hell do you do, you're never in your otfice?' I said, 'Of course I'm not, I've got 5,000 users out there: they're the people I'm working for . . . . ' "
Indeed in answer to a question about the criticism of MIS people getting too technical, he replied; "That's fight. The reason for that is because if you look around here you find that every single one of those guys is a programmer."
At Beta, the analyst estimated that over 50~o of his colleagues had programming as a background. In general, it seemed that the analysts interviewed were atypical. Most of them had not been programmers, whereas the majority of their colleagues had progressed through that discipline. This result was confirmed by the Data Processing managers who indicated that the path for promotion to the analyst position was generally through programming, which supports previous findings. This should be
borne in mind when considering the comments of the systems analysts. Knight [27] says that MIS specialists are essentially 'agents for change': firstly, in the creation and development of an idea, and then secondly, in the introduction and adoption of that idea. But currently, the emphasis is almost entirely on the former skill to the neglect of the latter. As Enid Mumford [34] notes: " . . . h u m a n relations and the ability to manage the human part of change are skill areas which are dealt with intuitively and unsystematically and this is in striking contrast to the technical side of computer systems where programmers and systems analysts continually stress the need for a logical approach."
This observation is borne out by other studies. Bjorn-Anderson and Hedberg [5] found a similar emphasis on technology by design teams. In their study of two banks both implementing large information systems they concluded that technological considerations dominated the design process from beginning to end. Both design teams put user attitudes bottom of their list of priorities in the formation of policy. The design of jobs and organizational structures were considered to be the responsibility of others. The authors account for this by pointing to the designer's fascination for technology and problem-solving. They conclude that a kind of self-selection is going on. People who become fascinated with technology choose positions where a technicist emphasis is required and rewarded. Many of today's designers were once programmers at an earlier stage in their career, a factor contributing to the technicist orientation of a large number of analysts and their possible perception of the implementation function as an aspect of 'dirty work'. In an earlier study of the implementation process, Mumford [34] found a great gulf exists between the systems designers and the clerical staff. The users told her: "They (the systems designers) don't give us much information. They don't tell us what is going to happen. It just happens and we are told about it." "They just give you a book to read. There is little explanation of the new things that are happening."
This technical emphasis is also observed in both the training of system designers and the literature surrounding MIS. In training, little effort is devoted to educating the system analyst in sensitivity to organisational variables [5]. Several authors have noted the disruptive behaviour of systems design staff in organizations.
Omega, VoL 13, No. 5
Rose has noted the differences in values and goals of designers and managers: "Conflict is structured into their relationship.... Resentment of the innovation is aggravated by resentment of the innovator: resentment of his rewards, his values, his manner, his vocabulary, his prestige as change-maker, his apparent freedom from 'normal company discipline', his promotability, his dress, his fickle loyalty to the firm, and, not least, his youth." [41, pp. 169-170] However, such disruptive behaviour could merely mean that what patron-client relationships exist that link designers to senior management does not strongly affect the new occupational culture and its attitudes. The assertiveness o f this new work category and its claims are perceived by other sections of management as claiming status and authority, which is not that o f mere hired hands. Emerging professions and sub-professions must inevitably find that they eat into claims made by dominant managerial groups and their auxiliaries for control in organizational contexts. The technical bias of the systems designer is probably the most cited reason in this variant of the human relations tradition for the dysfunctional behaviour observed. It is not surprising, therefore, that the most frequently proposed solutions concentrate on mitigating this bias in some way. For example, Enid Mumford [34] suggests: "If computer innovation is to continue to thrive then it seems important to bring the value position of the computer man somewhat closer to that of our 'civilized man', so that he no longer defines his role in purely technical terms." What is needed, it is claimed, is to get designers thinking in terms of "socio-technical systems" [18]. In this regard several authors concentrate on the background of the designers. This seems quite logical but is also quasi utopian in its implications that if the designer has a propensity to tinker with technology then a broader training in 'behavioural factors' is a worthy solution. A similar proposal is to recruit personnel from different backgrounds to the design profession, or to introduce professionals skilled in behavioural and organization factors into the design process [4, 5, 34]. The emphasis is on the civilizing through a low-level applied social science of an elite which to the contrary, believes in its own mission to civilize. This humanization ideal is part of a broader human relations' school.
397
In these human relations solutions the designers and not management are the target for criticism. Structural organizational conflicts are marginalized. "There is little recognition here that the data processing man's success may depend not on the line manager becoming more conversant with his language, but of his becoming more understanding of the problem and needs of the line manager." [34] Presumably an awareness of the essential unity of management vis d vis the control of the total labour process is supposed to more than balance the internal tensions of the management of management. She explores this further in discussing the power imbalance that exists between the MIS expert and the line manager. But, as Mumford acknowledges, currently, there is no structural incentive for the stronger party to change their patterns which lead to middle level resistance. Instead, she advocates a process of 'civilization' so that the designer no longer defines his job in purely technical terms. This has a parallel with the belief that the mission of the gentry was to civilize the 19th century industrial bourgeoisie through manners and culture. User involvement
It was revealing to see how the analysts worked with users in our study. At Gamma, the analyst was having some trouble designing a system: Analyst: "... I had one person, one female who had been in the company 10 years at that time, and she was the top person. Boy did she resist! Took a long time to win her over." Interviewer: "What kind of things did you do to win her over?" Analyst: "I kept asking her advice.... I had already figured it out but I wanted her to do it... and so on, as if she really was somebody. She was, in her own mind, because she had been there longer than anyone else, but as far as her manager was concerned she was just one of the girls.... But you have to work on these people, butter them up." This was given as an example of user involvement through pseudo democracy and manipulation. Apparently the design aid used in this firm had such involvement required throughout its stages. Clearly, though, the analyst was given complete freedom to interpret what was meant by 'involvement'. The analyst at Delta agreed that their design aid forced them to consider the users more by involving them in design and implementation. In fact he consid-
398
Newman, Rosenberg--The Politics of Organizational Control
ered 'People Skills' to be his most important characteristic: "I think it's the most important thing in this job. i f you can't relate well to a user, you can't draw out of him what he needs. Everything else you do is a waste."
At Beta, the analyst cited 'communication' as the most important skill. At Delta, the analyst was dealing with a wide range of users with differing experience in D.P.: " . . . you can go from the users who have no experience at all with data processing and trust you absolutely to do whatever you think is best for h i m . . , up to the user who has (had) a great deal (to do) with data processing projects in the past and is dogging you constantly to make sure you are not snowing him or giving him any bunk, and everything in between."
The usual relationship with the users seems to consist of the analyst in charge with the user contributing according to his skill and experience. The amount of user-involvement seemed to be largely up to the analyst. Most of the projects under the analyst at Delta were judged by him to be successful. The exception was the first project he worked on: "the user manager requested work to be done but didn't encourage much contact with people who have actually been using the output and also semi-withdrew themselves from what was happening. So the result of the project was a lot o f guess work by the analyst as to what was really wanted."
Hedley [23] found a similar result in his study where lack of management definition caused the system analyst to step in and fill the gap. Did the users not co-operate? The analyst at Delta again: " . . . t h e y pretty well had to co-operate because we replaced one system with another and they had no choice. They were complaining. A lot of grumbling went on for years for a good reason too because a lot of the stuff they got just wasn't suitable."
The system seemed to be technically acceptable; the users just did not use parts of it. Perhaps the most interesting case was at Alpha where they were implementing a large, materials management system over 2 years using the top-down approach. The analyst interviewed was the senior analyst on the project. He was obviously an exceptional person and, as stated elsewhere, he saw himself in contrast to his colleagues. For instance, he was never in his office. "What do you need an office for? I need somewhere to put my papers . . . . My job is out in the field."
He was also concerned to involve the users in the process:
"We have a very large user involvement. I like to do it that way as much as possible. Let them design it even, subject to us then going over it from a technical viewpoint."
It was clear that top management was supporting the project heavily. They had secured 35 users who normally had nothing to do with data processing, but now spent 75~ of their time on the project. Over 500 people in the firm had been interviewed. The analyst was spending a lot of his time in giving presentations to managers and others. Also the whole training process had been carefully considered: ' L . . we're going to let them do the training. We're going to train the managers and it's up to the managers and liaison to appoint somebody to train their own g r o u p . . , if a D.P. team goes out and starts trying to train people directly, you'll have a hell of a job on your hands."
His philosophy was to delegate training to the managers. They were the ones who should know their employees and therefore should be able to overcome any fears personnel might have of computers. In the context of information systems, BjornAnderson and Hedberg [5] give three reasons for involving client users in the design process. First, it develops supposedly 'realistic' expectations and reduces resistance to change (although the authors question the ethics of manipulating the users just to produce an effective system). Secondly, the user probably knows his job better than anyone else. If the designer needs user-expertise, then the obvious solution is to involve the user in the design process. Lastly, the authors advocate participation as advocates of liberal-democratic values. As they put it: "Because organizations ought to respect human dignity and basic human rights, all members should have a right to participate in decisions about changes that are likely to affect them." [5, p. 135]
The authors also suggest various tactics of getting users involved in the design process. They place at one end of the spectrum a passive involvement, with the designer taking all the initiative; at the other end, the designer teaches the users how to design, leaving the user effectively in control of the process with the designer as a technical consultant. There are many other methods of user participation and it is not the purpose of this paper to examine them in depth. Suffice it to say that the majority of writers in this area advocate user participation
Omega, Vol. 13, No. 5
as the main solution to the dysfunctional effects of systems design and implementation. [See 10, 15, 30, 32, 44]. Manipulation and pseudo participation are also wide spread practices, though the liberal democratic norms of the participation literature does not welcome such bad faith in organizational practices [42]. The literature requires the user to build his own system of democratic control and this, in turn, is possibly a form of manipulative rationality through democratic values. Additionally, a long-cited cause of dysfunctional behaviour resulting from new information systems is lack of participation in the control of the design process [6, 44]. We have seen that dominant coalitions can exhibit a tendency to avoid participation in the design process. Dickson and Simmons [15] claim that much of the blame for failure in information systems can be attributed to top management's strategic withdrawal. Hedley quotes one managing director as saying:
399
also be seen in the way the MIS group claim to authority is incorporated into actual practices. As we have argued, designers singled out technology as their major area of responsibility and this was often recognized as a legitimate claim by at least some levels of management. This is illustrated in the Bjorn-Anderson and Hedberg study: "The design teams were not rewarded for considering human needs, and budgets and time schedules were not planned to allow capacity for work design or reorganizing. Neither did they encourage---or even enable---members of the design teams to spend time evaluating and learning from the information systems they had installed. When one system was technically completed, another design project was waiting." 15, p. 133]
If the 'dominant coalition' of management is unable to dominate and determine the socialisation into the political system and cultures of the organization then the technicist orientation of designers will disrupt and distort the very implementation they welcome. However, the data could also be analysed by " . . . We found out too late that top management must noting the advantages of such a strategic withbe involved for the thing to be a success. It wasn't the Computer Director's fault; it was the failure of top drawal by top management as the informal management to realize that the computer is just as much alliance of interests between systems analysts the responsibility of top management. We didn't realize and top management at the expense of middle the implications quickly enough." [231 level managers. At middle or operation level management the same lack of participation is In matters of authority, systems personnel observed, but, whereas top management can have sometimes been able to move into areas of probably function without more than a cursory policy where managers fear to tread. But this is acquaintance with the new system, this is not not always a long-term successful strategy. true of middle management. They will be the Sooner or later the power resource base in this prime users of the reports of the new system area forces different layers of top management (e.g. stock control, production control and paysometimes reluctantly to get involved; the prob- roll), and they in turn will be potentially more lem is that it is usually too late to affect the subject to monitoring and control. Dickson and system's design process substantially. Hedley Simmons note that: [23] noted in his study how the systems person"When operating managers do become involved with nel filled the vacuum in authority left by mansystems (usually after a great deal of coercion), their behaviour often takes the form of aggression or agement: "What tended to happen in the absence of definition was that the people responsible for systems analysis made policy decisions about the various controlling factors almost without being aware of doing so. There was therefore the danger of long-term decisions being made in terms of their appropriateness to system design rather than in relation to more fundamental criteria. Thus real control tended to shift to the systems designers."
If the diverse layers were unable or unwilling to utilise the new system of formal rationality of the MIS, then the designer category will, by default, assume that authority even if individual members are reluctant to do so. This ambiguous refusal to take responsibility for the design can
OME. 13/5- C
projection."
[151 Up to this time, these managers tried to avoid any participation in the design process. It follows that if a clerical or operations supervisor or a departmental manager is known to avoid design participation, we should not be surprised to observe similar behaviour at the employee level. If management is not going to insist that design teams involve employees in the design, then participation is going to continue to be the exception rather than the rule. The opening of this section gave a scenario of some users'
400
Newman, Rosenberg--The Politics of Organizational Control
reactions to a new system when they h a d no chance to p a r t i c i p a t e in the design. I n the ' b a n k s ' s t u d y the a u t h o r concluded: "The projects did not take the users' interests into consideration. Most users were represented neither directly nor indirectly in the design processes, and the technological specialists had conceptions of users and users' needs that were not shared by the users themselves," [5, p. 1311 T h e y also f o u n d a similar lack in r e g a r d to training the users o f the new system: "... computer programming continued for several years, but the programming of people for new tasks--in the form of training--took place only during the last few weeks just before the technology was installed in branch Ofllces."
[5, p. 1281 Clearly, user p a r t i c i p a t i o n a n d task t r a i n i n g are closely allied. I f a designer does n o t e n c o u r a g e p a r t i c i p a t i o n , he is unlikely to e x p e n d a g r e a t deal o f effort to train the user to use the new system. H o w d o m a n a g e m e n t users view the systems analysts? G e n e r a l l y , in o u r s t u d y they seemed quite satisfied, a l t h o u g h there h a d been s o m e conflicts in the past. The m a n a g e m e n t user at Beta was p e r h a p s typical. In answer to a question a b o u t his r e l a t i o n s h i p with the c o m p u t e r d e p a r t m e n t he replied t h a t it was excellent. The p r o b l e m s h a d c o m e f r o m the user a n d a n a l y s t ' m i s u n d e r s t a n d i n g ' each other: "... I sat down for two hours of the session with the computer analyst.., at the end of the two hours I've had to say, 'Well if you're saying that now, you didn't hear what I said two hours ago.' And he's had to say the same to me. It's not one way, it's both ways. . . . I found another thing I've been able to do recently (and almost had head-on fights (over it)) (from what) we wrote out of the meetings, you'd think we both had been in different meetings." T h e m a n a g e r h a d learned by experience to ' n e g o t i a t e ' his e n v i r o n m e n t with the analyst. But this h a d been gained at some cost to himself: Manager: "... In order to get over the language barrier even though it might irritate them by my, in effect, doing a pre-design of the thing, I tried to see if I could use their terminology to explain to them what it is I want . . . . " Interviewer: "What was their reaction to that approach?" Manager: "Oh, polite," Interviewer: "Patronization?" Manager: "Yes, the individuals reacted to me with polite silence and grins on (their) faces... I'm sure there are analysts who consider it is their job to analyse the problem and come up with the answer and the hell with the end user as to what he really wants... But that was just with the first system. The other ones we have gone backwards and forwards, backwards and forwards and we've had compromises."
F i n a l l y , several writers p r o p o s e s o l u t i o n s t h a t require the user to change. Lucas, for e x a m p l e , suggests that: "Users have the responsibility to learn about information systems, contribute to their operation and development, and participate in making intelligent decisions about them." [301 T h e s a m e theme is t a k e n up in the b a n k s ' study. Users m u s t be m o r e r e a d y to p a r t i c i p a t e in design processes. T h e r e the a u t h o r s t a k e the 'user p o w e r ' c o n c e p t further: they a d v o c a t e m o r e d i s s e m i n a t i o n o f expert k n o w l e d g e in laym a n s ' terms (books, lectures, etc.). T h r o u g h collective n e g o t i a t i o n s users can then influence the design process. A N o r w e g i a n case is q u o t e d where several t r a d e unions have r e g u l a t i o n s g o v e r n i n g the use o f i n f o r m a t i o n a l systems so t h a t n o n - m a n a g e m e n t users are effectively represented in the process. It s h o u l d be noted, however, t h a t unions which represent supervisory staff have a l l o w e d such 'users' to restrict the a r b i t r a r y p o w e r o f t o p m a n a g e m e n t . T h e S c a n d i n a v i a n b a n k s ' s t u d y is, p e r h a p s , typical o f the solutions a d v o c a t e d : "Training, diversity, and consciousness raising, Jtogether with more diverse performance measures and supportive reward systems, can enable and encourage design teams to design with both human needs and technological possibilities in mind. Users must improve their resources to influence design processes, mere participation is not enough. Increased self-reliance, access to expertise, and legislative support, which grants users the right to share design activities, can strengthen users' influences over design processes. Participative designing can contribute to better designs when design teams share the major objectives. When there are many conflicting interests, information systems designing should be seen also as a political process, and designs should emerge as compromises between different organizational subgroups." [5, p. 139]
Top-down design--an imposed rationality T h e t o p - d o w n p r o j e c t at A l p h a was p a r t i c u larly rich b e c a u s e it involved m a n y o f the aspects a l r e a d y m e n t i o n e d as well as illustrating several new issues. " T o p - d o w n " uses a l a y e r e d c o n c e p t u a l a p p r o a c h to design which tends to d o w n - p l a y the existing m o d e s o f o p e r a t i o n in f a v o u r o f a r a t i o n a l solution. Such s o l u t i o n s c a n b r i n g a great u p h e a v a l to an o r g a n i z a t i o n . W h e n A l p h a a g r e e d to the p r o j e c t it c a m e as a r e c o g n i t i o n o f the i m p o r t a n c e o f m a t e r i a l s m a n a g e m e n t to their business. T h e first thing the a n a l y s t d i d was to ' c r e a t e the e n v i r o n m e n t ' for the system. W i t h such a large p r o j e c t it was
Omega, Vol. 13, No. 5
n o t surprising that several 'illogicalities' came to light in the old system [2]. First o f all they i n t r o d u c e d physical c o n t r o l s o n stock, something that had been lacking u p to that point. T h e n the a n a l y s t discovered vast discrepancies in the a m o u n t o f stock in h a n d . O n one item they held the e q u i v a l e n t o f 14,000 years o f stock whereas they carried only a b o u t 'ten m i n u t e s ' w o r t h of a high t u r n o v e r piece. There was at least 6 m i l l i o n dollars w o r t h o f excess stock: "... that is an enormous loss and I know damn well there is another 30-40~o that is unknown until we do the inventory around here, probably as we put the system up." H o w did they order stock u n d e r their old system? "They talk about it, a wonderful way of calculating the order quantity. It's 3 times the lead-time usage. Well, I mean on basic principles!.., they double everything!" I n addition, the staff had devised ways o f circ u m v e n t i n g the system: "What interested me was how much of this was going on. . . . It was awful! I couldn't believe the figures. I had to run the thing twice to make sure we hadn't made a mistake in the programme!" U n d e r the rules o f the c o m p a n y , cheques over a certain a m o u n t ($25,000) h a d to be s a n c t i o n e d by the board. The a n a l y s t had discovered people in supplies who, w h e n they discovered a discrepancy in the stock levels, w o u l d happily write-off in excess o f a million dollars in one day! As a final example, the purchase order form had to be redesigned to suit the new e n v i r o n m e n t ; previously they had been using different forms for different jobs. Clearly, if the purchase order is p r o d u c e d o n the lineprinter it would be beneficial to rationalize its design. The a n a l y s t explains: "We designed one form to cover all purchasing. . . . but then discovered.., we've cut printing costs by about 75~o..." Once the m a n a g e r s realized some o f the p o t e n tials for savings they were o n to him to do similar work in other areas. A l t h o u g h the t o p - d o w n a p p r o a c h held o u t great potential for the savings m e n t i o n e d above, it involved some i n h e r e n t p r o b l e m s with employees. The project m a n a g e r explained this very well: "Top-down is a troublesome approach for the users because the user is used to his function and so that means his view of looking at it is bottoms-up, really. Now when we look at it tops-down, especially in the initial stages,
401
he's busy saying, 'that's a function but where does my thing fit into this?'... It's a very difficult thing for him to relate to." F o r the t 0 P - d o w n a p p r o a c h to be successful, c o - o p e r a t i o n is n o t optional: it is essential. Sometimes, however, people did n o t w a n t to co-operate a n d the systems a n a l y s t felt he had to use some drastic action. As he explained: "we came to some areas.., where certain gentlemen reside and they are extremely difficult. They think they have to have everything done their own way... This went to and fro.., political battle, and in the end we said, "Right! Don't bother!, so we just chopped off the front end of the project.., and we said "Fine! When we fire this thing up, you do realize, gentlemen, that the only way you're going to order material is through this system so if you don't go through the system you won't get the thing because we ain't going to pay for it!' So they said, 'Oh. that's a good point'.., so we just walked off.., and waited for something to happen. They came back and said, 'Maybe we should'." Clearly, the user had little choice b u t to cooperate. Sometimes, the project involved m o v ing m a n a g e r s who resisted the new system: "In the managers it definitelygoes with age and how long people have been in the area...there you've got the.., when I say shop floor.., also some of the older managers. I get terrible resistance from them. One we've just had to move. In fact with every system that I've put • in, the managers had to move. They just block progress." This closely parallels M u m f o r d ' s earlier findings c o n c e r n i n g the designers' p o o r view o f the users: "The user cannot resist, he can only be reluctant." "People have nebulous ideas of what they want. They do not think through the details." "They don't understand the problems and think that any information they get from a manual system they can also get from a computer system." [34] a n d is confirmed by a later study o f the attitudes o f systems designers by H e d b e r g a n d M u m f o r d [22]. I n the stores area at A l p h a , a key one to the project, the senior m a n a g e r s were replaced: "The whole of the top management has changed. Three new, younger, dynamic guys, really keen to go. Tremendous! Because they really want to see the system go . . . . They're not going to sit around and listen to all this rubbish. One or two people have been moved sideways, 'promoted'.... I know (the changes) came from (one of the top managers)." A t the shop floor level he has also experienced resistance, a l t h o u g h n o t from everyone: "... The other area (of resistance) is on the shop floor. Those people we do get resistance from, (with) one or two exceptions. You get some brighteyed old fella who comes waltzing along and says, 'Oh that's terrific! Wish we had that in our day.' It's not very often. That's the exception."
402
Newman, Rosenberg--The Politics of Organizational Control
He gave an example of overcoming resistance to introducing a computer terminal depending on the person:
Interviewer: "Are they at the moment fully aware of what's going to happen and how the change is going to affect them?" Manager: " N o they're n o t . . . "
"You know you've got the young ones galloping up and saying, 'Oh! Can we have a terminal: let's go, I want to learn.' Good, that's terrific, but the stubborn o n e s - never. You've got to leave them in their own time and they'll come around. And once they realize what its really like they'll never admit it to you. They'll say, 'Oh well. This bloody old thing! I'll use it'. In fact they'll use it quite happily."
He then went on to explain that the top-down approach does not require a great understanding of the current system: it was going to be changed anyway. We then went on to the question of union involvement. He was not expecting any trouble. Did the unions know what was happening?
He also related a technique that one of his previous colleagues used to overcome resistance. The employees flatly refused to use the new stock system so the analyst went in one weekend, took all the stock cards and burned them. The employees then had to use the new system. Although such methods are drastic they probably form part of the on-the-job 'wisdom' that seems so prevalent in data processing work and demonstrates an enhanced claim to managerial authority. In the current example at Alpha, the analyst had the following 'practicalist' view of people which is also widespread in management. " . . . I have very little time for all this psychological stuff, as I call it . . . . (A person) is a leader or a follower. To me there are two categories that make s e n s e . . , that's the way everything has worked for 2 million years".
As an example of the implications of the rational, top-down approach to design we present without comment an extended extract from the conversation we had with the project manager. The new system was eventually going to cause some disruption at the clerical level. We were interested to see if any contingencies had been planned, especially as this division had experienced union problems before. Interviewer: "What about at the shop floor level? I notice there are manual files kept up here, rotary type of files. Will all that disappear?" Manager: "Yes." Interviewer: "And how are you going to handle that kind of changeover?". Manager: "The rotary files you speak of is a work order file of materials for plan requirements. That will all be contained in the computer . . . . " Interviewer: "What about changing the role in all this? Presumably they'll change from being a clerk to, say somebody that has to use a keyboard. Are there any problems associated with that that you can envisage?" Manager: "I don't know. Do you mean industrial relations problems or skill problems?" Interviewer: "Yes." Manager: " . . . To me t h i s . . . (is) just another tool. Using the tube (screen) is very simple. And this thing is just another tool. The information systems give them the same capabilities of a pencil. The transaction enables them to add, delete, or change information within it."
Manager: "As a courtesy we have told them we are doing this thing because naturally the people are going to ask questions. If the union office isn't aware of some major undertaking that's going to affect the employees, they're going to be put off. On the other hand, the union's role is to ensure good working conditions and good wages for their people. That's got nothing to do with what we're doing. This is a management fight . . . . Really, what we're doing is helping the union because what we're doing is going to increase the productivity of the p e o p l e . . . " Interviewer: "You haven't got to the stage where you are facing these problems?" Manager: " N o and we won't. I can anticipate a whole series of problems; I'll take them on with relish!"
One of the major direct sources of dysfunctional behaviour is the increase in these task redefinitions caused by new information systems. The literature seems to indicate that some of these changes are unintentional and are the result of the demand made by formal information systems in organizations and this was confirmed in the above case. In Bjorn-Anderson and Hedberg's study they gave an example where the design team realized the consequences of their design: "I realized that most of the users would turn into part-time key punchers . . . . The tellers' role will be changed. They'll need to know more about banking, especially about the work of the savings department." [5, p. 129]
Crompton and Reid [9] have documented how the deskilling of clerical work has in a local authority produced an 'antagonistic' attitude to information systems. For management, traditionally used to a comparative freedom and flexibility in their decision-making, the new technology could be massively disruptive, imposing a rationality constraint that few have been used to. Argyris [2] notes: "MIS can do to middle management exactly what job specialization does to lower level employees."
Argyris, in his article on why middle managers resit 'rational' information systems, gives us the following insight:
Omega, Vol. 13, No. 5 "The basic problem is that organizations are full of concealed dysfunctional actions and defences that are revealed by MIS."
403
groups is thus weakened by the imposition of MIS.
[21 In other words, much of the internal organizational 'rules of the game' may be functional to consensus construction among various actors and sub units but is not abstractly rational. The internal political system and its supportive professional and occupational cultures do not have overarching goals but they collectively constrain those actors who believe the organization does possess goals at different levels of operating. If the politics of senior management continue to demand more and better information systems, this demand for increased 'rationality' will reveal: " . . . a maze of cover-ups, elaborate fictions, and dist r u s t . . . v a l i d data for an MIS would reveal to many managements how much has been hidden from it all these years." [2]
It is a banal but accurate observation that for many important organizational actors the dysfunctional effect of such new systems is too high a price to pay for the supposed benefits? But this could continue, even if strains emerge, if the 'dominant coalition' demands increased formal rationality. MIS undercuts the existing 'rules of the game' as these are tied to the variety of weak or strong political systems in organizations which unify the various layers of management. Such a supposedly neutral act as one manager supplying data from his division to a common database (the corporate resource argument for databases [19]) finds, unless controls are explicitly built in, that data become the property of all units in the organization, increasing the potential of senior managers to control middle level managers especially. MIS as part of an overall attempt to overcome 'departmentalism' and 'particularism' through an insistence on the possibility of a general corporate strategy, threatens the power of subordinate managerial units to claim influence over the decisions formally taken by top management. The element of consensus negotiations between all power
3Not long after the interviews at the firm where the large scale materials management system was being implemented, the senior analyst left and the project was severely curtailed. The negative reactions to the topdown system were toe great to continue the project in its entirety.
C O N C L U S I O N S A N D DISCUSSION "... power has, in fact, passed to what anyone in search of novelty might be justified in calling a new factor of production. This is the association of men of diverse technical knowledge, experience, or other talent which modern industrial technology and planning require." John Kenneth Galbraith The relative deskilling of the middle management strata has been noted as a theme in the literature. The centralization of computer systems clearly has implications for the conditions of work of white collar supervisory workers but it is misleading to attribute these effects entirely to the machines themselves. It is the way the machines are allied to completely new ways of organizing human labour that will have the most significant impact on the work situation. Indeed, in the literature designed to help management with the introduction of computers it is constantly emphasized that mechanisation is insufficient. Machines are only efficient to the extent that they are incorporated into a reorganised labour process. The resistance to this reorganization is recognized but muted in the human relations literature of which Mumford and others are notable exponents. To a significant extent, the increased control of white collar and supervisory staff through information systems is seen as an improved mode of co-ordination of the enterprise and while they are often regarded with some hostility by middle management, top management can, and does, benefit from this service. Productionism and technocracy are both elements of Taylorism which older managerial groups may use on skilled and semi-skilled labour but would not welcome if it cuts into management as an 'art'. Attempts to control and limit control strategies upsets the informal mechanism and agreements between the lower level actors and supervisory levels in the front line. As we have found and as is confirmed by the literature, the values of system designers tend to be technologically oriented and the objective of rationality is highly prized. The designers tend to be younger than the managers they work with and, given the opportunity, they will seek new 'converts' to their way of thinking. Their goal is the design and implementation of 'good' systems and knowing that what they do is an improve-
404
Newman, Rosenberg--The Politics of Organizational Control
ment on the old technology. Despite all the organizations carried out in the wake of Leavitt friction they create or endure [41] they are and Whistler's prediction have been somewhat generally a highly satisfied group of specialists inconclusive. It is interesting to note, however, [34]. The only thing they have to fear is the that those studies carried out in financial orgrowth of managerial coalitions to block and ganizations, banking, insurance and the accontrol their work autonomy. Unfortunately, counting departments of large companies, have when it comes to dealing with these young tended to report an increase in the centralisation innovators, the same degree of satisfaction is of control [45]. In such a structural context the not always enjoyed by top management [13]. alliances between dominant coalitions of senior Like a number of new specialist occupational managers and systems analysts to establish and groups the designers are scientific 'elitists' who increase the span of control will not only affect can have a professional contempt for managers clerical staff but middle management as well. This will decisively affect the 'rules of the who have no scientific training or culture and possess managerial authority by claims to game' of the internal political system of or'experience' alone. As with many science spe- ganizations and push or lure specific groups of cialists the secular claim to both autonomy from employees into a diversity of patterns of collabmanagerial control and high status is implicitly oration or resistance. Neither labelling such legitimated by a quasi-religious belief that they patterns of resistance as 'irrational ludditism' or alone hold the key to unlocking the answers. believing that users can be manipulated into Most practitioners would readily admit that acceptance if in real terms the span of control is 'participation' can produce better designs. But affected, are adequate solutions and this has such solutions have been known and advocated been noted by Markus [31]. If we accept that the traditional techniques for over a decade, so the question arises, why is there no indication in the literature that the for implementation are inadequate because they ignore the issue of power which is at the heart situation is getting any better [5, 14, 16, 31]? The normative solution is orthodox and of the process [31, 39] then the implications are reflects a strong commitment to an abstract extensive. While top management may be attracted by learning through education. Organizations, it is suggested, should recruit social scientists and the potential benefits of sophisticated informatrain them as designers. The emphasis is clearly tion systems they need to be wary of the conon changing the designer. Unfortunately, this sequential strains and stresses produced in the approach ignores some of the organizational political system. Moreover, there is a tendency for new systems to disturb traditional patterns realities in which the designer operates. The computer's potential as a mechanism of of behaviour in unexpected ways. As with any storing, monitoring, processing and making major reallocation of internal resources--in this available information diminishes not merely the case not money but information and control value of white collar particularized knowledge over information--the disruption to orof the work process and his or her ability to ganizational life may be too high a price to pay exercise discretion in making decisions and for the benefits. The effect on middle management, while not in controlling the performance of work as Braverman [7] argues, but has strong effects on homogeneous, is likely to involve some loss of management. Clerical work is subject to a de- control. Middle management would do well to skilling and middle management loses understand some of the patterns of change significant autonomy and control. The advent required of them by the new systems. Whilst it of the computer then has served not merely is true that most managers are aware of the to deskill the clerical 'craft' but also to re- politicised nature of work and are adept at centralize aspects of control which were rela- securing corporate resources, the implications tively diffused throughout the organization. of major changes in information systems are Leavitt and Whistler [29] predicted that com- often inadequately grasped. When they are unputers would centralise power within the or- derstood, managers have sometimes reacted ganization and would severely attenuate the with blind resistance. Where small scale changes middle levels of managerial authorities. Empir- in information systems are made in individual ical studies of the impact of the computer on the departments the loss of autonomy can be min-
Omega, 1Iol. 13, No. 5
405
ment information system at the US post office. Minimised. However, in the case of large, top-down, neapolis, Minnesota. Management Information System information systems which cross departmental Research Centre, Working Paper 72-05. boundaries, while the scope for individual re2. Argyris C (1970) Resistance to rational management systems. Innovation 10, 28-35. sistance is muted, the potential for collective 3. Bariff L and Galbraith JR (1978) Intraorganizational action is greatly enhanced. A concerted coalipower considerations for designing information systion of middle managers could block the develtems. Acctng, Orgns Soc. 3(1), 15-27. opment process or at least influence the design 4. Bennis W (1965) Theory and method in applying behavioral science to planned organizational change. J. Apin more favourable ways [31]. plied Behavioral Sc. 1(10), 338-60. The effects of systems innovation upon cler5. Bjorn-Anderson N and Hedberg B (1977) Designing information systems in organizational perspective. In ical work practices is often to deskill staff if not Prescriptive Models of Organizations (Edited by Nysto eliminate jobs altogether. Some of these trum P and Starbuck W) North HoUand/TIMS Studies effects are an inevitable by-product of the new Management Sciences, Amsterdam. technology while others occur by default [5]. 6. Bostrum P and Stephen Heinen J (1977) MIS problems and failures: A socio-technical perspective. MIS Q. 1(3), Trade Union officers could exert more influence 17-32. on the design process and minimise its effects on 7. Braverman H (1974) Labor and Monopoly Capital. Monthly Review Press, New York. clerical staff by insisting on representation on the design team. Additionally, clerical staff 8. Burns T and Stalker GM (1961) The Management of Innovation. Tavistock, London. themselves can exercise more influence than is 9. Crompton R and Reid S (1983) The deskilling of clerical work. In The Degradation of Work? (Edited by Wood often realised [33]. Hutchinson, London. The recognition of a broader approach to 10. S) Crowe E (1981) The loneliness of the long distance implementing information systems would also programmer. Radical Sci. J. 8, 137-141. have an impact on the academic subject as 11. Crozier M (1964) The Bureaucratic Phenomenon. TavLondon. taught in further education and professional 12. istock, Dalton M (1959) Men who Manage. Wiley, New York. training establishments. The introduction of 13. Danziger JN (1977) Computers and the frustrated chief executive. M1S Q. 1(2), 43-53. case study material to supplement more tradiBrabander B and Edstrom AN (1977) Successful tional approaches may help future analysts in 14. De information systems development projects. Mgmt Sci. their appreciation of the organizational com24(2), 191-199. plexities they will have to operate within but it 15. Dickson GW and Simmons JK (1970) The behavioral of MIS. Bus. Horizons 13(4), 59-71. must be admitted that this is largely tinkering at 16. side Dutton WH and Kraemer KL (1978) Management the margin. utilization of computers in American local governments. Comm. A.C.M. 21(3), 206-218. It is likely that the significant differences inside organizational political systems will affect 17. Eason K (1977) Case studies on the impact of computer based information systems of management. Loughthe patterns of co-operation and conflict. A borough University, Loughborough. need for more field work based on studies of 18. Emery FE and Trist FL (1960) Social-technical systems. In Management Science: Models and Techniques (Edited systems analysts and how they relate to the by West Churchman C and Verhulst M). Pergamon internal 'rules of the game' will illuminate this Press, Oxford. little known area of managerial culture and 19. Everest GC (1974) Database management systems tutorial. In Fifth Annual Midwest AIDS Conference Prodecision making.
ACKNOWLEDGEMENTS
20.
The helpful and constructive comments supplied by the anonymous referees are gratefully acknowledged. After this paper was written and before it was accepted, one of its authors, David Rosenberg, died suddenly at the age of 41. In his academic career he was acknowledged as a thorough scholar by all his colleagues. David published extensively in the area of local government accounting, successfully combining his knowledge of sociology, accounting and political science to form a true interdisciplinary approach. He will be greatly missed by those who knew him.
21. 22.
23.
24.
REFERENCES 1. Anderson JC, Dickson GW and Simmons J (1973) Behavioral reactions to the introduction of a manage-
25. 26.
ceedings (Edited by Chervany NL) Vol. 1, Minneapolis, Minnesota. Greenbaum J (1976-77) Division of labour in the computer field. Monthly Rev. 28(3), 40-55. Gouldner A (1957-58) Cosmopolitans and locals: toward an analysis of latent social roles, 1 and 2. Admve Sci. Q. 2, 281-306 and 444 480. Hedberg B and Mumford E (1975) The design of computer systems. In Human Choice and Computers (Edited by Mumford E and Sackman M), pp. 31-59. North-Holland, Amsterdam. Hedley RA (1970) Organizational objectives and managerial controls: A study of computerisation. In Industrial Organization--Behaviour and Control (Edited by Woodward J). Oxford University Press, London. Hill S (1981) Competition and Control at Work. Heinemann, London. Hughes EC (1971) The Sociological Eye: Selected Papers, Aldine-Atherton, Chicago. Hickson D J, Hinnings CR, Lee CA, Schneck RE and
406
27. 28.
29. 30. 31. 32. 33. 34. 35. 36.
Newman, Rosenberg--The Politics of Organizational Control
Pennings JM (1971) A strategic contingencies theory of intra-organizational power. Admve Sci. Q. 16(2), 216-229. Knight K (1967) A descriptive model of the intra-firm innovation process. J. Bus. 40(4), 478-496. Kraft P (1979) The industrialization of computer programming: From programming to 'software production' 2. In Case Studies on the Labor Process (Edited by Zimvalist A). Monthly Review Press, New York. Leavitt H and Whistler J (1958) Management in the 1980s. Harv. Bus. Rev. 36(6), 41-48. Lucas HC Jr. (1975) Why Information Systems Fail. Columbia University Press, New York. Markus ML (1983) Power, politics and MIS implementation. Comm. A.C.M. 26(6), 430-444. McKinsey Company Inc. (1969) Unlocking the computer's profit potential. Comput Automn 18, 24-33. Mechanic D (1962) Sources of power of lower participants in complex organisations. Admve Sci. Q. 7(3), 349-364. Mumford E (1972) Job Satisfaction: A Study of Computer Specialists. Longmans, London. Myers CA (1966) (Ed.) The Impact of Computers on Management. M.I.T. Press, Cambridge, Massachusetts. Newman M (1985) Access to information: Strategies for prevention and promotion. J. Mgmt Stud. 22(2), 61-79.
37. Pettigrew A (t973) The Politics of Organizational Decision Making. Tavistock Publications, London. 38. Pettigrew AM (1975) Strategic aspects of the management of specialist activity Personal Rev. 4(1), 5-13. 39. Pfeffer J (1981) Power in Organisations. Pitman, Marshfield, Massachusetts. 40. Radnor M, Rubenstein AH and Bean AS (1968) Integration and utilization of management science activities in organizations. Ops Res. Q. 19(2), 117-141. 41. Rose M (1969) Computers, Managers and Society. Penguin, Harmondsworth, Middlesex. 42. Ross JE (1968) Data processing training for management: Objectives and curriculum content. Comput. Automn 17(9), 16-20. 43. Salancik G and Pfeffer J (1977) Who gets power--and how they hold on to it: A strategy---contingency model of power. Organl Dynamics 5(3), 3-21. 44. Swanson EB (1974) M.I.S.: Appreciation and involvement. Mgmt Sci. 21(2), 178-188. 45. Whistler T (1970) The Impact of Computers on Organisations. Praeger, New York. Newman, Faculty of Economic and Social Studies, University of Manchester, Manchester M I3 9PL, UK.
ADDRESS FOR CORRESPONDENCE: Dr M