Domains and cultures: the context of communication

Professor Tom Wilson
Department of Information Studies
University of Sheffield

The heading that. appears in the programme was intended to remind me what was required of this presentation; what I am actually going to do is to try to explore why it is so difficult to get research and information messages across in organizations.

We take it for granted, I think, that research and information activities do have some contribution to make to the effectiveness of organizations. It is very difficult, however to prove the link or even discover it. When times are good there is little incentive to do the necessary research, simply because the 'goodness' of the activities is taken for granted. And when times are hard, there's no money around to do the work, and everyone's busy simply trying to make that impact.

Before we can go further, however, we need to be clearer about the terms 'research' and 'information' - partly because there is some confusion over the words) and partly because words are important. I think that at least three activities are embraced by the two words:

1. internal monitoring of the organization's activities, requiring data analysis techniques, and, increasingly, the use of computers. In this activity we are trying to discover how the system functions;

2. intelligence gathering the external equivalent to monitoring scanning the environment for information on social and other trends that may affect policy developments, or management practice, or professional practice; and for the findings of external research that may have something to say for policy, management, or practice. In this activity we are seeing how the system relates to the environment; and

3. research discovering why the system -functions as it does and/or, what effect the system has with regard to the objectives set, and/or how it responds or fails to respond to environmental changes.

There is a tendency to confuse monitoring with research and those who direct organizations must come to terms with the need for all three of these functions to be performed and, preferably, be made explicit in the organizational structure, if the system is to operate effectively and adapt to a constantly changing environment. Organizations that. fail in one or other of these functions are likely, in my opinion (and with some empirical support) to be less effective than they might otherwise be.

And, of course, the three things are interrelated: intelligence gathering draws attention to needs for changes in monitoring and poses questions for research; monitoring draws attention to patterns or inconsistencies that. need explanation or comparison with trends and events elsewhere; research draws attention to the need for new intelligence and for different actions to be monitored, or monitored and analyzed in different ways.

But what is this thing that uses research results, intelligence and data this thing we call the 'organization' or the 'system'? Leaving aside a detailed examination of the debate, I have found three sets of ideas useful and stimulating in thinking about the role of research and intelligence. We need models of the kind I shall describe because the interaction between s research and intelligence (R & 1) section and the rest of the organization is complex and potentially fraught with opportunities for misunderstanding.

The models are those of Charles Handy (1978, 1985), Kouzes & Mico (1979), and Henry Mintzberg (1979). Handy develops the idea of organizational culture briefly, the 'boss' culture of the autocratic head, supervising everything and everyone; the role culture of the typical bureaucracy; the problem solving or 'task' culture embodied in a matrix organization; and the 'existentialist ', Dionysian, or individualistic culture, typified by the law partnership or the university.

All organizations, of course, are a mix of cultures: it is not too difficult to imagine a Probation Service which functions chiefly as a professional bureaucracy but which has, perhaps, the occasional Area Officer with autocratic, Zeus-like tendencies (such characters are even found in universities), desperately trying to manage a bunch of Dionysian probation officers, bent on doing their own thing for the ultimate good of the client, of course.

Naturally, the different cultures have different attitudes towards information and communication: the 'boss' is all-knowing; the role culture has information flowing up and down the lines of authority (chiefly down it seems to most of use); the problem solvers share information continually; the Dionysians talk to one another whenever they feel the need.

And where is the R & I section in this? Well, by inclination and function I think they are likely to see themselves as problem solvers, wondering how they got into the bureaucracy and what on earth they are doing there, and who these crazy people are all around them. How they can use the idea of organizational cultures to make sense of things is something I shall return to.

First, however, let us take a look at the idea of domains. Although Kouzes and Mico do not acknowledge it, the basis of the idea has a long history in the sociology of organizations. It is certainly found in one of the more approachable works of Talcott Parsons (1960) and, in essence, is the fairly trivial idea that human service organizations have levels or domains: the policy domain, the management domain, and the service domain. What Kouzes and Mico do with the idea, however, is to draw attention to the fact that the different domains involve different modes of work, and different measures of success, and, I would add, different demands for information as a. consequence.

Again, where is the R & I section in this? Nowhere, seems to be the answer. Domain theory is too simplistic, at least in its attention to structure. It ignores two other divisions which Mintzberg draws attention to; namely, the 'technostructure' which houses analysts who are concerned with designing the flow of work, setting standards of performance, and the like not particularly prominent in what Mintzberg calls the 'professional bureaucracy'; and the 'support staff' - very strong in a professional bureaucracy - which includes administrative support and, of course, R & D.

Clearly, the R & I section is a unit which serves the organization without being integrated into the line management structure. So which domain is it supposed to serve? All? Is that possible? How would it do that? For the policy domain it would need to produce strategy papers on long-term trends (or obtain such intelligence from outside) and its usefulness would be judged on the basis of how well it assisted policy formation. For the managerial domain it would need to monitor activity and draw attention to significant divergences in that activity which might need remedial or preventative managerial action. It would also need to gather environmental intelligence to help management understand the context within which things were happening. Its success would be judged on criteria of efficiency in performing these tasks.

And for the service domain? For those Dionysian probation officers? Well, only one thing really research and intelligence that will help me do my job but perhaps I exaggerate.

Organization theory is often castigated as having little to do with the real problems of managing organisations but I believe that the sets of ideas I have examined have great value for people trying to get to grips with the complexity of organizations and discovering how to do the things required of them in an effective manner. Not only is it difficult for research and information sections in all. kinds of organisations to discover how to function, it is also difficult for their managers to decide how they ought to.

Consider, for example, how management might use the ideas. I'm sure that managers here have already identified more messages than I am capable of suggesting, but I see at least four in what I've been saying.

First, to go back to the beginning: which of the three tasks I identified do you want R 8r I sections to do? All of them? Is that possible? Are the sections properly staffed to do all three jobs? What central task do you see them performing and have you identified the centrality of that task in job descriptions and section objectives? Does your R & I section know, in other words, what it is supposed to be doing? Lots of questions, but sometimes questions can be more useful than answers.

In particular, I would suggest that if management want monitoring and not research, that point should be made clear in job titles, job descriptions, and the names of the operational units. If you call a monitoring unit a research unit you will get people who want to do social research in the Probation Service and who will be very disgruntled if all they are expected to do is produce statistics. True, 'Research Officer' flows more readily off the tongue than 'Monitoring Officer' and having a 'Research and Intelligence' unit may seem more prestigious than having a 'Monitoring' unit but is it really what you want?

Secondly, does the concept of organizational cultures make some kind of common-sense to you? If so, you should be sensitive to the need for R & : I staff to have the flexibility to approach the different cultures in different ways and that takes time. There is no point in expecting that the same, undifferentiated mode of dissemination of, for example, research results, will get the message across to the different cultures.

Thirdly, does the idea of domains make sense? If it does then certain messages flow from the ideas. If activity is being monitored, which domain is the information intended for? Is it properly designed for people working in that domain, and does the research/monitoring section know which domain they are supposed to be addressing? If research is commissioned, are the results intended primarily to influence policy, or management practice, or operational practice? Is the research brief defined in those terms? Do the operational level staff from whom data are collected know that the work is intended to lead to change in their ways of working or not, as the case may be? Are unrealistic expectations being set up because these matter's have not been properly addressed? If an intelligence or information function is being performed who is it for? Everyone across the domains? Through the same media? If so, is it surprising if large numbers of staff never use the information?

Fourthly, does Mintzberg's idea of a five element structure for organizations, which I presented as an elaboration of the idea of domains, make sense? Again, if it does, messages follow. The R & I section is a 'support' section, on the edge of the main professional bureaucracy some might even say 'marginal' to the operations of that bureaucracy. However, if the R & I section is truly engaged in research, rather than in monitoring, it will need high level support if it is to have the appropriate kinds of access to the relevant people and research arenas. Does the existing structure provide that? You may recall that the other element of structure proposed by Mintzberg is the 'technostructure' the standards-setting part of the system. Mintzberg puts personnel training units in the technostructure as being concerned with performance standards - in that case training is divorced from R & I and I think there are compelling reasons for linking them, as I shall suggest in a moment.

Now let us look at the messages for the R & I section. First, the notion of cultures. As soon as it is recognized that different sections of an organization might embody different cultures in this way, it becomes obvious, I think, that the section must use different strategies for getting its messages across. Assume that the service domain in the Probation Service is predominantly Dionysian, then, to make sure that everyone knows what you believe he or she needs to know you have to tell each person individually. The normal bureaucratic lines of communication do not function very well in that culture.

Work in the USA on the social worker's understanding of research emphasises the nature of the communication problem at that level. Speaking of research studies shown to social workers the paper notes:

'...the findings of the study exerted a strong biasing effect on social workers' judgements of the research. When reading a study with positive findings of the effectiveness of casework treatment, respondents evaluated the study as methodologically superior, and generally as having greater implications for practice than the same study reporting negative results. This biasing effect was so great that social workers reading the study with positive results tended to believe that the treatment procedures described were similar to their own, while those confronted with negative findings tended to deny that this was so!' (Kirk & ; Fischer, 1976)

In my opinion research results with implications for practice can only be put. across effectively through training or through organization development. And, as that quotation suggests, that is all the more so if the findings of research are negative with respect to current practice. I believe strongly that where in-house research is concerned, there is a very great need for a formal link between research and training, otherwise the benefits of research will not be found.

Work that Marian Barnes and I undertook last year in social services, departments suggests that this lesson has not yet been learnt in that sector: 50% of the projects reported to us were intended to have an impact on practice, but training techniques were used for dissemination in only 4% of the cases and OD in a further 14%. (It also seems likely that the 4% was a subset of the 14%). The typical final report on the project (which I believe to be confused with the academic function of research) was produced in 87% of the cases (Barnes & . Wilson, 1986).

We can make similar statements about the other cultures identified by Handy; for example, let us assume that the service domain consists of a set of problem solving groups rather than of idiosyncratic individuals. Communication is so free in such groups that you can tell anyone arid the word will get round to everybody pretty quickly. The role culture is something of a problem. It ought to function in such a way that information fed to the top will filter down the bureaucracy to everyone (whether or not everyone actually needs to know is another matter of course) but bureaucracies do not function perfectly. Things go wrong. People can act as gatekeepers in two ways: they can keep the gate open, allowing information to flow through the system, or they can keep in closed. One needs to know the system in some detail to discover where the blockages are, and how to circumvent them. If you a working to a Zeus figure, of course, you had better make sure you tell him (or her) first.

The implications of domain theory and of Mintzberg's ideas on the structure of organizations for the R & I section mirror those for management, so I can be brief. The chief message of domain theory, I believe, is similar to that for cultures: the effective presentation of research results and the recommendations flowing from them, or of information and intelligence generally, demands different methods for the different domains. For two reasons: first, the domains operate according to different principles - equity at the policy level, service at the service level, and secondly, they have different work methods and standards of performance - standards of efficiency at the managerial level, professional autonomy at the service level. In other words, we are not dealing with a single organization, we are dealing with mini-organizations within a maxi-organization. And we need to understand them and pattern our modes of dissemination appropriately: final reports where they will be read, oral presentations where they are signalled, policy papers and discussion groups where they are needed, executive summaries where they are needed, and so on.

Can you do all of this? Of course not - not so long as the function of R & I is not properly defined. And, even when it is properly defined, would the staffing implications be properly taken into account? But at least analyzing the organization in this way enables you to say what you believe ought to be done, and you may have the opportunity to demonstrate that benefits follow if you do function in this more 'user-sensitive' fashion.

As for Mintzberg's ideas, well, again, it is a mirror image of the message to management. What kind of a function is expected of R & I? Is it a management support unit, or is it intended to contribute to organizational change and to the development of effective work practices? If it is the latter then its place is with training, and one can argue that research, rather than monitoring, ought to be concerned with performance standards and the identification of best practice.

Now. of course, organizations change, or are made to change (just as the Probation Service in various areas has had to adapt to the disappearance of the Metropolitan Counties) and the models that may be helpful in describing and understanding today's organizations may not be very useful for the future, or for pointing the way forward to change. Handy says as much in the first edition (1978) of his book, Gods of Management when he gives it the subtitle, 'How they work and why they will fail' (my added emphasis). The subtitle is changed in the second edition (1985) to, 'The changing work of organizations') but the message in the book is essentially the same:

'...the organizational imperatives (for increased size and greater consistency) in our present society are locked into an inevitable battle with the individual imperatives (for greater opportunity for personal expression and choice)... If our organizations are to survive, they must adapt their managerial philosophy to one which is better suited to the needs) aspirations and attitudes of individuals. In the new mix of gods which will result, Apollo will be less dominant and less inhuman. This will mean, however, a massive reorganization of the structure of organizations and of their work.' (Handy, 1985: 177)

Quite coincidentally, I was talking about matters of this hind at. a meeting at Birmingham University yesterday, and an the train back to Sheffield I got into a discussion with a social services researcher about the kinds of structures that might emerge if decision making responsibilities were devolved to the area level. We concluded that the areas would need to be bigger, to reduce the possibilities of diverse service standards being applied in different areas, and that some structure would be needed at the policy level to ensure consistency in priorities and standards. We hypothesised that the overall structure of the system would then be a matrix structure, with people representing both an area and a professional specialism. The policy board would represent the areas and the specialist teams and, in all probability, the middle rnanagement would disappear. The R & I unit, however, or something very like it, possibly allied to training would survive as a service unit to the policy-making board.

This is not a forecast for the future of human service organizations (because, of course, the first rule of forecasting is that forecasts are always wrong), it is simply to suggest that organizations can be designed and that in the future the Probation Service and its R & I sections may be faced with the need to redesign and that it might be a good idea to start now to analyse existing structures and ensure that functions are performed to the benefit of everyone: policy maker or probation officer, Zeus, Apollo, or Dionysius.

Finally, it is worth noting that although there is nothing as useful as a good theory, the application of theoretical ideas sometimes leaves a little to be desired. I came across a limerick while preparing this paper which I think sums up the situation nicely;

We all place a great deal of reliance
On the theory and practice of science
But the hopeful intentions
Of so many inventions
Can be quite buggered up in appliance


  • Barnes, M. & Wilson., T.D. (1986) The internal dissemination and impact of in-house research in social services departments. Research, policy and planning, 4, 19-24.
  • Handy, C. (1978) Gods of management. London: Souvenir Press.
  • Handy, C. (1985) Gods of management. 2nd ed. London: Pan.
  • Kirk, S.A. & Fischer, J. (1976) Do social workers understand research? Journal of Education for Social Work, 12, 63-70
  • Kouzes, J.M. & Mica, P.R. (1979) Domain theory: an introduction to organizational behavior in human service organizations. Journal of Applied Behavioral Science. 19, 449-469
  • Mintzberg, H. (1979) The structuring of organizations: a synthesis of the research. Englewood Cliffs, NJ: PrenticeHall.
  • Parrott, E.O., ed. (1984) The Penguin book of limericks. Harrnondsworth: Penguin.
  • Parsons, T. (1960) Structure and process in modern societies. New York, NY: Free Press.

This paper was originally given at a national conference of probation research and information staff in 1987