art of science

24
8 The analytical interview Relevance beyond reflexivity Kristian Kreiner Jan Mouritsen Clearly, in the general area of management and business studies, the am- bition to describe and analyse the details of practice is a legitimate one (Jönsson, 1996, 1998). This ambition has focussed on the multiple ways in which it is possible to make theoretical sense of managers’ realities (Jönsson & Macintosh, 1997). Often, this involves close collaboration between researcher and the empirical field in various ways where the re- searcher often is an interviewer, a discussion partner, an observer or a change agent. In this essay we propose and discuss an interviewing practice that we call ‘the analytical interview’. Interviewing is probably the most prevalent method for collecting data in the social sciences and arguably a best prac- tice already exists (see, e.g., Gubrium & Holstein, 2002). We do not in- tend to rehearse the content of such best practice; nor will we partake in the healthy deconstruction of the ways in which interviewing is represent- ed by scholars and practitioners (e.g., Holstein & Gubrium, 1995; Czar- niawska, 2001; Alvesson, 2003). Our aim is narrower in the sense that we focus on one particular type of interview practice (or interviewing strategy) that emphasizes collaborative analysis and construction of knowledge between an interviewer and a respondent. How is ‘analytical interviewing’ analytical and therefore distinct from other types of interviewing practice? Conventionally (according to, e.g., the Oxford English Dictionary), ‘analysis’ signifies the process of

Upload: asser-humblefella

Post on 29-May-2017

237 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Art of Science

8The analytical

interviewRelevance beyond reflexivity

Kristian KreinerJan Mouritsen

Clearly, in the general area of management and business studies, the am-bition to describe and analyse the details of practice is a legitimate one (Jönsson, 1996, 1998). This ambition has focussed on the multiple ways in which it is possible to make theoretical sense of managers’ realities (Jönsson & Macintosh, 1997). Often, this involves close collaboration between researcher and the empirical field in various ways where the re-searcher often is an interviewer, a discussion partner, an observer or a change agent.

In this essay we propose and discuss an interviewing practice that we call ‘the analytical interview’. Interviewing is probably the most prevalent method for collecting data in the social sciences and arguably a best prac-

tice already exists (see, e.g., Gubrium & Holstein, 2002). We do not in-tend to rehearse the content of such best practice; nor will we partake in the healthy deconstruction of the ways in which interviewing is represent-ed by scholars and practitioners (e.g., Holstein & Gubrium, 1995; Czar-niawska, 2001; Alvesson, 2003). Our aim is narrower in the sense that we focus on one particular type of interview practice (or interviewing strategy) that emphasizes collaborative analysis and construction of knowledge between an interviewer and a respondent.

How is ‘analytical interviewing’ analytical and therefore distinct from other types of interviewing practice? Conventionally (according to, e.g., the Oxford English Dictionary), ‘analysis’ signifies the process of

Art_of_Science2f.indd Avs7:153Art_of_Science2f.indd Avs7:153 05-11-26 14.08.4105-11-26 14.08.41

mw.lib
Pages 153-176 in: Tengblad, Solli and Czarniawska (eds.): The Art of Science, Liber/CBS Press, 2005. Published in Insights@CBS, 2006, nr. 18: http://frontpage.cbs.dk/insights/710008.shtml
Page 2: Art of Science

breaking down a phenomenon into its constituent parts; the phenomenon can then be understood as the synthesis of these parts. It also refers to the process of discovering the implicit and hidden principles that govern em-pirical phenomena. We use ‘analytical’ to signify a commitment to study dilemmas in ordinary organizational life and practice. In the British phi-losopher Gilbert Ryle’s (1954/2002:1) sense, dilemmas are characterized by:

lines of thought, which are not rival solutions of the same problem, but rather solutions or would-be solutions of different problems, and which, none the less, seem to be irreconcilable with one another.

Such different problems, or such conflicting and competing concerns, are the constituent parts of practice. Being explicitly or implicitly suspended between such concerns, practice develops in ways that become the focus of empirical observation and theory building.

Empirical work and theory building are not only the aims but also the content of analytical interviewing. Theory building involves the “sugges-tion of relationships and connections that had previously not been suspected, relationships that change actions and perspectives” (Weick, 1989). The interviewer engages the interviewee in the construction of new dilemmas and the exploration of their potential implications for cur-rent knowledge about current practice. In this sense, analytical inter-viewing is a collaborative effort (Ellis & Berger, 2002), a sort of conver-sation (Holstein & Gubrium, 1995), in which the classical distribution of interviewing roles is radically changed. Constructing and exploring di-lemmas, i.e., making separate concerns and problems conflict and com-pete for attention, is a strategy for the rendering of practice as problem-atic, and thus worthy of study. Presumably, practice will often appear as unproblematic, to actors as well as to socialized spectators. To consider some thing as natural or as rationally chosen is probably in many cases an acquired ability. By focusing on dilemmas we are reminded that the liti-gation between the conflicting and competing concerns might, under slightly different circumstances, have produced another practice. The inter est in the observed practice must build on the assumption that the practice could have been different. By exploring dilemmas the interviewer and the interviewee are able to construct the counterfactual image of practice that makes the factual practice significant. It further allows them

154 t h e a n a l y t i c a l i n t e r v i e w

Art_of_Science2f.indd Avs7:154Art_of_Science2f.indd Avs7:154 05-11-26 14.08.4105-11-26 14.08.41

Page 3: Art of Science

to contemplate how practice may change in the future, i.e., the current practice represents nothing but a quasi-resolution of the conflict between competing concerns (Cyert & March, 1963/1992).

The analytical interview redefines respondents’ contribution. What they say can no longer be treated as privileged conjectures about current practice and its justification. The aim is not to provide final answers but to give input to the analytical conversation. Of course, respondents con-tinue to volunteer pre-existing knowledge; more significantly, they engage with the interviewer in the production of new knowledge. We present five reasons for not treating the interviewee’s answers as final and privileged knowledge:

• Ignorance: Organizations tend to develop routines that over time erase their rationale (March, Schulz & Zhou, 2002). What once may have been a conscious act of engineering a balance between conflicting concerns may over time become a behavioural pattern learned through socialization. The interviewee’s understanding and account may reflect socialization more than experience, and he or she may be ignorant of the tensions that the organization has learnt to cope with in a routine manner.

• Tacitness: Much of what actors know about their practice they know tacitly. Being asked to give a verbal account of a potentially complex and contradictory practice reduces the account to a sim-plified and one-dimensional picture.

• Boundedness: One possible way of handling conflicting concerns is to create loosely coupled organizations. The interviewee may not necessarily know the practice of other parts of the organization, and may never have considered the links between his or her own practice with that of organizationally distant colleagues.

• Institutionalization: The interview situation has multiple structural and cultural antecedents (Briggs, 2002; Alvesson, 2003). The interviewee’s account may tend to drift towards something that the interviewer is believed to understand, expect and/or accept. Furthermore, the account is governed by logics of representation, not logics of practice (Czarniawska, 2001).

• Opportunism: The interview process is not only a passive repre-sentation of a piece of practice, but is also potentially an opportu-nity for an active and opportunistic presentation in order to pro-

c r e a t i v e a n d s k i l f u l r e s e a r c h p r a c t i c e 155

Art_of_Science2f.indd Avs7:155Art_of_Science2f.indd Avs7:155 05-11-26 14.08.4105-11-26 14.08.41

Page 4: Art of Science

duce a particularly favourable image of the organization and/or the interviewee.

All these five conditions are reasons for serious concerns about the validi-ty of the information gained from the interview, were we to consider the interviewee as the source of final truths. However, since the analytical in-terview treats such information as input to the exploration of new know-ledge, the issue of validity is of less saliency. The conversation between the interviewer and the interviewee will explore the plausibility of state-ments, i.e., it will test the degree to which the interviewee’s conjecture is “interesting, obvious, connected, believable, beautiful, or real, in the con-text of the problem they are trying to solve” (Weick, 1989:524). Thus, while validity is never an irrelevant concern, the analytical interview holds the promise of producing new knowledge that overcomes at least some of the above-mentioned reasons for questioning the validity of the collected data.

Having now indicated what analytical interviewing is, in the rest of the essay we discuss the types of strategy that analytical interviewing may adopt. We describe and discuss an interviewing practice that in itself must balance various concerns and lines of thought. Thus, our presentation of the analytical interview as a practice consists of pointing out the dilem-mas that need to be handled in the interview process.

The strategy

As a research strategy the analytical interview develops a new set of con-cerns about the whole process of interviewing. Stated as a series of dicho-tomies, it is possible, as we will show in the following sections, to think of alternatives to the process and purpose of the analytical interview compared with a view (for the present purpose, narrow) of what the con-ventional interview accomplishes. It is important to stress, however, that we use the dichotomies to illustrate the research strategy of the analytical interview. The concerns reflected in conventional interviewing are not irrel-evant, but possibly less relevant in the case of analytical interviewing. We admit to over-emphasizing the one concern of the many implicit, conflict-ing and competing concerns in connection with analytical interviewing. In practice, the interviewer in an analytical interview must constantly try to build a bridge between the concerns that we pose as dilemmas below.

156 t h e a n a l y t i c a l i n t e r v i e w

Art_of_Science2f.indd Avs3:156Art_of_Science2f.indd Avs3:156 05-11-26 14.08.4105-11-26 14.08.41

Page 5: Art of Science

Next, we analyse four dilemmas implicit in the process of inter-viewing.

Plan versus presence

In interview-based research, as in most other endeavours, it is a good idea to plan – to think before you act. The existence of an interview guide is the symbol of such thinking or planning ahead of the actual interview. The existence of an interview guide helps negotiate access to the inter-viewee, and it helps the interviewer to guide the dialogue towards a use-ful end within the presumably narrow time frame of the interview. Un-guided, the interviewee may easily run off with the dialogue and leave the interviewer with idle talk and useless data. In short, it represents good inter viewing manners to plan in advance what you want to ask the inter-viewee, the kinds of problems and issues to be discussed, and the type of knowledge to be brought home from the interview.

But there is a different concern that tends to be neglected in the call for planning the interview. The interview guide makes the interviewer think before the interview and is a tool for governing the interview process, but in some situations the guide may easily stand in the way of finding new knowledge. It is constructed on the knowledge we already have before we talk to the interviewee; and it is constructed in such a way that the inter-viewee may likely know, and can easily provide, the answers to our ques-tions. If the interview guide governs the dialogue there is little room for learning new things, for seeing new connections, for pursuing new ideas, and for thinking original thoughts. As the aim of an analytical interview is to explore the uncharted complexities of practice (as described above), listening and recording will not suffice. Hearing (including the inter viewer’s hearing the implications and suggestions of the interviewee’s answers, and the interviewee’s hearing the implications and suggestions of the inter-viewer’s questions) will be necessary. That scenario requires a presence of mind that sometimes the interview guide makes difficult, if not impossible.1 In pursuing an analytical interview strategy, the interview process changes from one of data collection to one of knowledge construction. Such know-ledge construction requires thinking, not only as a prerequisite for asking the question, but also as an effect of hearing the answer.

1 Ryle (1949/1990) discussed the difference between, e.g., listening and hearing in terms of ‘task verbs’ and ‘achievement verbs’. See Ryle (1954/2002).

c r e a t i v e a n d s k i l f u l r e s e a r c h p r a c t i c e 157

Art_of_Science2f.indd Avs3:157Art_of_Science2f.indd Avs3:157 05-11-26 14.08.4205-11-26 14.08.42

Page 6: Art of Science

Analytical interviewing legitimizes constructive violations of good in-terview manners. When interviews fail it is rarely because the interview guide is violated, but because it is not violated! The meandering dialogue may be bad interview practice, but it may also be a rational strategy for exploring new meanings. One way to cope with the dilemma between preparation and presence is to consider and treat the interview guide (the symbol of preparation) as an input to, and a premise for, the knowledge creation and the sense making during the interview – as opposed to treat-ing the guide as a driver of the interaction in, and a template for the out-come of, the interview. The extent to which such a coping strategy suc-ceeds can be measured by the frequency and quality of ‘follow-up’ ques-tions.

Interacts versus double interacts

The focus of the interview guide is on the interview questions, and its ambition is to provide an interesting and stimulating beginning to the di-alogue. Obviously, the rationale for interviewing in the first place is the answers solicited by the opening questions. It is the answers that are re-ported as data and that are discussed in the methodological literature in terms of validity and reliability. In a sense, the ‘final answer’, one that completes and terminates the dialogue, is the ideal. It fully answers the interviewer’s query. The interviewer can then jump to the next question in the interview guide.

However, when interviewing is considered an exploration into unchart-ed territories there are no final answers, only dead ends. In ‘analytical inter viewing’ the aim is to develop a dialogue with the respondent that can continue. The ‘follow-up’ questions are the interviewer’s major task. We suggest that the challenge in interviewing is not of a motivational character, i.e., dealing with the reluctant respondent. The real challenge is dealing with the respondent who willingly answers all the questions dili-gently. Such interviewees pressure the interviewer to ask more than mere-ly clarifying questions – or questions that take the dialogue to a new item on the interview guide. If the dialogue is to continue, the answer must be followed by a new question that adds new dimensions, e.g., by pointing out logical implications or theoretical generalizations, and by mobilizing other implicit or absent concerns in the respondent’s account.

The follow-up question is the true art of analytical interviewing. It has

158 t h e a n a l y t i c a l i n t e r v i e w

Art_of_Science2f.indd Avs3:158Art_of_Science2f.indd Avs3:158 05-11-26 14.08.4205-11-26 14.08.42

Page 7: Art of Science

received little attention in the literature, probably because it is hard to plan, and probably also because it requires insight, prior knowledge and skill on the part of the interviewer. In other situations, the interview guide may allow the researcher to have students or professional service firms collect the interview data. Analytical interviewing is different; it requires expert knowledge and requires the researcher to be present, simply be-cause analysis is no longer desk work subsequent to data collection. The capability for posing inspiring follow-up questions requires process skills (interview training and experience), theoretical breadth (new domains of knowledge and experience to be mobilized) and self-confidence (the will-ingness to risk jumping spontaneously on new opportunities for explor-ing linkages to new domains of knowledge and reality without being able to foresee the consequences and the reactions). Fortunately, it is exactly the inspired follow-up questions that motivate the interviewees to be come full participants in the analytical interview. While the interact (question-answer sequence) is easy to plan and control, the double interact (ques-tion-answer-follow-up sequence) is virtually impossible to plan and con-trol.2 What one wants to ask next as an interviewer is to be seen as a gen-uine response to the answer received.

Mono-directional learning versus multi-directional learning

Conventionally, the structural setting of interviewing is assumed to be asymmetrical. The respondent knows something that the interviewer wants to learn. In the conventional interview situation, the only incentive for the respondent to share his or her knowledge with the interviewer is the symbolic recognition of being the knowledgeable superior. However, the discussion above suggests that the character of such social exchange may change in the course of the interview. The conversational character of the interview, stimulated by follow-up questions, suggests that the in-terview process is much more than just an explication and documenta-tion of what the interviewee already knows. By adding the knowledge and creativity of the interviewer, the process may constitute a learning opportunity for the interviewee. He or she may easily realize new connec-tions and linkages, not necessarily the ones introduced by the interviewer, but ones he or she produces in response to the interviewer’s input and conjectures.

2 For the distinctions among act, interact and double interact, see Weick (1969/1979).

c r e a t i v e a n d s k i l f u l r e s e a r c h p r a c t i c e 159

Art_of_Science2f.indd Avs3:159Art_of_Science2f.indd Avs3:159 05-11-26 14.08.4205-11-26 14.08.42

Page 8: Art of Science

That is to say, the analytical interview allows the respondent to be sur-prised by his or her own statements and yet to be able to reflect rational-ly on the surprise so that it is not just a random occurrence. The analyti-cal interview aims to make the respondent think and consider. It aims to get findings ‘beyond’ common sense, and it aims to create new insight rather than confirm what was already known before the interview. The respondent becomes engaged in theorizing over the possible meaning and character of organizational practices previously taken for granted. He or she tests various accounts for their plausibility and helps select the more “interesting, obvious, connected, believable, beautiful, or real” (Weick, 1989:524) ones. Thus, the social situation that started out asymmetrical-ly may conclude, in successful cases, as a much more symmetrically collab-orative one, a multi-directional learning process in which all contribute with their own resources and skills.

Proving a point versus pursuing new meaning

The central concern with method is often related to validity; i.e., ques-tions about whether the interview guide helps to collect information that correctly portrays the phenomenon under study. These concerns assume that the world is there a priori and that the role of the researcher is not to stand in the way and prevent reality from emerging. Because the reality emerges with the researcher as a medium, he or she is a potential source of bias and error. When the researcher can be suspected of having an in-terest in the outcome of the interview, e.g., because he or she has a point to make, a theory to prove and/or a prejudice to legitimize, the risk of manipulation seems high. That risk may seem even higher in the case of analytical interviewing as it is outlined above.

The situation is no different when the interviewee may be suspected of having a different interest in the outcome – only the direction of the ma-nipulation is reversed. Validity and reliability are at risk, whoever is ma-nipulating the other.

But the premise that the world exists a priori and that the aim of the interview is to allow this reality to emerge are different from the assump-tions we made for the analytical interview. Manipulative intentions may exist from the outset, but since the process will meander outside the con-trol of both the interviewer and the interviewee, such intentions may soon become irrelevant. If the exploration into uncharted territories of

160 t h e a n a l y t i c a l i n t e r v i e w

Art_of_Science2f.indd Avs3:160Art_of_Science2f.indd Avs3:160 05-11-26 14.08.4205-11-26 14.08.42

Page 9: Art of Science

knowledge is taken seriously there exists no prior aim or point to prove that will taint the knowledge constructed. This is not to claim no source of bias and error exists. But the sources of such problems lie probably more in incapability and ignorance than in the motives and manipula-tions of the interviewee and the interviewer. When the analytical inter-view ing succeeds, the interviewer and the interviewee are both theorists, collaborating on construction of new knowledge. They may be con front-ed with the same confusion and difficulties that Ryle (1954/2002:7–8) ascribes to theorists,

A theorist is not confronted by just one question, or even by a list of questions numbered off in serial order. He is faced by a tangle of wriggling, intertwined and slippery questions. Very often he has no clear idea what his questions are until he is well on the way towards answering them. He does not know, most of the time, even what is the general pattern of the theory that he is trying to con-struct, much less what are the precise forms and inter-connexions of its ingredi-ent questions… Unlike playing-cards, problems and solutions of problems do not have their suits and their denominations printed on their faces. Only late in the game can the thinker know even what have been trumps.

The methodological costs of involving an interviewer and an interviewee in the knowledge production are not of a type related to validity and reli-ability. They are opportunity costs – the knowledge domain that was not mobilized; the line of thought not pursued; the connections discovered only in retrospect; the ideas discarded too quickly; and the failure to re-cognize the implications of phenomena. Even if the judgment used in se-lecting a conjecture as a premise is made not on validity, but on the vari-ous criteria of plausibility mentioned above (Weick, 1989), judgments may be poor, conservative and unimaginative.

Thus, the analytical interview is more susceptible to the risks of non-participation and incapability, both of which will reduce the knowledge being produced, than to the risks of poor validity and reliability.

Fieldwork examples and illustrations

It is undoubtedly true that the written and published literature reports in-terview answers more than interview questions, and hardly any follow-up questions in the sense discussed here are publicly documented. The re-

c r e a t i v e a n d s k i l f u l r e s e a r c h p r a c t i c e 161

Art_of_Science2f.indd Avs3:161Art_of_Science2f.indd Avs3:161 05-11-26 14.08.4205-11-26 14.08.42

Page 10: Art of Science

searcher is silenced as a matter of good methodology and the active co-construction of meaning in the course of an interview would probably fail to pass beyond the editor’s desk. Descriptions of collaborative and active interviewing do exist and accounts of conversation-type interviews can be found occasionally (Ellis & Berger, 2002). However, while analyti-cal interviewing is a conversation-type interview, all conversation-type in-terviews are not analytical. The ones reproduced in Ellis & Berger (2002) certainly are not. Thus, we are forced to rely on our own research.

Learning about discipline

Our first example concerns an interview with a construction manager. In the interview, we inquired about the difficulties of being a construction manager, mixing his observations and reflections on his practice with the underlying tensions and uncertainties of his job. The illustration is meant to suggest that analytical interviewing may take unexpected turns and may lead to discussions and reflections that often are assumed to be the researcher’s exclusive domain.

At one point in the interview the construction manager mentions a re-current situation that causes him much difficulty and anxiety. When arriv-ing at a construction site he sometimes finds his men taking a break out-side break hours. He admits that his first inclination is to rush in and kick them out into the site. Such lack of discipline deserves punishment, he suggests – but then he reflects on a possible alternative reason for why his men are not working. He suggests they may have sacrificed their offi-cial break to finish some pressing task – a gesture of diligence that de-serves praise, not punishment – and yet, he finds himself typically choo-sing the punishment alternative. After having developed the dilemma in the conversation between the respondent and the interviewer, the latter asks a series of questions that all aim to introduce the complexity and un-certainty of construction work as a new premise for the conversation. The central issue is whether, and under what circumstances, the construc-tion manager is more dependent on the work gang’s discipline than on their diligence. This new issue gives the respondent the opportunity to further reflect on his own practice in light of the context – and to link new aspects of that context to the analysis, e.g., the differences between the gangs, the nature and use of the bonus systems, and the culture of collaboration (see also Kreiner, 1989). The follow-up questions intro-

162 t h e a n a l y t i c a l i n t e r v i e w

Art_of_Science2f.indd Avs3:162Art_of_Science2f.indd Avs3:162 05-11-26 14.08.4305-11-26 14.08.43

Page 11: Art of Science

duced more general work-organizational aspects into the conversation, but did so in a manner that built directly on the interviewee’s own ac-count. The questions allowed him to extend his own lines of thought and made the conversation develop new information and reflections that the interviewer could not have foreseen when asking the question. The result was a much richer understanding of the intricacies of managing construc-tion work that transformed the gangs from a monolith to a complex work unit that could react in different ways to the same stimulus. Here, via the analytical interview, the construction manager could explain how a rule should work in different situations, and indeed could recognise that it would be folly to assume a rule should be implemented in the same way in all situations. To follow a rule is an ambiguous blessing.

Learning about an accounting systems for representing knowledge

This example is drawn from an interview conducted in the process of find-ing out what intellectual capital is and how it works (see e.g., Kreiner & Mouritsen, 2003, Mouritsen & Flagstad, 2005). It is an interview with a HR manager who established an accounting system involving a broad set of both financial and non-financial indicators which was intended to re-present knowledge and its effects on organisational decision-making and value creation. In the specific instance, the HR manager used the EFQM model as a template to organise the indicators that expand the convention-al financial focus in many reporting systems and to add indicators about results about employees, customers and society. The following is an ex-cerpt from the dialogue between the HR manager (the respondent) and the interviewer.

Respondent: In our organisation we use the EFQM model to structure our reporting because we know that satisfied employees lead to happy customers who then are loyal, and this will benefit our bottom line. So employee satisfac-tion is important to us.

Interviewer: Does this mean that when numbers are red, then you invest heavi-ly in HRM practices?

Respondent: Well, [pause] it does not really work this way. Listen, … the financial numbers create a space for our other objectives.

c r e a t i v e a n d s k i l f u l r e s e a r c h p r a c t i c e 163

Art_of_Science2f.indd Avs3:163Art_of_Science2f.indd Avs3:163 05-11-26 14.08.4305-11-26 14.08.43

Page 12: Art of Science

This example shows that the interviewer is not only interested in getting already established information concerning the logic and role of indica-tors about knowledge. The example also says that knowledge requires people to be motivated because otherwise why would they develop and share knowledge. Finally, the example says that knowledge production and sharing is a good thing for the business. This would then have illus-trated the HR strategy of considering people as the main value-creating factor. It would have said that happy employees produce happy custom-ers, leading to good financial results. This line of thought could likely have been followed up on by confirmative questions, such as, “Can you give an example?” If examples were given, the commitment to the origi-nal line of thought would simply be reconfirmed.

Rather than seeking reconfirmation of one line of thought, the analyti-cal interview seeks to introduce a conflicting or competing concern, and keeps the conversation open by challenging the interviewee with a differ-ent line of thought. In this case, this is achieved by simply drawing the claimed causalities to their logical conclusions and (as it turns out) out-side the boundaries for their validity. The interviewer’s interest is not only to understand the details of a people-centred strategy, which is well known from the literature, but also to find the boundaries and dilemmas implicit in such a strategy. Therefore, the interviewer asks a question that is critical because it is theoretical; the question explores the boundaries of the idea of a people-centred strategy. This is a different type of question because, besides being descriptive, it is also contemplative and quite like-ly a highly relevant management dilemma in the HR manager’s ordinary activities. This analytical question tries to produce a logical conclusion and points out a (theoretical) tension in the logic behind the respondent’s claim. The interviewer produces a need for the respondent to develop some of the conditions further that make the claim of the people-centred strategy relevant, even if these conditions are not (currently) in place. Suddenly, the respondent is able to see and report how, under certain conditions, the relationship may be the opposite of his initial proposition. Good financial results are the precondition for being able to talk about employee and customer satisfaction. Not vice versa. This is a new line of thought that apparently surprises even the respondent. The trick is that the interviewer, by engaging in a dialogue about dilemmas or tensions in-formed by theory, is able to develop both theoretical and practical know-ledge.

164 t h e a n a l y t i c a l i n t e r v i e w

Art_of_Science2f.indd Avs3:164Art_of_Science2f.indd Avs3:164 05-11-26 14.08.4305-11-26 14.08.43

Page 13: Art of Science

The respondent’s new answer significantly develops what is at stake in a people-centred organisation. The respondent obviously knows that there are problems in implementing a people-centred strategy and that sometimes even the HR manager does not believe that the firm pursues such a strategy. It could easily be true that the people-centred strategy is not the precondition for value growth in the firm; it could be that finan-cial value is the precondition for a people-centred strategy. Compared with most literature on people-centred organisations, this is an anomaly; it is a new finding. The HR manager was brought to a point where he could formulate a new line of thought that he didn’t know he knew – and in doing so, he learned something more profound about the nuances of the organizational practice.

Understanding statistical information: Taking charge of correlations

The next example of analytical interviewing involves the manager of a large consulting firm. The consulting firm had analyzed data on non-fi-nancial measures and the resulting lagging indicator of financial effects, such as profitability per employee. The idea was that the non-financial information should function as a predictor of the development in financi-al results. Literally, the correlations showed that the most profitable divi-sions in the firm are the ones with old employees, old customers, low em-ployee satisfaction, inflexibility and little cooperation. These numbers told a very unconventional story about the strategy of a consulting firm because consulting firms seldom wish to become rigid and uncommitted to staff development. The following interview sequence followed the pre-sentation of the statistical information:

Respondent: We really have problems because our analysis shows that our humanistic network credo does not work.

Interviewer: So, now you are changing towards a more rigid organisation?

Respondent: No, we have to analyse what the correlations mean and then restore our credo.

The respondent interprets the statistical information and starts by ac-knowledging the tension in the firm’s HR oriented strategy. However, when the interviewer points out the logical conclusion of the results, namely creating a firm less caring about its employees, less flexible and less cooperative, in order to achieve higher profitability, the respondent is

c r e a t i v e a n d s k i l f u l r e s e a r c h p r a c t i c e 165

Art_of_Science2f.indd Avs3:165Art_of_Science2f.indd Avs3:165 05-11-26 14.08.4305-11-26 14.08.43

Page 14: Art of Science

forced to discard the study and discredit the numbers to protect the stra-tegy. The numbers of the EFQM model suddenly present no general truth, but a conditional truth, subject to local conditions. The respondent’s move may seem inconsistent, but it does restore trust in the received wisdom that non-financial numbers are indices of future financi-al results (see e.g., Kaplan & Norton, 2004). The respondent only needs to find the uncontrolled variables that make the numbers look odd.

Probably, the interviewer’s question was not asked with the expecta-tion that the respondent would affirm the consulting firm’s intentions to change the organization towards a more rigid form. The question was keyed by taking the study seriously and by implicitly invoking another in-stitutional framework that has to do with the proper ‘look’ of a modern firm. Faced with two conflicting expectations, the respondent chooses to discredit the study in order to maintain legitimacy.

The follow-up question was analytical in character because it linked two concerns together and therefore defined a dilemma for the respond-ent. We could imagine the interview continuing by exploring, in this study, why institutional legitimacy was more important than trust and why the measures were used not only by this firm but by firms through-out Europe.

Financial analysts’ appraisal techniques

We have tried to teach analytical interviewing in some PhD courses. One assignment for the participating PhD students was to conduct an inter-view with a respondent.3 The following two excerpts from actual inter-views illustrate the difference between the conventional interview and the analytical interview. Both PhD students interviewed a respondent with knowledge about financial analysts’ work. Consider first the convention-al interview that is adapted a bit, but that really took a very short time. The assignment calls for seven minutes of interviewing, but the student could not fill the time slot.

Interviewer: I study how venture capital is allocated. As analyst of a successful venture capitalist firm, what methods do you use to evaluate prospective firms?

3 This is a teaching situation where PhD students were asked to interview a professor of finance who had substantial knowledge about financial analysts. The finance professor accepted role-playing a particular person whom he knew in the financial community.

166 t h e a n a l y t i c a l i n t e r v i e w

Art_of_Science2f.indd Avs3:166Art_of_Science2f.indd Avs3:166 05-11-26 14.08.4305-11-26 14.08.43

Page 15: Art of Science

Respondent: None, I use my intuition.

Interviewer: But surely you must have some method to get to your conclu-sion??!!

Respondent: No, I write the analyses after I have made up my mind.

Interviewer: But, really, finance theory tells that you should …. Thanks for your time

In this interview, the interviewer strongly holds to the belief that the prior theorisation of the situation has to be maintained because the line of questioning is fixed in the questionnaire and is concerned with what is known at the outset of the interview. Clearly, the interviewee rejects the prior finance theory, and the interviewer is not pleased. The interview also shows that the interviewer learns absolutely nothing about the role of the financial analyst in the production of the market for knowledge. The interviewer is handicapped by the interview because it just ‘proves’ that the initial theorising is wrong, but nothing is put in its place; there-fore, there is no learning from the interview.

An alternative line of inquiry comes closer to the analytical interview-ing approach and produces a very different account.

Interviewer: I study how venture capital is allocated. As analyst of a successful venture capitalist firm, what methods do you use to evaluate prospective firms?

Respondent: None, I use my intuition

Interviewer: I know that you are highly esteemed … So, how do you learn about prospective firms?

Respondent: I first flip through the application, but what is really important is that I visit the firm, that I talk to his wife, that I know about his personal finances…

Interviewer: Fascinating, so the whole process is purely your ‘senses’? It is a leap of faith?

Respondent: Not quite. I also listen to friends knowledgeable about the tech-nology and I search around to find possible competitors … Then my firm requires me to write a report

In this sequence the interviewer senses that something is developing which was not expected, but which might be significant. The first answer is a dead end, and the interviewer retreats to requesting information about the analyst’s intelligence activities on prospective firms that he later

c r e a t i v e a n d s k i l f u l r e s e a r c h p r a c t i c e 167

Art_of_Science2f.indd Avs3:167Art_of_Science2f.indd Avs3:167 05-11-26 14.08.4405-11-26 14.08.44

Page 16: Art of Science

evaluates intuitively. For example, talking to an applicant’s wife does not elicit the kind of information required, according to the textbooks. How-ever, allowing the interviewer to characterize the evaluations claimed builds on intuition as a leap of faith. The allusion to randomness and subjectivism provokes the analyst to reconstruct the things that he does and which now appear to constrain the randomness and subjectivism, e.g., talking to friends and colleagues. We, and possibly the analyst him-self, learn that intuition is exercised as the backdrop of the mobilization of a network of resources, like wives, friends, competitors, etc. Without such a network in place, intuition will not work. The interviewee’s first answer is technically correct, but it is a very narrow account of the ‘truth’ since it does not include the context in which intuition is exercised, and on which the sensibility of intuition is totally dependent. A much richer picture is revealed, for the interviewer as well as for the analyst. Although the interviewer knew all the elements from the start, the way they were put together to produce meaning was an outcome of the interview, not a pre-existing insight ready to be explored. Arguably, the interview even contributes to a new theory that compared with conventional finance theory sees the role of calculation very differently (e.g., Hägglund, 2001; Mouritsen, Munk Nielsen, Lindhart & Stakemann, 2005).

These small examples show what the analytical interview can produce. It can develop new ideas and theories, putting lines of thought, concerns and pieces of information together in unexpected ways. As a result, the interviewee in the analytical interview is forced to handle the conflicting concerns, and thereby richness is added to the account of organizational and individual practices.

The challenges and promises of analytical interviewing

We have described and illustrated the analytical interview in detail and have illustrated how it may be conducted. Our discussion of the analyti-cal interview is now conducted at a more general level where we try to reflect on its effects compared with other interview approaches. We need to emphasize that the analytical interview is an addition to existing ap-proaches rather than a completely new reformulation of what intervie-wing processes accomplish. It has many close cousins, e.g., the active in-terview and the collaborative interview, with which analytical inter-

168 t h e a n a l y t i c a l i n t e r v i e w

Art_of_Science2f.indd Avs3:168Art_of_Science2f.indd Avs3:168 05-11-26 14.08.4405-11-26 14.08.44

Page 17: Art of Science

viewing shares, among other things, the conversationalist and dialogical approaches (Holstein & Gubrium, 1995; Ellis & Berger, 2002; Kvale, 2005). Further, we support Alvesson (2003) in his attack on neopositi-vism and romanticism. We also share in the appeal of reflexivity to him and others. Thus, modification to existing themes in the literature is more important to us than the claim that the analytical interview is a complete-ly new affair. But there are variations and we will start by taking stock of such variations. We consider our contribution to have re-introduced and given voice to aspects of the research interview that have laid dormant in the existing literature. Let us highlight some of these aspects.

Qualities of the analytical interview

First, while accepting the need to plan and prepare before the interview, we emphasized the need for presence, i.e., the ability to go with the flow of the conversation. We advocate for presence in two senses. On the one hand, the researcher needs to be present in person. He or she is not im-portant for asking the questions that may be prearranged in the interview guide, but he or she is vitally important for asking the follow-up ques-tions. The analytical training, the breadth of experience and the grasp of the existent literature that he or she represents are necessary resources for the meaningful conversation and the knowledge creation that we call analytical interviewing. On the other hand, physical presence has to be accompanied with a presence of mind. The interviewer must seize the moment, make new connections, see logical implications and produce contrasting interpretations. Only in this sense will he or she become re-sourceful in the interviewing process. In short, in our account the pendu-lum has swung away from previous ideals of the anonymous interviewer who could be substituted for with hired hands in the field meeting. The co-production of knowledge inherent in analytical interviewing totally depends on the unique contribution of the researcher – unique, that is, in relation to the unique contribution of the interviewee.

Second, we emphasized having a strategic intent behind the questions asked during the interview. Obviously, we ask questions to receive an-swers. But the answers we get may facilitate or stop the ongoing conver-sation. Thus, we may ask questions also to obtain answers to the follow-up questions. The success of the interview may be more dependent on the quality of the follow-up questions than on the quality of the answers. The

c r e a t i v e a n d s k i l f u l r e s e a r c h p r a c t i c e 169

Art_of_Science2f.indd Avs3:169Art_of_Science2f.indd Avs3:169 05-11-26 14.08.4405-11-26 14.08.44

Page 18: Art of Science

analytical interviewing moves the focus from the single interact to the double interact.

Third, the role of the respondent is to surrender information and in-sights about the organizational practice under study. However, the ana-lytical interviewer will not be content to document the information and insights that the respondent possesses at the start of the interview. This would be to abandon the possibility that he or she can learn something in the course of the interview. Since the role we designate for the interview-ee is one of an active participation in the knowledge production, the mark of success is a more knowledgeable interviewee at the end of the in-terview. Learning is not a by-product of the conversation, but the focal objective, for the interviewer as well as for the interviewee. Thus, the in-terview should be conducted in ways that force the respondent to think, to reflect, and to learn more of the reality that he or she is already know-ledgeable about.

Finally, we emphasized the risks of manipulation, both in terms of the interviewer manipulating the interviewee to confirm his or her prejudicial theories, and of the interviewee manipulating the interviewer to relay his or her biased and self-serving presentation. But we also emphasized the fact that the analytical interview is an exploration into uncharted and un-planned territory. It is not clear what the spoils will be, nor is it easy to guard against or manipulate such unknowable outcomes. Thus, the ana-lytical interview does not ignore the risk of manipulation and biases, but promotes a process that renders manipulative intent ineffectual.

Such are some of the modifications that we have claimed as character-istics of analytical interviewing. We will discuss two central implications of these modifications subsequently.

The move away from the researcher’s desk

The researcher’s desk is, in many ways, a safe haven for analysis and re-flection. Alvesson (2003:30–31) points out that this is where the hard work begins:

Instead of relying strongly on the researcher to optimize the interview as a technique or tool and/or to work hard in interview encounters at getting inter-viewees to be honest, clear, and consistent, the message expressed here is rath-er that the hard work should be conducted at the desk … Fieldwork is, of course, important, but the complexities and pitfalls involved call for careful, ongoing reflection.

170 t h e a n a l y t i c a l i n t e r v i e w

Art_of_Science2f.indd Avs3:170Art_of_Science2f.indd Avs3:170 05-11-26 14.08.4405-11-26 14.08.44

Page 19: Art of Science

Alvesson discusses eight metaphors for thinking about the interview in order to spur reflexivity at the researcher’s desk. Clearly, this is im-portant. However, in our view, the hardship during the interview has little to do with “getting the interviewee to be honest, clear, and consistent”, as Alvesson claimed above. It is much more about getting the interviewee to be creative, analytical, and reflective – and much more about having interesting data for the researcher to reflect upon when he or she sits at a desk. We see no reason to assume that reflection requires a desk – that re-flection is necessarily a separate process to be located in time and space. Reflexivity could also be an approach and a state of mind by which ordi-nary processes are conducted, including interviewing. There is no reason that the interview questions could not be phrased with an acute aware-ness of the eight metaphors – no reason why the conversation should not benefit from the mixture of inspirations and interpretations all provide. Additionally, there is no reason that the interviewee should not equally be (or become) reflective in answering the interview questions and contri-buting to the conversation.

If the researcher’s desk is the designated place for reflection and theo-rizing, we propose to bring that desk into the field and have the interview take place around it. When the desk is returned to the researcher’s do-main, the chances are that it will be covered with much richer data for the researcher.

Dilemmas as the motor of creativity and analysis

We have pointed out that metaphors can be a creative force in the researcher’s reflections, and also in the interview process. Rather than metaphors, in our discussion of the analytical interview, we chose dilem-mas as the motor of creativity and analysis. Let us return to reflect on the implications of this choice.

The dilemma as a central element of practice introduces implicitly the notion of choice. By pointing out the conflicting and competing concerns we are also reminded that the litigation between them may result in a va-riety of decisions. The decision taken is significant (i.e., says something important about practice) exactly because something else might have been decided. As Giddens (1991) explicitly points out, in high modernity societies, individual choice making is always in progress with a view to-ward the transformation of the world and the individual’s role in it. The

c r e a t i v e a n d s k i l f u l r e s e a r c h p r a c t i c e 171

Art_of_Science2f.indd Avs3:171Art_of_Science2f.indd Avs3:171 05-11-26 14.08.4405-11-26 14.08.44

Page 20: Art of Science

autobiography of the individual is developed by choice-making that, in principle, involved options that were not pursued in the actual course of events.

We conclude from these reflections that empirical inquiry is not just about what practice is factually observable. To interpret these factual practices, we need to know and understand the alternative practices that were not chosen. Interviewing about organizational practices also in-volves interviewing about the reality that is not there – the counterfactual reality. In this counterfactual reality, the interviewee probably has no pri-vileged knowledge. Indeed, the interviewer may actually be better equip-ped to imagine and describe the counterfactual reality, and thus, we find another argument for collaborative effort in the research interview.

The dilemmas introduce the notion of the counterfactual reality that becomes the main instrument for exploring the richness of the actual practice. They introduce an important source of inspiration for the inter-view questions, namely the idea of opportunity costs. The actual practice has many justifications, but so have the other concerns that receive less support by the practice studied. These ‘neglected’ concerns remain valid concerns that are either neglected at a cost (the justification of which could guide the further questioning) or are coped with in other ways (which would take the interview to new domains of organizational prac-tice). Here we begin to imagine an analytical tree where the pursuit of the hidden costs of current practice represent the narrower and narrower choice of alternatives, and where the alternative domains of organizatio-nal practice represent the separate branches of the organization, each with its internal logic and dynamics. The creativity potentiality lies in pursuing the logics and justifications to more discriminating nuances, as well as in jumping from one branch to another.

Validity and plausibility in knowledge production

It is clear from our discussion that we wish to supplement validity and re-liability with another notion of good research. Weick’s (1989; 1998) idea of disciplined imagination and reflexivity is about plausibility (rather than abstract truth) and appeals to us because it does not say that non-validity and non-reliability are relevant alternatives. Plausibility does not suggest that the findings are ‘un-scientific’ but rather that they make a lot of sense within the analytical world that has been constructed through

172 t h e a n a l y t i c a l i n t e r v i e w

Art_of_Science2f.indd Avs3:172Art_of_Science2f.indd Avs3:172 05-11-26 14.08.4505-11-26 14.08.45

Page 21: Art of Science

the interview. Weick (1989:524) pays attention to accounts of relevant actors that are “interesting, obvious, connected, believable, beautiful, or real, in the context of the problem they are trying to solve”. These cha-racteristics require the world of affairs to be present because how other-wise could a finding, a conjecture, be of such quality? The constructions to which such qualities can be attached are real in the sense that they in-corporate “considerable past experience with related problems that the theorist [both interviewer and respondent, we may add] brings to bear on the conjecture” (Ibid.: 525). The idea of plausibility is about the affairs of the world but it does not presume that verification and validation can help us to find all possible and realistic states of affairs. Many possibili-ties are not realised in practice, but this does not make them less real as options for decision-making and as elements to be considered in desig-ning organisations and actions.

We suggest that the analytical interview is a research strategy where the production of knowledge and the collection of data are intermingled. Knowledge is a construction but not in the sense of being ‘simply made up’. It is a construction in the sense that it flows from the interview where both interviewer and respondent may grow in wisdom. No one is able to understand his or her position in the world completely, and even if he or she were, there would be tacit knowledge that cannot be expressed in words. Therefore, there is no reason to think that the respondent, by being allowed to speak, will tell merely what he or she knows. Nor would it be prudent to think that he or she would tell something of theoretical interest, either to the researcher or to him- or herself.

The analytical interview is a means to make the dialogue continue in an interesting way. It is a means for capturing the theoretically interesting in an interview because interesting empirical evidence is not only a matter of the ‘size’ of empirical evidence, but also of whether it has been spun in theoretically interesting networks of claims, counter-claims, factual and counterfactual realities, understandings and reinterpretations. The evi-dence is a piece of singular data; it is also an explanation with a thesis, replete with contradictions, omissions, exceptions and breakdowns. This is what is at stake in an analytical interview. It may be, therefore, how we can learn; how we – both interviewer and the interviewee – actually be-come wiser. Analytical interviewing is a challenge to conventional, and still legitimate, concerns with validity and reliability because surely know-ledge is constructed and not only discovered through the interview. And

c r e a t i v e a n d s k i l f u l r e s e a r c h p r a c t i c e 173

Art_of_Science2f.indd Avs3:173Art_of_Science2f.indd Avs3:173 05-11-26 14.08.4505-11-26 14.08.45

Page 22: Art of Science

yet this concern is not really as problematical as one should imagine be-cause even if the analytical interview is not primarily concerned with being certain – it is not primarily afraid of creating knowledge that is wrong – it is more concerned with developing insight – with developing new propositions. Therefore, it may be that the logic and discipline of the analytical interview, as it is firmly anchored in theoretically interesting propositions even if they may change in the course of the interview, are more relevant for the production of substantive knowledge because data are not separated from analysis. Reflection is not postponed to the researcher’s desk, as it may be in conventional interview research. Reflex-ivity is achieved in situ, and therefore it may contain many of the ele-ments that make the proposition strong. Reflexivity is part of the inter-view, as well of the work that follows it.

The analytical interview is no panacea. It requires considerable theo-retical insight. As it requires the researcher to be able to shift positions during the interview, a certain theoretical maturity is necessary. Inter-viewing is not about assuming a ‘blank sheet’ or a non-involvement, al-lowing data to just move through the body of the researcher. The inter-view begins with the premise that both researcher and respondent are knowledgeable about the situation they are discussing, and they therefore know enough to construct a series of plausible dilemmas that inform this situation. Without such resources and efforts, it would not be possible to learn. There would be little that resulted from the interview that contrib-utes to the respondent’s knowledge of the situation, or adds to the litera-ture.

In presenting our account of analytical interviewing, we do not wish to describe it as a specific method. We share Czarniawska’s (2002:747) view that “there is no one thing that a researcher must, should, or can do . . . Much more important than a specific interpretive or analytic technique is the result – an interesting recontextualization.” There is much more to the construction of interesting knowledge than interviewing, including analytical interviewing. Without the ability to specify the route to inter-esting results, we must face the challenge of devising strategies that are conducive to such achievements. We have argued that analytical inter-viewing is such a strategy.

174 t h e a n a l y t i c a l i n t e r v i e w

Art_of_Science2f.indd Avs3:174Art_of_Science2f.indd Avs3:174 05-11-26 14.08.4505-11-26 14.08.45

Page 23: Art of Science

References

Alvesson, Mats (2003). Beyond neopositivits, romantics and localists. A reflexive approach to interviews in organisational research. Academy

of Management Review 28(1): 13–33.Briggs, Charles L. (2002). Interviewing power/knowledge, and social in-

equality. In Jaber F. Gubrium & James A. Holstein (Eds.), Handbook

of Interview Research. Context & Method. Thousand Oaks: Sage, 911–922.

Czarniawska, Barbara (2001). Is it possible to be a constructive consult-ant? Management Learning 32(2): 253–266.

Czarniawska, Barbara (2002). Narrative, interviews, and organizations. In Jaber F. Gubrium & James A. Holstein (Eds.), Handbook of Inter-

view Research. Context & Method. Thousand Oaks: Sage, 733–749.Cyert, Richard M. & March, James G. (1963/1992). A Behavioral Theo-

ry of the Firm. Cambridge, MA: Blackwell.Ellis, Carolyn & Berger, Leigh (2002). Their story/my story/our story. In-

cluding the researcher’s experience in interview research. In Jaber F. Gubrium & James A. Holstein (Eds.), Handbook of Interview Re-

search. Context & Method. Thousand Oaks: Sage 849–875.Giddens, Anthony (1991). Modernity and Self-identity. Self and Society

in the Late Modern Age. Oxford: Polity Press. Gubrium, Jaber F. & Holstein, James A. (2002) (Eds.). Handbook of In-

terview Research. Context & Method. Thousand Oaks: Sage.Holstein, James A. & Gubrium, Jaber F. (1995). The Active Interview.

Qualitative research methods series, No.37. Thousand Oaks: Sage.Hägglund, Peter (2001). Företaget som investeringsobjekt. Stockholm,

EFI, Stockholm School of Economics.Jönsson, Sten (1996). Accounting for Improvement. Oxford: Pergamon.Jönsson, Sten (1998). Relate management accounting research to mana-

gerial work! Accounting, Organizations and Society, 23(4):411–434.Jönsson, Sten & Macintosh, Norman B. (1997). Cats, rats, and ears:

Making the case for ethnographic accounting research. Accounting,

Organizations and Society, 22 (3–4): 367–386. Kaplan, Robert S. & Norton, David P. (2004). Strategy Maps. Convert-

ing Intangible Assets into Tangible Outcomes. Boston: Harvard Busi-ness School Publishing.

c r e a t i v e a n d s k i l f u l r e s e a r c h p r a c t i c e 175

Art_of_Science2f.indd Avs3:175Art_of_Science2f.indd Avs3:175 05-11-26 14.08.4505-11-26 14.08.45

Page 24: Art of Science

Kreiner, Kristian (1989). Culture and meaning: Making sense of conflict-ing realities in the workplace. International Studies of Management &

Organization, 19(3): 64–81.Kreiner, Kristian & Mouritsen, Jan (2003). Knowledge management as

technology – Making knowledge manageable. In Barbara Czarniaw-ska & Guje Sevón (Eds.), The Northern Lights: Organization Theory

in Scandinavia. Malmø: Liber & Copenhagen, Copenhagen Business School Press, 223–247.

Kvale, Steiner (2005). Inter Views. An Introduction to Qualitative Re-

search Interviewing, 2nd ed. London: Sage. March, James G., Schulz, Martin & Zhou, Xueguang (2000). The Dy-

namics of Rules. Change in Written Organizational Codes. Stanford: Stanford University Press.

Mouritsen, Jan, Munk Nielsen, Jakob, Lindhart, Jesper & Stakemann, Benedicte (2005). The capital market and the knowledge economy: the analysts’ inscriptions and the knowledge based company, Present-ed at 1st Workshop on Visualising, Measuring, and Managing Intangi-bles and Intellectual Capital, Ferrara, Italy, Oct 18–Oct 20.

Mouritsen, Jan & Flagstad, Kirsten (2005). The making of knowledge society: Intellectual capital and paradoxes of managing knowledge. In Barbara Czarniawska & Tor Hernes (Eds.), Actor-Network Theory

And Organizing. Malmø: Liber & Copenhagen: Copenhagen Business School Press, 208–229.

Ryle, Gilbert (1949/1990). The Concept of Mind. London: Penguin Books.

Ryle, Gilbert (1954/2002). Dilemmas. Cambridge: Cambridge University Press.

Weick, Karl E. (1969/1979). The Social Psychology of Organizing. Read-ing, MA: Addison-Wesley.

Weick, Karl E. (1989). Theory construction as disciplined imagination. Academy of Management Review, 14 (4): 516–531.

Weick, Karl E. (1998). Theory construction as disciplined reflexivity. Academy of Management Review 24(4): 797–806.

176 t h e a n a l y t i c a l i n t e r v i e w

Art_of_Science2f.indd Avs3:176Art_of_Science2f.indd Avs3:176 05-11-26 14.08.4505-11-26 14.08.45