Option #2: Student Developed Case in Project Metrics You have had many experiences in this class with measuring, metrics, KPIs, dashboards and scorecards. You had the opportunity to interview a projec

The ‘mangle’ perspective focuses on the dynamic intertwining of people, technology and concepts; this can provide the basis for an insightful narrative describing the reality of the project in terms of the planned approach, the problems met and the outcomes. Two examples are given, one of a conventional model building exercise, the second of a ‘soft OR’ intervention: both describe projects conducted by practitioners for commercial purposes. It is concluded that, by using the mangle perspective, the OR case writer can winnow the wheat from the chaff in order to write a succinct informative narrative, a narrative that could be utilized by behavioural OR (BOR) researchers. It is further concluded that BOR researchers should engage with ‘practice theory’ to deepen their understanding of what actually happens in projects.

Journal of the Operational Research Society(2017)68(5),507–520. doi:10.1057/s41274-016-0011-8; published online 7 October 2016 Keywords:practice of OR; process of OR; case studies; behavioural OR; soft OR 1. Introduction Social science methodology is anchored by a number of basic precepts that are rarely questioned by [academic] practitioners. One precept that is central to the logic of analysis is the idea of having cases. Ragin (1992, p. 1).

The aim of the paper is to suggest how case studies can be written to serve the purpose of those undertaking research into the process of OR; in other words, serve the purposes of BOR research. I was first alerted to the special demands of writing case studies when I submitted a paper for publication in Interfaces(Ormerod,1996). The editors made their specific requirements for case studies very clear. I had previously published a case study of the same project in another journal (Ormerod,1995). Comparing the two accounts, it is clear that theInterfaces’narrative is the more compelling read. For example, when it comes to describing the action, the earlier paper starts: ‘The approach proposed by the consultants was …’ (Ormerod,1995, p. 281). Note the use of the passive voice and the impersonal reference to ‘consultants’, despite the fact that I was referring to myself. By contrast in theInterfaces paper I say ‘I organized the project as…’ (Ormerod,1996, p. 105). The second account is simpler, more direct and morelikely to hold the reader’s interest. When I subsequently published cases in other OR journals (references to these can be found in Ormerod,2005), I found the requirements equally demanding, but they were not specifically aimed at the writing of case studies. I was able to revert to my previous habits; the narratives became less personal and the passive tense was often used in preference to the active. In retrospect, I no longer achieved the immediacy, the clarity and the pace of an Interfacespaper.

In 1998, there was some correspondence in theJournal of the Operational Research Societyabout the writing of case studies. I suggested some editorial guidelines for practice papers as follows:

To be of interest [in a ‘practice paper’] there needs to be something new, knotty, substantial or interesting that arises from implementing, or attempting to implement, some aspect of OR. The practice paper should:

(i) describe the wider context (ii) identify the roles of the authors and other actors (iii) describe the process and progress (both managerial and technical) of the intervention including blind alleys (iv) identify the problems with implementation (v) describe and where possible quantify the outcome of the intervention whether successful or not *Correspondence: Richard Ormerod, Warwick University, Coventry CV4 7AL UK.

E-mail: [email protected] Journal of the Operational Research Society (2017) 68,507–520 ª 2016 The Operational Research Society. All rights reserved. 0160-5682/17 www.palgrave.com/journals (vi) be written in the first person singular or plural (vii) be an interesting story to read. (Ormerod,1998, p. 1223).

I was looking at the issue from the perspective of practitioners who needed guidance and encouragement to write about their experiences. Gerald Midgley (an academic) gave the perspec- tive of researchers who wanted case studies to provide suitable research material; this second perspective is particu- larly pertinent here:

Theory also allows practice to be seen from a variety of different perspectives. If we are prepared to look at a problem situation or application through several different theoretical ‘lenses’, this increases the insights that become available to us.…it might be useful for the Journal to make a special effort to encourage the submission of papers on the process of OR, with an emphasis on work that illustrates insights into process with examples from practice. (Mid- gley,1998, pp. 1219–1220).

I have had doubts about whether such case studies can be achieved by practitioners in practice. In the OR domain, the Interfacespublication process ensures a certain quality of the writing but does not necessarily result in the sort of description of the process of OR that Midgley and other researchers would welcome. One of the problems is that whereas technical descriptions can usually be relatively brief, process descrip- tions are difficult to write succinctly. The difficulty faced by a practitioner who has been immersed in the hurly burly of organizational reality for the duration of the project is: what should be put into a journal paper and what should be left out?

Organizational change is inevitably complex (see for instance, Pettigrew,1985), and the interests of the readership are varied.

The usual solution for papers submitted to OR journals, is to stick to the technical facts; the process aspects are thereby eliminated or so curtailed as to provide little insight.

How can the situation be improved? How can better (in terms of process description suitable for research purposes) case studies of journal length be written? In other words can a case study be written which is more than a technical account but less than a full account of the process, a narrative that describes the twists and turns, the confrontations and compromises, and the successes and setbacks. One approach, that of Andrew Picker- ing, seems to offer a way of achieving these aims in a reasonably succinct way. Pickering refers to the dynamic process of action, reaction and interaction as ‘mangling’ (Pickering,1995). The question addressed in this paper is: could this approach or something like it be used to write more informative yet concise OR case studies for use in BOR?

The term behavioural operational research (BOR) has recently been introduced into the OR literature by Ha¨ ma¨ la¨ inen et al(2013). In an introduction to a special edition of the European Journal of Operational Researchin 2016, the joint editors say: ‘Events such as the 2014 IFORS and 2015 EURO conference streams on BOR, the creation of a BOR nationalinterest group sponsored by the UK OR Society, and the launch of a BOR website portal hosted by Aalto University (bor.aalto.fi) are all a clear testimony of the revival of the BOR agenda within our community’ (Franco and Ha¨ ma¨ la¨ inen, 2016). It is clear from the special edition papers that the definition of BOR is still the subject of debate (see for instance, Becker,2016; Brocklesby,2016; Franco and Ha¨ ma¨ la¨ inen,2016; Velez-Castiblancoet al,2016; White, 2016). Much BOR is likely to focus on the way that psychological and organizational behaviour can be represented in models. Here the interest lies in research into the practice of OR and the contribution that well-writtenpractitionercase studies can make to such research. Thus, action research and case study research conducted byacademicsare not the focus of attention (Benbasatet al,1987).

The structure of the paper is as follows. The next section introduces the ideas of the mangle perspective as developed by Pickering. The following two sections are narrative accounts of two projects derived by applying the mangle perspective.

The first relates the development of a mathematical model of international coal trade. The second tells the story of helping a large UK supermarket company develop their information systems strategy. Both accounts talk about the people involved, their disciplinary habits, the networks they belong to and the material objects they come into contact with. The mangle at work is thereby illustrated. The fifth section discusses some methodological problems associated with descriptions of practice and the prospects of suitable practi- tioner case studies being written for BOR research into the process of OR. The sixth section suggests some research directions.

In a previous paper, the argument was made that technical OR cases written by practitioners could be made more informative by adopting the mangle perspective (Ormerod, 2014). Here the aim is to provide (i) a second example of applying the mangle perspective to a technical OR modelling example and (ii) explore whether the same approach can equally help practitioners write more insightful ‘soft OR’ case studies. All three cases address the hypothesis that the mangle perspective applied to OR case study writing results in a richer and more insightful account, potentially useful for BOR research.

2. The mangle perspective 2.1. The theoretical foundations The mangle perspective is an extension of the academic research programme (or school of thought) known as the sociology of scientific knowledge (SSK) (Bloor,1976).

Pickering starts from the idea that practice should not be viewed in terms of facts and observation, but should rather be approached from the perspective that scientific practice involves actions through time. This he refers to as the ‘performative idiom, an idiom capable of recognizing that the 508 Journal of the Operational Research SocietyVol. 68,No. 5 world is continually doing things and so are we’ (Pickering, 1995, p. 144). Such a view requires the concept of agency:

who or what motivates and controls the forward momentum of action? Actor–Network Theory (ANT) (Latour,2005) expands the human-centred view of SSK with the claim that material, machinic things (such as radar and computers) can also be taken to provide agency (use of both SSK and ANT in the OR domain can be found in Keys,1998). Pickering goes a step further allowing agency to reside in ‘concepts’ as well.

The concepts that Pickering refers to are to be found in the practice culture of the researchers (we are considering here operational researchers, in other words OR practitioners) under consideration; the culture will be influenced by the discipline in which the practice is embedded and, of course, wider social beliefs. However, it is the narrower concept of research/ practice culture that Pickering focuses on. The concepts constituting this research/practice culture influence the steps that, out of disciplinary habit, researchers and practitioners take to solve problems.

As a project progresses, difficulties are invariably encoun- tered; people or circumstances block the intended path. If a project is to come to a successful conclusion these problems have to be overcome somehow. Pickering describes this as the ‘dialectic of resistance and accommodation’; as problems are met (resistance) and overcome (accommodation), one or more variables (material, human, cultural) are changed. The insight of the ‘mangle’ is that the change in one variable causes all the elements to change – a mangling, a dynamic intermingling, occurs; none of the elements are taken to hold constant in order explain changes in the others. The mangle is Pickering’s implementation of his preferred performative idiom.

Pickering’s conceptual framework can be located within the theoretical perspective of social theory referred to as ‘social practice theory’. The close connection between Pickering’s mangle perspective and the practice theory school can be gleaned from Reckwitz (2002, p. 259): ‘Practice theory… encourages a shifted self-understanding. It invites us to regard agents as carriers of routinized…complexes of bodily movements, of forms of interpreting, knowing how and wanting and of the use of things’. As explained by Nicolini (2013, p. 162), contemporary theories of practice, developed by a range of social theorists, have ‘elaborated and refined Heidegger and Wittgenstein’s initial intuition that phenomena such as knowledge, meaning, identity, power, language, social institutions and transformation are ‘‘housed in’’ and stem from the field of social practices. Although their views are only partially aligned, so that there is no such thing as a unified Heideggerian and Wittgensteinian practice-based approach or school, these authors have all contributed to a common project, or school…for all these authors the starting point for theorizing human affairs is social practices and their connec- tions, and not well-formed individuals [under-socialized rational agents, for instance rational economic man] or overarching systems’.Nicolini concludes: ‘In spite of building on the shoulders of giants, practice theory is still in its infancy and whether it will ever become a powerful bandwagon is yet to be determined.

Most importantly, the practice approach is still largely untested.

As suggested by Latour, the proof of the approach will be in the capacity of future texts to represent practice in a rich insightful way. Therefore, the way forward is, first and foremost, to develop the approach by using it’ (Nicolini,2013, p. 240). 2.2. The template for projects The second part of the mangle perspective is based on the empirical observation that projects usually follow similar patterns during their lifetimes. The idea is that the case writer should identify (i) periods of fairly routine activity involved in getting the job done (following disciplinary habit) and (ii) significant events that impinge on routine activity and give rise to a mangling. Thus, the writer is asked to think about the dynamics of the project, to identify significant events, and to note the consequences.

Pickering suggests that projects usually consist of three types of activities: bridging, transcription and filling. Bridging involves creating a bridgehead, a concept space in which the ‘model’ will be built (Pickering,1995, p. 129). He uses the term model in a very general sense – it could mean a mental model, a conceptual model or a mathematical model. Tran- scription is the copying of established moves from an existing system into the new space created by the bridgehead (p. 116).

This means rooting around for something in one’s experience, in not too dissimilar circumstances, which could be used as a starting point: a point of departure. This could be a technical method or a process of intervention that you had previously used. You copy your old model or approach into the chosen conceptual space (the bridgehead), thus creating a new system, or ‘base’ model. Work then has to start to adapt this base model to the new project circumstances; this work is what is meant by filling. Pickering describes the process of filling as ‘completing the new system in the absence of any clear guidance from the base model’ (p. 116). For example, you could take a scheduling model, which you had developed in your last project, and tailor it to meet the requirements of the new client. In my experience that this is how most of us proceed most of the time. The approach may turn out to lead to a dead end, in which case one looks for a new starting point, a new bridgehead and base model.

2.3. The application of the mangle perspective In seeking to apply the mangle perspective when writing a case study, it should be remembered that the idea is neither to force fit the narrative into the mangle template, nor to highlight evidence of mangling when it was in practice unimportant; after all, the aim is not to confirm or reject the academic theory Richard Ormerod—Writing practitioner case studies to help behavioural OR researchers ground their theories 509 of mangling but to write case studies that succinctly describe the progress of the project in an informative way. It is essential to avoid distorting the narrative except in the direction of a clearer description of what happened and why – that is the aim of good case writing. The power of the mangle approach lies in its ability to help the writer recognize and reflect on underlying causes, connections and dynamics.

The OR reader may well find that in Pickering’s original exposition the language of the mangle approach is unfamiliar and the intent, in the context of OR practice, not altogether clear. In part this is because of its roots in sociology and the use of the language of sociology, but it is mainly because it is cast at a very high level of abstraction to cope with the many and varied types of investigatory projects that Pickering has in mind. My task here is to bring the mangle concept down into the OR domain. One example has already been published (Ormerod2014, pp. 1248–1253); the project in question was the development of a ‘what…if’ model of the UK energy economy. This seemed to work well and conclusions were drawn. In the next section, I will use a second example of technical OR modelling to illustrate the application of the mangle concept and to examine the earlier conclusions. I will then test the approach on the Sainsbury’s case involving soft OR. The cases are intended to contribute to the testing of practice theory in general and the mangle perspective in particular by demonstrating that, by taking a mangle perspec- tive, the practice of OR can be described in a richer and more insightful way.

3. The development of a world coal model The case presented here describes the process of model development. The mangle perspective is used to highlight the key episodes. In brackets it is explained how the perspective has been applied such that, when the bracketed comments are all deleted, what remains is the case narrative of the process of intervention. There is no existing technical case study of this project; at the time the technical aspects of the modelling were not considered sufficiently interesting (complex or innovative) to warrant publication. However, theuseof the model has been discussed in Ormerod and McLeod (1984) and Ormerod (1997). A full-blown case study would bring together the process aspects (described here), the technical aspects and the use aspects.

3.1. The need for a model is identified In the early 1980s, I was the Deputy Director of the Central Planning Unit (CPU) of the UK’s National Coal Board (NCB).

The NCB (later the British Coal Corporation) ran the nationalized coal industry covering all the coal mining and related activities in the UK. Politicians, geo-political analysts and corporations who were in the business of producing,trading and using energy (for instance, oil companies, shipping companies and electricity generators) were interested in whether international trade in steam coal would grow rapidly to (i) fill the gap left by dwindling US oil supply supplies, (ii) loosen the grip of OPEC (a cartel of oil producers including Saudi Arabia) on world energy supply and (iii) meet the needs of rapidly industrializing developing countries such as China.

There already existed some international trade in coal but this was mainly in specialized coking coal for the steel industries, not in steam coal for electricity generation. Nevertheless, from a low base, steam coal trade was starting to grow rapidly and from the NCB’s perspective there might be an opportunity to export some coal. However, more strategically, the growing supply of cheap steam coal, primarily from Australia and South Africa, posed a major threat to the NCB’s home market.

I was faced with a growing flow of reports, statistics and general intelligence about the growth of coal trade coming across my desk: this included reports on who was doing what, their declared intensions and published projections of how the position might develop. I was also having discussions with key players and analysts as part of my day-to-day activities. I was finding it difficult to make sense of it all. I concluded that, as a coal company, the NCB needed to understand more about the underlying economics of coal trade. As I was responsible for commissioning the development of strategic models (as a client), I was in a position to do something about it. CPU needed to include a world coal model in its armoury of strategic models. Such a model would provide a framework to help us structure and make sense of the intelligence we were receiving.

The new model would be developed for CPU by OR scientists in the strategic modelling section of the NCB’s Operational Research Executive (ORE). The model would thus become part of a suite of models already developed by ORE to support CPU’s strategic thinking (Plackettet al,1982). One of the intended purposes of the model would be to develop projections of international steam coal prices within scenarios of world economic and energy futures. We used these scenarios to provide parameters for our models of UK energy flows. These models were being used to support the NCB’s financial submissions to the Department of Energy and other bodies (Ormerod and McLeod,1984; Ormerod,1997). More directly, the projected international price of coal would be used to evaluate our investments in production capacity. 3.2. A base model is established The strategic modelling team had earlier developed a multi- period LP model to help CPU understand the underlying economics of UK energy flows. The model, called ITEM (investment and technology in the energy market) was proving useful in that it provided a test bed for exploring assumptions and testing ideas. We wanted the new model to fulfil a similar role. [We had created a bridgehead, a conceptual space within 510 Journal of the Operational Research SocietyVol. 68,No. 5 which the new model would be built; the concept being to develop a test bed based on economic thinking]. We decided to adopt the multi-period LP approach of ITEM. [ITEM would be taken as the ‘base’ model].

3.3. Filling begins Martin Plackett, and his OR team started the development of the model with (i) each coal mining area able to produce coal and transport it to a port at a cost, subject to the physical constraints of mine and port capacity, (ii) each exporting port able to ship coal to any other port to meet the demands of importers at a cost and (iii) the objective function taken as minimizing the total cost of meeting the demand for coal. [The initial transcription]. We then considered the links between time periods and used the dual-equilibrium method to account for the long term. [The transcription completed]. Work could then start on populating the model with data. [The process of filling in]. For instance, we obtained reports on coal reserves and the cost of mining and moving the coal to port in the different coalfields within the major exporting countries. This enabled cost structures to be approximated for inclusion in the model. [Routine activities for operational researchers]. At this time, LP packages on PCs were being introduced providing us with an order of magnitude improvement in ease of use. The technology thus did not give rise to rise to any issues. [One material aspect which was present in the energy model case (Ormerod,2014) was thus eliminated].

3.4. Model building proceeds During this period, Martin Plackett moved from ORE into the Central Planning Unit (CPU) and John McLeod in ORE took over as leader of the strategic modelling team. [A change that could trigger a mangling]. I had previously worked in ORE and we all shared a desire to develop a model whose assumptions were transparent, factually based, and followed some logical reasoning. [Each of the main actors had a background in ORE, schooled in its research/practice culture, its way of thinking, its habits and its ethos; thus no resistances were generated as a result of conflicts of disciplinary culture, no accommodations were required and no mangling occurred]. This uniformity was not necessarily a good thing as ‘groupthink’ could prevail; an injection of economic concepts was needed. While the OR modellers concentrated on developing and running the model, I became the conduit for economic thinking. As part of my day-to-day activities as Deputy Director of CPU, I participated in various networks of energy economists and had access to, and general dealings with, economists in Government and the large energy corpo- rations. [Both the UK Government and energy companies are collectivities (a collectivity is a group of people who think of themselves as belonging to an identifiable social unit) withaction programmes (a programme of activities undertaken by a collectivity) from which they viewed the energy markets and the players interacting within it; their perspective was generally informed by economics and, of course, politics]. I was therefore in a good position to provide the economic perspective. 3.5. A resistance is met and accommodated The model could now be used to explore the market structure and its properties under various sets of assumptions (scenar- ios). [During this period of routine activity, the day-to-day activity is dictated by the requirements of running the model; the model, a machine-like object, has temporarily become the actor in control]. In my view, the results did not seem very realistic when compared to actual buyer behaviour; it seemed that too much had been left out. [A resistance had been met].

One of the issues of the day was the decision by the Danish Government to limit dependence on South African coal (in the era of apartheid) by restricting the proportion of coal imported from South Africa in any 1 year. [The Danish Government is a collectivity with an action programme from which it inter- preted the actions of the South African Government, another collectivity with its own action programme]. We tested the introduction of constraints to limit the dependence of Denmark on imports from South Africa. This was simple to do and seemed to make the results a bit more realistic. We therefore introduced the more general ability to constrain flows from any source to any country. Now we were able to model the desire of coal buyers to spread their bets. The results were much more plausible. [An accommodation had been achieved].

3.6. The model is used The model could now be run under different assumptions of demand, discount rate and the various constraints; these assumptions were chosen to be consistent with possible economic and energy scenarios. As new information came in (for instance, a requirement for low sulphur coal to combat acid rain, the opening up of new sources for coal, or the announcement of new port capacity), new questions could be asked, new insights gained. [A period of relative calm ensued, during which disciplinary habits took over but with some disturbance caused by changes in the real world of coal trading].

Gradually, the model came to be accepted internally for long-term planning purposes and it also achieved some recognition externally; for instance, a report by the Economic Assessment Service of the International Energy Agency’s Coal Research Centre was based on runs of our model (Long,1982).

We also used the model to provide projections in support of the NCB’s evidence on the building of a new nuclear power station at the Sizewell B Inquiry; these projections were Richard Ormerod—Writing practitioner case studies to help behavioural OR researchers ground their theories 511 accepted by the inspector and his advisors to the detriment of the case made by the nuclear proponents; however, it was a Pyrrhic victory: The NCB, with the help of coal trade model, won the battle but this did not prevent consent being given to the CEGB to build the first PWR nuclear power station in the UK. [The CEGB had met the NCB’s resistance and accom- modated it]. Sizewell B is now supplying electricity to the UK national grid (Ormerod,1997).

The model was not used to guide the NCB’s short-term sales tactics in the international steam coal market as the managers involved pointed out that the crucial short-term dynamic of contractual transactions was not captured by the model. [A resistance was met]. ORE was commissioned to develop a new systems dynamics model. [An accommodation was sought]. This was in accord with the original purpose of the model which was to provide a test bed; the model described here, being based on the economic concept of optimal market clearing, was never going to meet the practical demands of short-term operations.

The world coal model example illustrates the application of the mangle concepts to a technical project and confirms the finding in the earlier paper (Ormerod,2014) that the mangle perspective could be applied to good effect. However, Pickering intended the mangle to apply to all sorts of projects including those where people and their interaction were centre stage, in other words those that had little technical content.

The Sainsbury’s case, to which I will now turn, describes a project of this non-technical type (Ormerod,1996).

4. The Sainsbury’s case The Sainsbury’s project has previously been written up twice:

the first case concentrated on the conduct of the project up to its termination, by which time a 5-year strategy had been agreed (Ormerod,1995); the second case also considered how the strategy was faring after 5 years had elapsed (Ormerod,1996).

The two cases explained the process followed, the methods used and the decision-making process but said little in detail about the behaviour of those involved as the project progressed. What follows reports on some of the crucial events that in practice determined the direction, progress and outcome of the interven- tion. Once again the narrative is informed by the mangle perspective. A complete case study would integrate these events into the overall process described in the original case studies.

4.1. The background The Sainsbury’s IS strategy development project was started in January 1989, it lasted for about 9 months, and was generally considered to be a success. As a consultant working for the PA Consulting Group, I led the project throughout. The story starts in late 1988 when we conceptualized the project, discussed our ideas with the client and won the contract.In the late 1980s, Sainsbury’s was the largest supermarket chain in the UK. Its management had a general aversion to the use of consultants but several consultancies were engaged by their IT/DP department to help systems development. In late 1988, they decided to develop an information systems strategy and invited the incumbent consultancies to bring forward their proposals. PA’s retail expert, who had been looking for an opportunity to build a relationship with Sainsbury’s, got wind of the project. He obtained their permission to submit a proposal even though it seemed likely that the consultancies they were already using would be preferred. [A resistance to our involvement was met]. I would be part of the proposal team. 4.2. A bridgehead is established The PA proposal team met. Sainsbury’s had indicated that they wanted the development to be, in the jargon of the day, ‘user- led’. We agreed between us that if we were to win this assignment, we needed to propose a process that encouraged users to be creative and innovative in their thinking about how to use IT to create competitive advantage. I agreed to design the intervention and lead the project should we win. I had previously helped several companies develop IS strategies, but these projects were neither user-led nor designed to encourage creativity and innovation. One of the main challenges for consultants engaged in strategy development is to ensure that key people in the client organization take ownership of the strategy and commit to implementing it. I needed to find a way of achieving this elusive ownership and commitment. Given the requirement to involve users, and it is users who would ultimately be in charge of implementing the strategy, it became clear that participation was the key; I needed to design aparticipative processthat involved those responsible for, and involved in, operating and developing Sainsbury’s business operations. But this was a difficult challenge that I was not sure I could meet. [A resistance was met, the difficulty of designing a participative process without previous experience; this would be gradually overcome as our proposal was developed (a bridgehead established), the methods to be used chosen and the need for events such as workshops identified (the results of filling)].

In designing a participative process, I decided to draw on my past experience (when at the NCB) of involvement as a participant in two projects that had harnessed the ideas and efforts of others. The first, the Energy Systems Project, was conducted by the International Institute for Applied Systems Analysis (IIASA) (Ha¨ fele,1981; Andereret al,1981; see also Ormerod,2013). The second, the World Coal Study (WOCOL), was conducted by MIT (Wilson,1980). In addition, when working in the NCB, I had project managed an investigation into information systems activities across all departments, conducted by external consultants; the consul- tants had used a participative approach. These three projects 512 Journal of the Operational Research SocietyVol. 68,No. 5 provided me with examples of the design and management of participative processes. As project leader my role would be to orchestrate and facilitate such a process. [A bridgehead had been established based on the concept of participation].

4.3. Participative and IS strategy development methods are used to establish a base model The basic project design (starting with business strategy and ending with a list of systems to be built) drew on my experience of developing IS strategies elsewhere but this time I needed to make the process participative. What methods could I use?

At a meeting, held at our request to help us prepare our proposal, Angus Clark, a Board Member who was the sponsor of the project, emphasized that they wanted a cadre of senior managers drawn from the major departments to be at the centre of the project. He also suggested there should be a steering committee on which he, David Quarmby, Board Member and Joint Managing Director, and others would sit. I raised the question as to whether it was acceptable for me to lead the project given that I had no retail experience. [A resistance was anticipated]. I was relieved when he answered they would actually prefer it, as it would be easier to get me into the Sainsbury’s way of thinking if I came without preconceptions. [The anticipated resistance didn’t material- ize]. In other words, they would provide the retailing knowledge; I would provide my experience of developing strategy. [Two different disciplines would be brought into contact].

We won the contract and the project started in early 1989.

[The initial resistance to our involvement had been overcome].

I was to work closely with the Director of the IT/DP Department, Alan Jacobs (Departmental Directors reported directly to Board Members). In order to form the group that would be at the centre of the project, I requested that senior managers from the main departments be nominated on the basis that I could call on their time for 2 days a week to work on the project. This would allow me to work closely with them. The participants would be selected on the basis of their potential to contribute to the strategy project and take systems development forward in the future; in other words, the idea was that they would get involved in strategy formulation and lead the subsequent systems development and implementation.

The chosen participants would also necessarily be involved in the consequential changes to business processes and organi- zational structure. The issue was discussed at board level but not all agreed to the proposals. [Some resistance was encountered]. The smaller departments had apparently sug- gested that they should also provide participants but the main departments, who did not want their influence watered down, resisted this proposal. [Power was exercised]. The compromise reached was that some of the less important departments would provide participants and each main department wouldprovide additional participants. [An accommodation had been reached]. There would be 16 participants in all, twice as many as I had anticipated.

I returned to the issue of deciding which methods to use to support the investigation. I was drawn to soft OR methods because they were designed to engage participants in the investigation of complex, messy issues (Rosenhead and Mingers,2001); perhaps they could provide the conditions for creativity and innovation to flourish. I decided to incorporate soft OR methods in the project design. [I would transcribe both soft OR methods and my previous experience in IS strategy development into the conceptual space ‘partic- ipation’ to provide a base model; this was a decision, a free choice]. The fact that I now had a group of 16 participants to work with meant that I had to rethink the structure and content of the project. [A mangling occurred].

I decided to break the group into four task forces. Once a month, we would have a one-day workshop in which everyone would be involved. I would structure and facilitate these. The first workshop would be introductory. The second would engage the participants in a variety of exercises such that by the end of the workshop they would (i) be formed into inter- disciplinary (inter-departmental) teams, (ii) have sufficient grasp of the method they would use (soft systems methodol- ogy; see Checkland,1981) and (iii) be sufficiently motivated to carry the project forward in a creative way. Clearly, this second workshop would be a make or break event. At subsequent workshops, the four task forces would report back, new ideas would be introduced and plans would be drawn up for activities over the next month. Later, attention would turn to evaluation; this would need to involve Board Members. 4.4. The participants resist involvement At my request, Alan Jacobs proceeded to write to each participant, setting out the project aims and inviting them to attend a workshop on a given date. [Designing and facilitating workshops when an issue needs to be aired was one of my disciplinary habits as a consultant]. The majority wrote back to say that they were otherwise engaged and nominated a substitute. This was clearly not what I had in mind. I wanted the selected managers to share the experience of the workshop, to hear what each other had to say, to understand their different points of view, and to become a team capable of undertaking the difficult tasks ahead. [My action programme had come into conflict with the individual action programmes of the invitees].

The use of substitutes would undermine the approach I was taking; it needed to be nipped in the bud. I therefore asked Angus Clark [I transferred agency from Jacobs to Clark]to write a second letter emphasizing the importance of their personal involvement in the workshop. [The fact that the letter came from a Board Member was enough to make it clear to the invitees that it was a board-level action programme that they had been invited to participated in]. The response this time Richard Ormerod—Writing practitioner case studies to help behavioural OR researchers ground their theories 513 was dramatically different. All but one of the invitees found that their diaries could be rearranged to allow them to attend. I had, of course, taken advantage of the hierarchical structure of the organization to get the project off on the right footing. [The power of a Board Member was sufficient to whip senior managers into line by issuing no more that a polite invitation; the first resistance to participation had been overcome]. I would not allow a substitute for the one non-attendee and suggested that he should be taken off the project. [I resisted].

However, I was persuaded to give him some one-to-one tuition in order to allow him rejoin the group. [My concerns were accommodated]. He became one of the groups most active and enthusiastic supporters.

The first workshop was held in one of Sainsbury’s meeting rooms in the basement of the head office building; the room was bleak and windowless, and too big for the number of people involved. It was far from ideal but it would be convenient for the participants. Despite misgivings I was persuaded that it was the appropriate choice given that it costs nothing. [Keeping costs down was embedded in the Sainsbury’s retailing culture; their strap line was: Good food costs less at Sainsbury’s]. During the workshop, two interconnected problems emerged. First, to my surprise, many of the participants did not know each other. [I had wrongly thought I would be tapping into well-established networks]. This could easily be put right by getting them to work together in mixed teams at the next workshop. [A deliberate mangling]. Second, when I invited them to discuss business strategy they firmly refused [a second resistance to participation had been met]. They explained that strategy was a matter for Board Members only; their role was to put into operation the strategy handed down from above. In my view, it was necessary for the participants to achieve a clear under- standing of the business strategy. IS strategy should support, or rather, given the strategic potential of IS, be co-determined with, business strategy. They needed to engage with business strategy if they were to develop an IS strategy. However, as I did not have anything in writing to show them there and then, they would not accept my assurances that they had permission to discuss strategy. [Here we see a clash of organizational decision-making cultures. Sainsbury’s was bureaucratic and hierarchical; it was based on discipline, respect for authority and limited (functionally necessary) communication between departments. On the other hand, I was trying to introduce a non- hierarchical approach with the emphasis on communication across organizational boundaries at all levels]. At the end of the workshop, the participants went away sceptical and unenthused. [The second resistance to involvement was not yet accommodated].

At this point, Alan Jacobs wanted to abandon the partici- pative, user-led approach and do something more conven- tional. [A further resistance was encountered]. Nevertheless, I insisted that we proceed with the second workshop. I would try to recover the situation. The second workshop was always going to be difficult, but now the odds of success weresubstantially reduced. [The resistance had been temporarily deferred]. The participants needed to become committed and enthused and, of course, to trust their own capability of tackling strategic as well as operational issues. I had to change things dramatically. I decided to make two crucial changes.

[These were free choices not determined by disciplinary habits]. First, I insisted on changing the location. I chose the ‘Boardroom Suite’ at the top of the Dorchester Hotel in London’s Hyde Park Lane. The location was prestigious, but it would be unfamiliar, less convenient to get to, and very expensive. Second, I insisted that for part of the day, whenever was convenient, David Quarmby and Angus Clark should join the meeting, neither to assert their authority nor to give their explicit permission to discuss strategy, but to join in the activities I had planned: cognitive mapping (Eden,1982)of Sainsbury’s business strategy based on interviews I had held with individual Board Members. I asked each of them to come into the room (preferably at different times to create surprise and drama), to take off their jackets, join one of the groups and get stuck in with whatever they were working on at the time.

Furthermore, I asked them not to make any authoritative statements about strategy even if requested to do so (as media spokesmen they were practiced in responding to questions without answering them). I was demanding behaviour that was totally counter-cultural, but my idea was that it would dramatically make the point that they were there to participate on equal terms. For the participants, I hoped that the experience of the new way of working would be convincing and memorable. [I wanted to give the mangling pot a vigorous stir by making it clear to participants that power had been devolved to me, and via me to them].

The workshop exceeded my expectations. [Jacob’s earlier resistance had now been accommodated]. To reinforce the success, we produced a ‘book’ of the event including all the diagrams and bullet point presentations produced by the participants, as well as anything that I had handed out. From then on, the workshop was referred to as the ‘Dorchester Experience’. [The location had been a material object affecting the action and contributing to the mangling; the book itself became an object in the project network of people, objects and disciplinary cultures]. From this point on, there was never again a problem with enthusiasm and commitment.

[The resistance to participation had finally been overcome].

An unexpected consequence of the workshop was that the two Board Members had thoroughly enjoyed the experience and were now confident that the approach would achieve their aim of developing a new cadre of senior mangers committed to taking the company forward; it had also reinforced their conviction that the culture in the organization needed to be changed. During the remainder of the project, there were, of course, other problems; but, given the willingness of all concerned, these were relatively easily overcome. [A project culture/discipline had been established which ensured contin- ued progress].

514 Journal of the Operational Research SocietyVol. 68,No. 5 4.5. One further mangling is encountered Later in the project, in order to examine the hard and soft impact of the proposed strategy, I had planned a series of evaluation debates between task force members, the steering committee and other board members; discussions were also held with the IT/DP managers to discuss costs and implemen- tation. The idea was to come up with judgemental estimates of the impact and feasibility of the new systems, estimates on which those involved would be happy to base their decisions.

The estimates would include both financial (i.e. costs and benefits) and non-financial elements (for instance, competitive advantage gained and impact on staff morale). Meetings to facilitate these debates took place. However, the steering committee advised that for each proposed system all the evaluation data (including the non-financial) should in the end be expressed in financial terms. This would help the Board authorize the expenditure proposed. [A resistance was met].

The evaluation system was therefore redesigned and re- applied. [The managements’ cultural habit of basing decisions on financial evaluation asserted itself].

4.6. The project appreciated Five years after the project had finished David Quarmby, the Joint Managing Director, commented:

The approach introduced a more open and participative method of decision-making. Old constraints were dis- carded and new ideas were explored creatively…we learnt a lot from the way that the strategy was developed. It was a real education for management. It opened their [the participants’] eyes to new and innovative ways of understanding the business, devel- oping IT solutions and making good decisions (quoted in Ormerod,1996, p. 123).

5. Discussion 5.1. The methodological problems of using case studies for research: the practice turn A fundamental question for case study writers is: what methodology should case study writers adopt? This is a question that social scientist have been asking themselves for many years. Gerring (2004) comments: ‘Practitioners continue to ply their trade but have difficulty articulating what it is that they are doing, methodologically speaking. The case study survives in a curious methodological limbo’ (p. 341).

Methodological issues are clearly of great concern for academic researchers attempting to further their studies by engaging with practitioners in context and using the writing of cases to record their observations and analysis of what they found. But practitioners who write case studies are unlikely to be unduly concerned about such issues.The cases considered in this paper are reports by OR practitioners, that is people involved in interventions con- ducted for commercial, charitable or other non-academic purposes. These case studies are in effect application descrip- tions or witness statements taking the form: (i) we were presented with these circumstances; (ii) an aim was agreed to address/mitigate/solve this problem or issue; (iii) what we observed was…; (iv) this is what we did; (v) our actions gave rise to these problems, so we changed tack…; (vi) we continued tacking until it was agreed the project should be brought to an end; (vii) the project was judged to be successful/unsuccessful by…and (viii) we have learnt….

Academics can potentially make use of these case study reports as texts to be studied, interpreted and incorporated into a theoretical argument. The challenge they face is to make what they can of the implicit methodology and biases of the account written by an OR practitioner. Projects are constituted by people interacting with each other; fundamentally, they are social activities. Thus to address the methodological challenge of interpreting case studies of practice academics turn in the first instance to social theories. Social theories of practice draw on, or have some relation to theoretical stances, from which theories of action in practical and everyday social settings are explored. MacIntyre defines practice as a:

Coherent and complex form of socially established co- operative human activities through which goods internal to that form of activity are realized in the course of trying to achieve those standards of excellence which are appropriate to, and practically definitive of that form of activity, with the results that human powers to achieve excellence and human conceptions of the ends and goods involved, are systematically extended.

(MacIntyre,1981, p.187).

Schatzki explains that the ‘practice approach’ can be demarked as analyses that develop an account of practices and treats the ‘field of practices’ as the place to study the nature and transformation of their subject matter. The practice approach (its ontological stance) takes the social world to consist of embodied, materially interwoven practice centrally organized around shared practical understanding; for instance, actions are embedded in practices, just as individuals are constituted within them, and institutions and structures develop out of practice (Schatzki,2001, p. 3). Practice approaches oppose approaches such as rational choice theory, structuralism, systems theory, semiotics and many strains of humanism and poststructuralism (pp. 1–2). Practice thinkers abandon the traditional conception of reason as an innate mental faculty; rather reason varies between practices, derived from and operating within each practice. An advantage of the practice approach for the BOR researcher is that practices can be observed – it is what people do; on the other hand, beliefs, motivations and norms are internal to individuals and cannot be observed directly (Barnes,2001, p 17). Richard Ormerod—Writing practitioner case studies to help behavioural OR researchers ground their theories 515 A close examination of social practice theory would seem to be fertile ground for BOR research. Good starting points for researchers who wish to develop theories and conduct empirical research into OR practice are provided by Turner (1994), Schatzkiet al(2001), Nicolini (2013) and Sandberg and Tsoukas (2011).

5.2. The suitability of practitioner case studies for BOR research Research into behavioural issues involves the use of quantitative and qualitative methods for the analysis of sociological phenomena (Strauss,1987). As mentioned in the introduction to the paper, the aims and scope of BOR have yet to be extensively debated and agreed but there seem to be two main streams: (i) the development of models of behaviour suitable for inclusion in OR decision-making models and (ii) research into the behaviour of OR scientists engaged in the process of OR.

Presumably the main sources of data for BOR researchers studying the process of OR will be (i) data derived from experiment and observation including practitioners’ reflections on their experiences captured in interviews and surveys, (ii) researchers own experiences of conducting interventions based on the assumption that their experiences are not too different from those of practitioners and (iii) practitioner case studies.

The question addressed here is whether cases in this third category can be written in such a way that BOR researchers interested in the process of OR find them useful.

Practitioners, as a rule, are not motivated to write case studies (Ranyardet al,2015) and OR case studies, which tend to stick to the technical facts, are usually of little help to BOR researchers.

A technical case study normally provides an objectified description of a technical system; however, when a case study focuses on human and conceptual as well as the material aspects, the researcher faces the problem of dealing with structures of meaning that are ambiguous and open to interpretation. Whether case studies (such as the UK energy model in the previous paper on the mangle and the international coal trade model and Sainsbury’s cases presented here) written by practitioners, can ever be more than just ‘story telling’ from a researcher’s point of view, is a moot point; in other words: when we make the step from a scientific case study of a technical application to one that focuses on behavioural aspects, we will be confronted with methodological challenges. Case studies written by ‘soft OR’ scholars could be, on the face of it, more useful; but of course, in that case the intervention is conducted by researchers not practitioners; the aims, expectations and constraints are rather different. BOR researchers have to accept that practitioner case studies will lack a proper epistemological basis; they have to be understood as witness statements, an application description with the potential for bias. But are they nevertheless a source of admissible evidence?

In 1979, a full issue of theAdministrative Science Quarterly was devoted to qualitative methods. It included a damningcritique of qualitative methods by Miles (1979), a well-respected researcher. In response Yin (1981) declared his aim was to ‘reaffirm the role of the case study as a systematic research tool’ (p. 58) and it is his views that have generally prevailed. The case research strategy is according to Benebasatet al(1987,p.369) particularly appropriate for certain types of problems: those in which research and theory are at their earliest, formative stages (Roethlisberger,1977), and sticky, practice-based problems where the experiences of the actors are important and the context of action is critical (Bonoma,1983). In the early formative stages of new ideas and methods, case studies can be used to document the experience of practice, the knowledge accumulated by practitioners engaged in a process of learning by trial-and-error.

Academic researchers can then formalize what has been learnt (Christenson,1976). In summary, Benebasatet al(1987,p.370) suggest that there are three reasons why case study research is a viable information systems research strategy. Reformulated in terms of OR cases written by practitioners, the three reasons are as follows:

(i) The practitioners involved in an intervention learn about the application of OR methods, techniques and models in a natural setting, providing researchers with the possibil- ity of generating theories from practice experience; (ii) The case method allows practitioners to describe ‘how’ and ‘why’ they took decisions in a particular context; this may help researchers understand the nature and com- plexity of the processes taking place.

(iii) A case approach is an appropriate way for practitioners to report on applications in subject areas or contexts where few previous applications have been reported; this helps the researcher develop an understanding of the limits of application, the difficulties met and how they were overcome.

By reflecting on a well-written case study, practitioners and researchers alike may be able to conclude what worked well, what did not and what might work well in future interventions. From a BOR research perspective, case studies of ‘soft OR’ interventions are potentially more fruitful than technical OR cases; in these accounts, process comes to the fore (a number of ‘soft OR’ cases can be found in Rosenhead and Mingers,2001; for ‘soft OR’ cases written by academics and published as journal papers, illustrating a variety of approaches to case writing, see: Ackermannet al, 1997,2014; Brocklesby,2009; Midgleyet al,1998; Winter, 2006; White,2009). But of course, the bulk of OR is technical OR and must be addressed by BOR research if it is to be worthy of the name.

In the related domain of manufacturing and service man- agement, case studies are written and used (see for instance, Meredith,1998; Voss,et al,2002) and others see a role for them in behavioural research (see for instance, Gans and Croson,2008; Zhaoet al,2013). Studying the prospects for behavioural research in operations management, Bendolyet al 516 Journal of the Operational Research SocietyVol. 68,No. 5 (2006) make no mention of case studies, but on the whole researchers are optimistic about the prospects.

We can conclude from this discussion that there is a general appreciation in both social science and domains closely related to OR such as information systems and manufacturing and service management, that case studies provide one viable way of conducting behavioural research.

5.3. The ‘dance of agency’, power and organizational change Does using the use of mangle perspective help us learn more about organizations, how the managers act and interact when involved in an OR interaction? Pickering describes the movement of agency between people, objects and disciplines as a ‘dance of agency’. With agency comes the power to act and control. This can be most easily understood in terms of people. A good example can be found in the Sainsbury’s case when, as I explain, I arranged for Angus Clark to write a second letter after Alan Jacobs’s letter had failed to elicit the hoped for response from the participants (they said they would send substitutes). In this micro-scenario, agency dances around as follows: (i) I ask Jacobs to write a letter to participants; (ii) Jacobs writes the letter; (iii) the participants respond; (iv) I ask Clark to write a second letter; (v) Clark writes the second letter and (vi) the participants respond. Each step attempts to follow the engrained hierarchically institutionalized relationships between Clark (Board Member), Jacobs (Departmental Direc- tor) and participants (senior managers including Departmental Directors) with my additional intervention (an external consultant granted temporary power to act subject to limits determined by the sponsors). The established power structure initially fails to deliver the required result because the participants do not report to Jacobs; they are drawn from other departments (silos) in this inter-departmental initiative.

Once I had overcome the initial resistances and gained the trust of the participants and the sponsors (Angus Clark and David Quarmby, both Board Members) I took over control of agency (within the limit agreed by the Board of 2 days a week of each participant’s time). This enabled me to grant the participants the power to act (within limits): they had to act as a team, document their findings, report back and so on. In this way, we can understand how the temporary organizational structure of the project, based on participation, was embedded in an existing hierarchical organization. This points to the fact that considerations of power are vital in the design and conduct of interventions. How will power be obtained and used to achieve the client’s ends via the intervention? Where can powerful support be found and how can powerful resistance be accommodated?

We can conclude from the example given here that the focus on episodes in which resistances are met and resolved provides insight into the exercise of power in the progress of an OR intervention. Consultants need to be adept at understanding thelocation and use of power in the host organization; sometimes they need to harness it, sometimes to navigate around it and in the last resort it may be necessary to confront it. 5.4. The efficacy of the mangle perspective Pickering uses case studies in his book to demonstrate how the concept of mangling can advance our understanding of scientific work both past and present (Pickering,1995). The previous paper on the mangle approach illustrated how the concept could be usefully applied to a technical OR project (Ormerod,2014). However, the experience of doing so was different from Pickering’s in two respects: first, while Pickering found he didn’t need to use the network concept as proposed in Actor–Network Theory, in the UK energy model case it was found to be essential; second, it was also useful to think in terms of collectivities as actors and more widely to think in terms of actions, interests and preferences stemming from individual and collective action programmes, an approach proposed by Boothroyd in the early days of research into the process of OR (Boothroyd,1978; see also Ormerod,2010).

People engaged in an action programme pursue the interests of that programme as they perceive it and they interpret the actions of others (who are engaged in different programmes) from the perspective of their own programme. Any individual will be involved in many action programmes at the same time (at work, home and in the community) but only one, or perhaps a few, primarily those that relate to work programmes will be relevant to the case study in question. For instance, a line manager responsible for a particular factory (programme one) might also be in charge of a project to improve safety across several factories (programme two); a manager in a local authority may in his or her spare time be involved in a campaign against a development affecting his or her home, family or community.

In the present paper, two cases have been discussed. In the first, the development of a world coal trade model (a case of no particular technical interest), the intermingling of disciplinary approaches (in this case OR and economics) was important.

Networks, collectivities (organizations) and action pro- grammes also proved to be useful in understanding actual events. In the second case, the development of an IS strategy at Sainsbury’s, the same concepts proved useful, reinforcing the previous conclusions. In particular, Pickering’s concept of ‘culture’ illuminated the interaction between the non-hierar- chical approach, which it was proposed to introduce, and the organization’s deeply engrained hierarchical habits; as dis- cussed above, focusing on the ‘dance of agency’ allows the role of power to be followed as the intervention unfolds.

Another conclusion is that applying the template forces the author (practitioner) to concentrate on the link between people’s dispositions and their actions. For instance, when recalling how, at the beginning of the two modelling projects, Richard Ormerod—Writing practitioner case studies to help behavioural OR researchers ground their theories 517 the modelling got going, it was necessary to think about the models that were created in relation to the experience, habits and recent assignments of the modellers. This provided an explanation of how the conceptual bridgeheads were created and the base model established.

Overall, we can conclude that the mangle approach can be usefully applied across a variety of OR projects, hard and soft, technical and non-technical, quantitative and qualitative by focusing on key episodes which affect the direction, progress and outcome of a project. A recent paper by Ha¨ ma¨ la¨ inen and Lahtenan (2016) gives us one example of a behavioural research interest that is being pursued, namely the ‘path dependence’ of OR Modelling; no doubt this behavioural research topic will be pursued further. However, more gener- ally, the future focus of interest of BOR researchers is difficult to anticipate and will undoubted vary greatly from researcher to researcher. Nevertheless, provided with the type of case descriptions which the mangle perspective encourages, BOR researchers should be in a better position to answer questions about the process of OR: what worked well? What did not work well? What lessons has the practitioner learnt? Topics might include, for instance, the choice of consultants and participants with different backgrounds and competences, the organizational arrangements for the involvement of participants, the handling of ethical issues, or the appropriate use of power.

6. Research directions 1. BOR researchers need to develop an understanding of the theories of social practice and explore how they can be applied to OR practice.

2. In time, BOR researchers will need to be questioned as to whether they find that case studies written from the mangle perspective are indeed more useful.

3. There are limitations to the use that can be made of accounts by practitioners working without some involve- ment of an academic. BOR researchers need to seek opportunities to collaborate with practitioners, as some already do, acting as observer, or being actively involved in the intervention, or helping to write the case up.

4. More existing case studies need to be rewritten using the mangle perspective to add weight to the hypothesis that the mangle perspective can be usefully applied in a variety of contexts by a variety of (academic and commercial) practitioners.

5. There is a need to monitor and evaluate published case studies in terms of the way they are written and the use they make implicitly or explicitly of the mangle and other social practice perspectives (Actor–Network Theory for instance).

In addition, PhD supervisors could encourage those students who are proposing to include case studies in their research to consider the mangle perspective. As BOR develops and if confidence in the mangle perspective increases, I would hopeJournal editors and referees will encourage the authors of case studies to consider the contribution they can make to BOR by bringing out certain features such as those brought to the fore by applying the mangle perspective. 7. Conclusions A recent paper (Ormerod,2014) introduced the idea of applying Pickering’s mangle concept to the writing of OR case studies. The aim was to help authors write papers of publishable length while doing justice to both the technical and process aspects of practice. To test the merits of the idea, the mangle approach was applied to a technical OR case.

This paper, in seeking to explore the relevance of the mangle perspective for BOR research, places it in the context of social science practice theory and has further tested it on two projects, the development of the NCB world coal trade model and the development of an information systems strategy at Sainsbury’s supermarkets. The first case confirms that the mangle concept can be usefully applied to technical projects, even ones that are not technically interesting. The second case illustrates how the approach can also be applied to a non-technical, qualitative, ‘soft OR’ cases. The Sainsbury’s case also illustrated the way that Pickering’s focus on the ‘dance of agency’ enabled the important issue of power to be surfaced.

Thus, there is support for the conjecture that the mangle perspective results in a better description of which actors did what and why in both technical and non-technical cases. It does this by identifying and describing key episodes of interaction that determine the direction, progress and outcome of the project. The power of the mangle approach lies in its ability to help the case writer recognize and reflect on underlying causes, connections and dynamics; BOR researchers should as a result be in a better position to address questions about the process of OR. What worked well? What did not work well? What lessons have the practitioners learnt? What lessons can BOR research abstract and feed back to practitioners?

References Ackermann F, Howick S and Quigley JL (2014). Systemic risk elicitation: Using causal maps to engage stakeholders and build a comprehensive view of risks.European Journal of Operational Research238(1): 290–299.

Ackermann F, Eden C and Williams T (1997). Modeling for litigation:

Mixing qualitative and quantitative approaches.Interfaces27(2):

48–65.

Anderer J, McDonald A and Nakicenovic N (1981). Energy in a finite world: Paths to a sustainable future. Cambridge, MA: Ballinger.

Retrieved October 2, 2013, fromhttp://webarchive.iiasa.ac.at/ Admin/PUB/Documents/XB-81-202.pdf.

Barnes B (2001). Practice as collective action. In: Schatzki TR, Cetina KK and Savigny EV (eds).The Practice Turn in Contemporary Theory(pp. 17–28). Abingdon: Routledge. 518 Journal of the Operational Research SocietyVol. 68,No. 5 Becker KH (2016). An outlook on behavioural OR – Three tasks, three pitfalls, one definition.European Journal of Operational Research249(3): 806–815.

Benebasat I, Goldstein DK and Mead M (1987). The case research strategy in studies of information systems.MIS Quaterly11(3):

369–386.

Bendoly A, Donohue K and Schultz KL (2006). Behavior in operations management: Assessing recent findings and revisiting old assumptions.Journal of Operations Management24(6):

737–752.

Bloor D (1976).Knowledge and Social Imagery. London: Routledge.

Bonoma TV (1983). A case study in case research: Marketing implementation. Working Paper 9-585-142, Harvard University Graduate School of Business Aministration, Boston, MA.

Boothroyd H (1978).Articulate Intervention: The Interface of Science, Mathematics and Administration. London: Taylor and Francis.

Brocklesby J (2009). Ethics beyond the model: How social dynamics can interfere with ethical practice in operational research/manage- ment science.Omega37(6): 1073–1082.

Brocklesby J (2016). The what, the why and the how of behavioural operational research – An invitation to potential sceptics.European Journal of Operational Research249(3): 796–805.

Checkland PB (1981). Systems Thinking, Systems Practice. Chich- ester: Wiley.

Christenson C (1976). Proposals for a program of empirical research into the properties of triangles.Decision Sciences7(4): 631–648.

Eden C (1982). Cognitive mapping.European Journal of Operational Research36(1): 1–13.

Franco LA and Ha¨ ma¨ la¨ inen RP (2016). Behavioural operational research: Returning to the roots of the OR profession.European Journal of Operational Research249(3): 791–795.

Gans N and Croson R (2008). Introduction to the special issue on behavioral operations.Manufacturing & Service Operations Man- agement10(4): 563–565.

Gerring J (2004). What is a case study and what is it good for?

American Political Science Review98(2): 341–354.

Ha¨ fele W (1981). Energy in a finite world: A global systems analysis.

Cambridge, MA: Ballinger. Retrieved October 2, 2013, fromhttp:// webarchive.iiasa.ac.at/Admin/PUB/Documents/XB-81-203.pdf.

Ha¨ ma¨ la¨ inen RP, Lahtenan TJ (2016). Path dependence in operational research – How the modeling process can influence the results.

Operations Research Perspectives3: 14–20.

Ha¨ ma¨ la¨ inen RP, Luoma J and Saarinen E (2013). On the importance of behavioural research: The case of understanding and commu- nicating about dynamic systems.European Journal of Operational Research228 (3): 623–634.

Keys P (1998). OR as technology revisited.Journal of the Opera- tional Research Society49(2): 99–108.

Latour B (2005). Reassembling the Social: An Introduction to Actor- Network-Theory. Oxford: Oxford University Press.

Long R (1982). Constraints on International Trade in Coal. IEA Coal Research Economic Assessment Service Report No, GS/83, London.

MacIntyre AC (1981). After Virtue: A Study in Moral Theory.

London: Duckworth.

Meredith J (1998). Building operations management theory through case and field research.Journal of Operations Management16(4):

441–454.

Midgley G (1998). Theory and practice in operational research.

Journal of the Operational Research Society49(11): 1219–1220.

Midgley G, Munlo I and Brown M (1998). The theory and practice of boundary critique: Developing housing services for older people.

Journal of the Operational Research Society49(5): 467–478.Miles MB (1979). Qualitative data as an attractive nuisance: The problem of analysis.Administrative Science Quarterly24(4):

590–601.

Nicolini D (2013). Practice Theory, Work, and Organization: An Introduction. Oxford: Oxford University Press.

Ormerod RJ (1995). Putting soft OR methods to work: Information systems strategy development at Sainsbury’s.Journal of the Operational Research Society46(3): 277–293.

Ormerod RJ (1996). Information systems strategy development at Sainsbury’s supermarkets using ‘‘soft’’ OR.Interfaces26(1): 102–130.

Ormerod RJ (1997). OR models assist the Sizewell B public inquiry - the NCB’s use of linear programming.OR Insight10(3): 2–7.

Ormerod RJ (1998). JORS editorial policy.Journal of the Operational Research Society49(11): 1221–1223.

Ormerod RJ (2005). Putting soft OR methods to work: The case of IS strategy development for the UK Parliament.Journal of the Operational Research Society56(12): 1379–1398.

Ormerod RJ (2010). Articulate intervention revisited.Journal of the Operational Research Society61(7): 1078–1094.

Ormerod RJ (2013). Collaborative research in energy: How the US- USSR initiated a research project 40 years ago.Journal of Research Practice9(1), Article V5.http://jrp.icaap.org/index.php/ jrp/article/view/389/314. Accessed 23 January 2015.

Ormerod RJ (2014). The mangle of OR practice: Towards more informative case studies of ‘technical’ projects.Journal of the Operational Research Society65(8): 1379–1398.

Ormerod RJ and McLeod J (1984). The development and use of the NCB Strategic Model.The Statistician33(1): 35–49.

Pettigrew AM (1985). The Awakening Giant: Continuity and Change in ICI. Oxford: Blackwell.

Pickering A (1995). The Mangle of Practice: Time, Agency, and Science. Chicago, IL: University of Chicago Press.

Plackett MW, Ormerod RJ and Toft F (1982). The National Coal Board strategic model.European Journal of Operational Research 10(4): 351–360.

Ragin CC (1992). Introduction: Cases of ‘‘what is a case?’’ In: Ragin CC and Becker HS (eds).What is a Case? Exploring the Foundations of Social Inquiry(pp. 1–17). Cambridge: Cambridge University Press.

Ranyard JC, Fildes R and Hu TI (2015). Reassessing the scope of OR practice: The influences of problem structuring methods and the analytics movement.European Journal of Operational Research 245(1): 1–13.

Reckwitz A (2002). Towards a theory of social practices: A development in cultural theorizing.European Journal of Social Theory5(2): 243–263.

Roethlisberger FJ (1977).The elusive phenomena. Boston, MA:

Harvard Business School Division of Research.

Rosenhead J and Mingers J (eds) (2001). Rational Analysis for a Problematic World Revisited: Problem Structuring Methods for Complexity, Uncertainty and Conflict. Chichester: Wiley.

Sandberg J and Tsoukas H (2011). Grasping the logic of practice:

Theorizing through practical rationality.The Academy of Manage- ment Review36(2): 338–360.

Schatzki TR (2001). Introduction: Practice theory. In: Schatzki TR, Cetina KK and Savigny EV (eds). The Practice Turn in Contem- porary Theory (pp. 1–14). Abingdon: Routledge.

Schatzki TR, Cetina KK and Savigny EV (eds) (2001). The Practice Turn in Contemporary Theory. Abingdon: Routledge.

Strauss AL (1987). Qualitative Analysis for Social Scientists.

Cambridge: Cambridge University Press.

Turner S (1994). The Social Theory of Practices: Tradition, Tacit Knowledge, and Presuppositions. Cambridge: Polity Press. Richard Ormerod—Writing practitioner case studies to help behavioural OR researchers ground their theories 519 Voss C, Tsikriktsis N and Frohlich M (2002). Case research in operations management.International Journal of Operations & Production Management22(2): 195–219.

Velez-Castiblanco J, Brocklesby J and Midgley G 2016. Boundary games: How teams of OR practitioners explore the boundaries of intervention.European Journal of Operational Research249(3):

968–982.

White L (2009). Understanding problem structuring methods interven- tions.European Journal of Operational Research93(3): 823–833.

White L (2016). Behavioural operational research: Towards a framework for understanding behaviour in OR interventions.

European Journal of Operational Research249(3): 827–841.

Wilson CL (1980). Coal – Bridge to the future: Report of the world coal study. Cambridge, MA: Ballinger.Winter M (2006). Problem structuring in project management: An application of soft systems methodology.Journal of the Opera- tional Research Society57(7): 802–812.

Yin RK (1981). The case study crisis: some answers.Administrative Science Quarterly26(1): 58–65.

Zhao X, Zhao X and Wu Y (2013). Opportunities for research in behavioral operations management.International Journal of Pro- duction Economics142(1): 1–2. Received 19 October 2015; accepted 26 May 2016 520 Journal of the Operational Research SocietyVol. 68,No. 5 Reproduced with permission of copyright owner.

Further reproduction prohibited without permission.