Description
Marketing management is a business discipline which is focused on the practical application of marketing techniques and the management of a firm's marketing resources and activities. Globalization has led firms to market beyond the borders of their home countries, making international marketing highly significant and an integral part of a firm's marketing strategy.
Industrial Marketing Management 39 (2010) 118–128
Contents lists available at ScienceDirect
Industrial Marketing Management
Critical realism in case study research
Geoff Easton ?
Department of Marketing, Lancaster University Management School, Lancaster University, Lancaster, LA1 4YX, United Kingdom
a r t i c l e
i n f o
a b s t r a c t
Case study research is the most popular research method for researchers in industrial marketing. However despite a number of attempts the problem of satisfactorily justifying the use of case research remains. This paper proposes critical realism as a coherent, rigorous and novel philosophical position that not only substantiates case research as a research method but also provides helpful implications for both theoretical development and research process. The article describes the critical realist approach due to Sayer and develops a general application of a critical realist approach to case research. An example of its use in practice is presented using a case study of the development of a buyer–seller relationship after the installation of a new MIS system. © 2009 Elsevier Inc. All rights reserved.
Article history: Received 1 May 2007 Received in revised form 12 June 2008 Accepted 20 June 2008 Available online 24 September 2009 Keywords: Case study research Critical realism MIS Causality
1. Introduction Case study research is, probably, the most popular research method used by industrial marketing researchers. This may be, in part, because of the nature of the subject. The main units of analysis are organisations and relationships, which are dif?cult to access, and complex in structure in comparison with, for example, consumer markets. As a result a case study of a single, or a small number, of such entities can provide a great deal of, largely qualitative, data which can be written up as a case study, offering insights into the nature of the phenomena. But how do we know that what is written as a case analysis represents the “truth”? How can we justly claim that we know, in some fundamental sense, what it is that we have researched? Often the use of the case method is not justi?ed at all in the resulting published work. When it is warranted, the justi?cation is made on the grounds of the interesting and novel nature of the results, exempli?cations of particular phenomena or applications of speci?c concepts or models. Few authors of case based papers offer a defence of their choice of the case method on formal epistemological grounds. Such reticence is hardly surprising. Making truth claims based upon such seemingly limited data is clearly a daunting prospect. This discretion is at least partly due to the dominance of the epistemologically positivistic underpinnings of most academic research in marketing. Clearly the sample size in any case study research project is never going to be large enough to qualify for the use of statistical inference. However a number of writers on case research as a research method seem to take a positivist position without perhaps realising it. For example, Eisenhardt in a widely cited paper offered
? Tel.: +44 1524 592917. E-mail address: [email protected]. 0019-8501/$ – see front matter © 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.indmarman.2008.06.004
the following advice on the number of cases to be used. “Finally, while there is no ideal number of cases, a number between 4 and 10 cases will usually work out well” (Eisenhardt (1989). The justi?cation for this statement is based on her experience with case research and is implicitly about increasing the number of cases as a way of ?nding the same results in each case. However there are a number of reasons why positivism would not be the answer anyway. Its de?ning feature is its nomothetic epistemological stance which implies that there exist regularities or law-like generalisations in material or social settings that provide the basis for both explanation and prediction. This regularity allows positivists to believe that they can make causal statements. If two events occur in sequence regularly then one is said to explain the other. However this simple and elegant formulation has any number of problems which makes its use in any research situation problematic. The most crucial problem is that constant conjunction of elements or variables is not a causal explanation or indeed an explanation of any kind. It is simply an atheoretical statement about the world. It doesn't answer the question why? Sometimes interpretivism, in one of its many forms, is used as a way of defending the use of case studies. Researchers simply interpret cases placing the weight of the research on authentic ways of data capture and sensitive and detailed data analysis. In general interpretivists deny the possibility of knowing what is real and reject the possibility of discerning causality. They can only provide their own interpretation. What is not clear in the interpretivist approach is by what standards one interpretation is judged to be better than another. It is even more problematic when the interpretations are particularistic since this would appear to rule out not just regularity as a criterion but also any form of comparison. In complete contrast the original pragmatists, such as Dewey, James and Pierce argued that it is the uses to which truth is put that
G. Easton / Industrial Marketing Management 39 (2010) 118–128
119
are important. Pragmatism espouses usefulness but only speci?cally and in context. Truth is what is useful to people researching in a ?eld, what helps the research project, what can be accepted and defended, what is open to criticism and renewal. It is a linguistic convention, a sort of shorthand that helps us to achieve our various objectives when researching and theorising. Clearly pragmatism can provide a very powerful justi?cation for the use of case studies since case studies as a research method offers the possibility of studying a problem de?ned situation in great detail. Indeed many action research and illuminative evaluation studies are, in effect, based on single cases of, usually, organisations. However there have been very few papers published in industrial marketing where the data describe problem solving situations and none that use pragmatism as a justi?cation for the validity of the interpretation. For a more extensive discussion of philosophical orientations and case research see Easton (1995), Easton (1998). In this paper I argue that critical realism offers a way forward. Critical realism is, by philosophical standards, a relatively new approach to ontological, epistemological and axiological issues. In a search of the ISI Web of Science database there were 334 papers which included critical realism in the title or abstract. Of those only 42 were in journals that could be said to be in management or organisation studies (and mostly in the latter). Again, of those 42 only 4 papers were in marketing journals and I was author or coauthor of 3 of them. Only 2 of the 334 dealt with case research and critical realism and I was a co-author of both of them. The fundamental tenet of critical realism is that we can use causal language to describe the world. Since all philosophical positions rely on assumptions they can only be ultimately judged pragmatically, not in the limited sense used by pragmatists but in terms of our beliefs that they result in better explanations. One powerful pragmatic argument in favour of critical realism is that it is performative. Critical realists assume that there is a real world out there. However there is no way that such an assumption can ever be proved or disproved, as social constructivists, pragmatists and even positivists are ready to argue. But this assumption is surely performative. In other words we behave as if it was true, as if the world was real. In general this supposition works, especially for the physical world. For example no constructivist would dare to say any longer that the world is totally socially constructed since that is in itself a realist statement. “In both everyday life and social science, we frequently explain things by reference to causal powers” (Sayer, 2000, p.14). Critical realism mirrors the language and procedures we routinely adopt and the explanations that we create. We use causal language without thinking. Critical realists argue for the use of causal language with thinking. Critical realism is particularly well suited as a companion to case research. It justi?es the study of any situation, regardless of the numbers of research units involved, but only if the process involves thoughtful in depth research with the objective of understanding why things are as they are. The paper is structured as follows. It begins with a discussion of the nature of case research and proceeds to a description of critical realism. It then offers an examination of the implications of adopting a critical realist justi?cation of case research and continues with an example of a critical realist case analysis involving the creation of a buyer–seller relationship through the (problematic) implementation of a new Management Information System (MIS). The issue of the generalisability of case research is then discussed, managerial implications are suggested and the paper concludes with a summary of the issues covered. 2. What is case research? This is a very dif?cult question to answer since, in practice, the label is often attached to quite disparate forms of research method. “The term ‘case study’ may refer to several different epistemological entities” (Mitchell (1983) in Verschuren (2003 p.122).
“Indeed the case study is probably best understood as an ideal type rather than a method with hard and fast rules. Yet the fact that the case study is fuzzy round the edges does not mean that it doesn't have distinctive characteristics” (Gerring, 2004 p.346). However sampling mode surely de?nes case research. A case is a single instance; a sample of one. Once the decision to use case research has been made a set of constraints and opportunities are realised. The key constraint is its low (statistical) representativeness. Although it is possible to research several cases this is not done in order to increase the sample size in the conventional sense. The logic of generalisability is totally different for case research and this will be dealt with later in the paper. A single case study must be able to stand on its own. The key opportunity it has to offer is to understand a phenomenon in depth and comprehensively. Research questions are de?nable in terms of the questions; who, what, where, how and why (Yin, 1989, p.18; Yin, 2003. p.5). Case studies are more suited to how and why questions which can be explanatory in nature. “This is because such questions deal with operational links needing to be traced over time, rather than mere frequency or incidence” (Yin, 1989 p.18; Yin, 2003, p.6). Case research allows the researcher the opportunity to tease out and disentangle a complex set of factors and relationships, albeit in one or a small number of instances. This is a process of iterative–parallel research which “…implies a continuous moving back and forth between the diverse stages of the research project” (Verschuren, 2003). The ?exibility that case research allows in this respect is one of its major advantages and one that is not shared by, for example, survey based methods. Case research can therefore be de?ned as a research method that involves investigating one or a small number of social entities or situations about which data are collected using multiple sources of data and developing a holistic description through an iterative research process. 3. Critical realism 3.1. Provenance Critical realism assumes a transcendental realist ontology, an eclectic realist/interpretivist epistemology and a generally emancipatory axiology. While critical realism is a relatively new orientation it is being taken up in many disciplines including economics (Lawson, 1997), sociology (Sayer, 2000; Layder, 1990), criminology (Pawson & Tilley, 1997), geography (Proctor, 1992; Yeung, 1997), linguistics (Nellhaus, 1998), religious studies (Robbins, 1999), history (Steinmetz, 1998), psychiatry (Hanley, 1995), social work (Houston, 2001), ecology (Trosper, 2005), environmental studies (Bania, 1995), law (Hanson & Yosifon, 2004), information studies (Wikgren, 2005 ), media studies (Lau, 2004), interdisciplinary science studies (Dickens, 2003) and management (Ackroyd & Fleetwood, 2004). There are a number of differing views and approaches to realism (Hunt, 2003). This paper espouses the version due to Sayer since his account of critical realist ontology is the most detailed and comprehensive and so makes it easier to demonstrate how it can provide both a philosophical justi?cation for case research and a guide to its use in practice (Sayer, 1992). In addition Hunt recognises Sayer as the key ?gure in the critical realism movement (Hunt, 2000 p286). Therefore in what follows the phrase critical realism means Sayer's version of the ontological position. 3.2. Basic assumptions of critical realism Sayer sets out what he regards as the 8 key assumptions of critical realism in the following extract from his book. 1. “The world exists independently of our knowledge of it. 2. Our knowledge of the world is fallible and theory-laden. Concepts of truth and falsity fail to provide a coherent view of the
120
G. Easton / Industrial Marketing Management 39 (2010) 118–128
3.
4.
5.
6.
7.
8.
relationship between knowledge and its object. Nevertheless knowledge is not immune to empirical check and its effectiveness in informing and explaining successful material practice is not mere accident. Knowledge develops neither wholly continuously, as the steady accumulation of facts within a stable conceptual framework, nor discontinuously, through simultaneous and universal changes in concepts. There is necessity in the world; objects—whether natural or social— necessarily have particular powers or ways of acting and particular susceptibilities. The world is differentiated and strati?ed, consisting not only of events, but objects, including structures, which have powers and liabilities capable of generating events. These structures may be present even where, as in the social world and much of the natural world, they do not generate regular patterns of events. Social phenomena such as actions, texts and institutions are concept dependent. We not only have to explain their production and material effects but to understand, read or interpret what they mean. Although they have to be interpreted by starting from the researcher's own frames of meaning, by and large they exist regardless of researchers' interpretation of them. A quali?ed version of 1 therefore applies to the social world. In view of 4–6, the methods of social science and natural science have both differences and similarities. Science or the production of any kind of knowledge is a social practice. For better or worse (not just worse) the conditions and social relations of the production of knowledge in?uence its content. Knowledge is also largely—though not exclusively— linguistic, and the nature of language and the way we communicate are not incidental to what is known and communicated. Awareness of these relationships is vital in evaluating knowledge. Social science must be critical of its object. In order to be able to explain and understand social phenomena we have to evaluate them critically” (Sayer, 1992, p.5).
fundamental nature and capabilities of the things we research rather than simply their measurable properties. It requires a shift from epistemology and methodology to ontology. 3.4. Causal powers and liabilities Entities have causal powers and liabilities. Causality is, of course, a subtle and disputed concept which Sayer attempts to capture by a process of interpolation using what he describes as an “ordinary” (arguably pragmatic) account of causality. “To ask for the cause of something is to ask ‘what makes it happen’, what ‘produces’, ‘generates’, ‘creates’ or ‘determines’ it, or, more weakly, what ‘enables’ or ‘leads to’ it” (Sayer, 1992, p.104.) Sayer also argues, “…..particular interpretations (of causality) can only be justi?ed in terms of their compatibility with our most reliable beliefs…” Put another way, they rely on an assumption which, together with other assumptions, create a system of thinking about the world that we ?nd acceptable. We must have reason to believe that bodies that we study have powers or liabilities to cause events to occur. They make things happen. Using an example from the case study used in the latter part of the paper, MIS have the powers to change organisations in any number of ways. Similarly organisations have the powers to modify new MIS that are in the process of being implemented or in the ways in which they operate. A liability may be regarded as a susceptibility to the action of other entities, for example particular kinds of organisations may be liable to have particular MIS implementation problems. The bene?t of this conceptualisation is that it focuses attention on three key questions. What are the entities that de?ne our research ?eld, what are their relationships and what are their powers and liabilities? 3.5. Events Events or outcomes are what critical realists investigate, that is the external and visible behaviours of people, systems and things as they occur, or as they have happened. However it is important to acknowledge that most social science research methods create data that are reported rather than directly observed. Descriptions of the events that occur during the implementation of an MIS are rarely experienced at ?rst hand or recorded in a way that is close to the event. Particular attention is paid to processes in critical realist accounts, especially those that produce and reproduce the ordering of events and social institutions. Again MIS are systems that reproduce themselves and are reproduced by other systems in place in organisations. Critical realists also believe that the non occurrence of an event when one is expected not only requires explanation but may also provide very useful insights. Failure of an MIS to provide some of its designed outcomes would be an example of such a non occurrence. Again this is a subtly different approach to that normally adopted in social research. It places the focus, at least in the beginning, on that which we can mostly clearly discover. 3.6. Structure of entities Entities will usually be structured. Structure is “...a set of internally related objects or practices” (Sayer, 1992, p.92). For example an organisation may be considered to comprise a series of other entities (departments, people, processes, resources) all of which can affect one another. Structures are nested within structures. For example entities can be organisations that have departmental structures and relations and, within them, individuals who have particular characteristics such as gender and psychological structures. Gender forms part of the internal relations of a person if gender is to be regarded as a necessary part of the structure that is being built.
Points 1, 4 and 5 set out the key critical realist assumptions about ontology. Critical realists propose an ontology that assumes that there exists a reality “out there” independent of observers. A naïve realist epistemology would assume that this reality can be readily accessed. This is a view often espoused by researchers in the natural sciences because of their ability to measure accurately and their access to controllable and / or closed systems. However these conditions rarely occur in social systems. As a result critical realists accept that reality is socially constructed and points 2, 3, 6 and 7 spell out this proposition thus creating a tension between these apparently contradictory views. However critical realists resolve the tension by arguing that the world is socially constructed but not entirely so. The “real” world breaks through and sometime destroys the complex stories that we create in order to understand and explain the situations we research. 3.3. Objects/entities Objects, or more generally entities, provide the basic theoretical building blocks for critical realist explanation and can be such things as organisations, people, relationships, attitudes, resources, Management Information Systems (MIS), inventions, ideas and so on. They can be human, social or material, complex or simple, structured or unstructured. Entities stand in contrast to the idea of variables that dominates most social research traditions. Variables are measures of things and not the things themselves. “Similarly, the concept of variable that is used in quantitative analysis is an indifferent one as regards causal explanation: variables can only register (quanti?able) change, not its cause” (Sayer, 1992, p.180). This modi?cation radically alters the way that we need to think about theory. It directs our attention to the
G. Easton / Industrial Marketing Management 39 (2010) 118–128
121
This conceptualisation suggests that there are two alternative ways in which MIS can be treated theoretically. The ?rst is to include MIS as a part of the structure of an organisation. The implication would then be that the study of an MIS might consider it to be primarily concerned with organisational behaviour as it relates to one particular part of its structure having interesting and important effects on its behaviour. A second approach would be to see the relationship between an organisation and an MIS as external so that one does not encompass the other but is related to it. This might be a way to deal with changes in MIS. 3.7. Emergence In the language of critical realism, entities may be analysed at a number of different levels of aggregation. A crucial critical realist assumption concerns the existence of emergence in such situations. The properties of entities at a higher level of aggregation are not necessarily understood through a summative process or, working from the top down, a reductionist approach. They emerge from those of the lower level but are not easily derived from them. For example MIS systems have emergent properties that are more than, and different from, the sum of their constituent parts. Similarly entities at a higher level such as organisations cannot simply be reduced to the summation of their components such as MIS. The implication is that in choosing a level of analysis one accepts that ready access to other levels is not necessarily easy. The current inability of biochemists and physiologists to ?nd a biochemical model of consciousness provides a case in point. Emergence must always involve some element of connectedness. For example biology could be regarded as an emergent from chemistry in the sense that it is the combinations and connections among atoms and molecules that create and sustain biological processes. Similarly the social world is only understood through the connections between the people that comprise a society not by studying the individuals in isolation. Closer to home the properties of organisations stem, in part, from the connections among the individuals and groups they contain. 3.8. Necessary relations Critical realists argue that there are two kinds of relationships among entities; necessary and contingent. As Sayer writes “…the relation between a slave and a master is necessary, in that what the object is dependent on its relation to the other; a person cannot be a slave without a master and vice versa” (Sayer. 1992, p.89). In terms of MIS, for example, organisations and MIS have a necessary causal relation since one cannot exist without the other. Organisations are managed and involve information systems, though not necessarily ones that are electronically based. MIS must exist within organisations. Necessary relations are not tautologies but derive directly from the nature of the bodies involved. The relation between entities and the events they cause will usually be a rich and varied one. Thus the elements of the relation are not simply given by the mutual de?nition. MIS and organisations have many ways of relating to and affecting one another. Entities are, however, de?ned in terms of their necessary relations. A person must be able to use or ignore an MIS. For complex entities there may be any number of relations that de?ne both it and the other entities to which it is related. Necessary relations are not inevitably rigid though changes in one entity will, of necessity, lead to changes in the other. Necessary relations do not have to be particularly important since this depends on the nature and objectives of the research being carried out. Of course that does not mean they do not exist. None of these ideas are likely to be particularly surprising. Research builds up theories that comprise a number of concepts
that refer to particular entities and specify the relations among them to create a theory or theoretical framework. But it is clear that in doing so we are relying on the relations among entities to hold the whole edi?ce together. In other words the de?nitions become referential and interdependent. Change in one body leads to change in another body with which it has necessary relations. That does not mean that there have to be regular changes though there may be. The change may also lead to the body changing its nature (e.g. from a national to a multinational corporation). Internal relations do not have to be symmetric. One body may be not be able to exist without the other, for example an organisation cannot exist without individuals but individuals can exist without, at least, formal organisations. While bodies may de?ne relations it is equally likely that there will also be an element of mutual de?nition. A Chief Information Of?cer will be de?ned in terms of the necessary relations he or she has with other board members, other institutions, certain employees, etc. We rely on these referential and interdependent relations to underpin our theoretical understandings. 3.9. Contingent relations A contingent relation occurs when “It is neither necessary nor impossible that they stand in any particular relation” (Sayer, 1992, p.89). Put at its simplest this distinction recognises that entities can have some relations (necessary) that will affect one another and some (contingent) that may affect one another. The implementation of a new MIS system may be affected by the home country of the system supplier or it may not. There is no necessity because the relation is a contingent one. However contingent relations are different from necessary relations only in the nature of those relations. “……the contingently related conditions are never inert, but are themselves the product of causal processes and have their own causal powers and liabilities” (Sayer, 1992, p.140). The theoretical framework chosen governs the difference between necessary and contingent. All events need to be explained by a combination of necessary and contingent relations. If all relations were contingent then each explanation would be unique and incapable of contributing towards anything by way of generalisation. Sayer puts it this way. “…Structures can therefore be said to be ‘invariant under certain transformations’, that is, they can continue to exist while their constituents undergo changes in attributes which are not relevant to their reproduction” (Sayer, 1992, p.94). This is crucial because it recognises that there will and indeed must be invariance in the system. Building theory and progress then become possible. 3.10. Context Distinguishing between context and contingency is possible. The former offers a simpler, less well-articulated version of the latter. Context is simply “relevant circumstances”. It is a very general concept and says little about the relationship between the focal entities and the environment except that it is (possibly) relevant. In the case of critical realism the entity should not only be de?ned but the form of the causal relationship clearly set out. The question should be “in what ways may the external contingency affect the events that have occurred?” 3.11. The structure of causal explanation The most fundamental aim of critical realism is explanation; answers to the question “what caused those events to happen?” In Fig. 1 causal explanation is structured in terms of the relationships among the concepts that have so far been discussed. This is a formal statement of the critical realist structure of explanation. Less formally a very simple example demonstrates the
122
G. Easton / Industrial Marketing Management 39 (2010) 118–128
Fig. 1. The structure of casual explanation (Sayer, 1984).
most basic form that such a formal explanation can take. Objects (a salesperson) having structures (knowledge and personality traits etc) and necessarily possessing causal powers (to persuade a buyer, who is another object) and liabilities (to be rejected by technical buyers, to get tired towards the end of the day) will, under speci?c condition c1 (the buyer has a need for the product and the offering is suitable) result in an event e1 (a sale), or alternatively under speci?c condition c2 (the buyer has a need for the product but the offering is not suitable) will result in an event e2 (no sale). In practice such formal explanations will not normally be possible because of the complexity of real world behaviour but they do provide a logical framework to guide case researchers. However critical realists argue that there should always be competing explanations since different interpretations of the data are necessary to ensure that the “best” current interpretation is made. But there should always be room for revision through the normal processes of academic presentation and critique. 3.12. Mechanisms Mechanisms are “nothing other than the ways of acting of things” (Bhaskar, 1978. p. 14). A causal explanation is one that identi?es entities and the mechanisms that connect them and combine to cause events to occur. Other writers (Tsoukas, 1989; Keat & Urry, 1975; Layder, 1990) use somewhat different terms such as “deep structures”, “deep processes” or “generative mechanisms”. The term mechanism has problematic connotations since it implies clear structure and invariance in operation, something that critical realists would reject. A better portmanteau term would be deep generative processes and structures. However mechanism has become embedded in the language community and so will be used hereafter but with the caution that it should not be taken literally. Mechanisms are at the heart of causal explanation but also generate a constant source of debate. Perhaps the simplest way of regarding mechanisms is that they are ways in which structured entities by means of their powers and liabilities act and cause particular events. “When activated, particular mechanisms produce effects in “conjunctures”, which may be unique. According to conditions, the same mechanism may sometimes produce different events, and conversely the same type of event may have different causes” (Sayer, 1992, p.116).
A simple example of a mechanism would be a log jam or tipping point. When a buyer and seller have agreed on almost everything (structure and conditions) there is always the possibility of a deal breaker emerging. The crucial deal breaker condition then solely determines whether the event will take place. Recognising that the deal breaker is actually a deal breaker rather than just a part of the players' negotiation strategy then becomes crucial. An important aspect of mechanisms in the critical realist tradition is that they offer a rich source of explanatory devices. They do not need to be linear additive as required by statistical models or logicorational as in box and arrow diagrams. Instead they can be linguistic in nature and metaphorical (Easton & Araujo, 1993). A brief survey of research in the social sciences produced this sample of source metaphors in use; marriage, garbage can, contagion, room and key, photography, war, dance, fortress, jazz, life cycle, tourist, theatre, cancer and computer. However, “A generative metaphor is not merely an ornate expression of similarities and analogies its author was already aware of, but is the source of new perceptions of similarity and analogy, picking out similarities and analogies that were unknown until the metaphor pointed them out and thereby brought them to the author's attention” (Lewis, 1996 p.493). At a more formal level, various models or less well speci?ed theories such chaos, catastrophe and complexity theory could be adopted. For example Kendall and Kendall (1993) elicit 9 metaphors that were used by information systems users and in systems development methodologies offering the potential to use one or more of them in developing explanations for MIS events.
3.13. Epistemology Critical realists accept that our world is, of course, socially constructed but argue that this is not entirely the case. They construe rather than construct the world. Reality kicks in at some point. “Critical realism acknowledges that social phenomena are intrinsically meaningful, and hence that meaning is not only externally descriptive of them but constitutive of them (though of course there are usually material constituents too). Meaning has to be understood, it cannot be measured or counted, and hence there is always an
G. Easton / Industrial Marketing Management 39 (2010) 118–128
123
interpretative or hermeneutic element in social science” (Sayer, 2000, p.17). The difference between critical realists and social constructionists lies in the acceptance of the possibility of knowing reality in the former case and its rejection in the latter who, in general, concentrate instead on uncovering the constructions that social actors make. Bhaskar (1978) has suggested that critical realism has a strati?ed rather than ?at ontology and this has major epistemological implications. The strata are the empirical, the actual and the real. The empirical domain is where observations are made and experienced by observers. However events occur in the actual domain and may be not observed at all or may be understood quite differently by observers. There is a process of interpretation that intervenes between the two domains. Events occur as a result of mechanisms that operate in the real domain. It is not the case that the real or actual cannot be observed but simply that it may not always be capable of being observed. We see just the tip of an iceberg but that doesn't mean that the invisible three-quarters is not there or is unconnected to what we see. Observation is fallible. It is unlikely to reveal completely and lead to a full understanding of any social situation. Since there can be no de?nitive criteria to judge the “truth” of a particular version, critical realism relies on the researcher to collect further data that helps to distinguish among alternative explanations and on the community of researchers to debate them thoroughly. As a result criticality within a discipline becomes essential since only by seeing the same data through the different theoretical lenses employed by different researchers can understanding of some of the features of the real world occur (Woodside & Wilson, 2003; Woodside, Pattinson, & Miller, 2005). 3.14. Research process Critical realists further distinguish themselves from other schools of thought in terms of the way that they believe that the process of research should be conducted. Pierce, ultimately a pragmatist, ?rst suggested in 1911 what the alternatives might be; induction, deduction and retroduction. Retroduction is a “…mode of inference in which events are explained by postulating (and identifying) mechanisms which are capable of producing them…” (Sayer, 1992, p.107). Put another way “...there is a concern in realist accounts, with a different form of inference to the more common induction and deduction. Whereas the latter are concerned with movements at the level of events from the particular to the general and vice versa, retroduction involves moving from a conception of some phenomenon of interest to a conception of a different kind of thing (power, mechanism) that could have generated the given phenomenon” (Lawson, 1997, p.236). Retroduction means “moving backwards” and that is what the process involves. It asks “What must be true in order to make this event possible?” Abduction is a closely related process which has already been suggested as appropriate for case research (Dubois & Gadde, 2002). 4. Critical realism and case research Sayer argues that critical realism is relatively tolerant with respect to different research methods. “Compared to positivism and interpretivism, critical realism endorses or is compatible with a relatively wide range of research methods, but it implies that the particular choices should depend on the nature of the object of study and what one wants to learn about it” (Sayer, 2000, p.19) (Fig. 2). He suggests that there are two broad types of research method; extensive and intensive. The former employs large scale surveys, formal questionnaires and statistical analyses, looks for regularities, patterns and similarities, accepts given taxonomic categories, privi-
Fig. 2. Critical realist view of causation (Sayer, 2000, p15).
leges replication and has restricted ability to generalise to other populations and limited explanatory power. The latter focuses on individual agents in context using interviews, ethnography and qualitative analysis, asks the question “what produces change?”, employs causal groups, produces causal explanations which are, however, limited to the situation studied so that testing is by corroboration. However he points out that, “the extensive/intensive distinction is not identical to the survey-analysis / case study or ethnography distinction; extensive methods might be used within a single case study; intensive approaches do not have to be limited to single cases and can use other methods besides ethnographic ones.” These descriptions af?rm that case research is an intensive research method according to Sayer's description and is therefore entirely consistent with a critical realist ontology. However it is important to consider how this might work in practice.
4.1. Critical realist case method The ?rst task in any research project is to decide on the phenomenon to be studied. A critical realist case approach is particularly well suited to relatively clearly bounded, but complex, phenomena such as organisations, interorganisational relationships or nets of connected organisations. It is less well suited, for example, to the study of individual behaviour, although that can be done, or to situations which are largely characterised in quantitative terms such as perceptions and attitudes tracking data. The boundaries of the phenomena, for example organisation or net, must be determined although it is not uncommon for those boundaries to be changed during the course of the research. For example, determining causality may require that the researcher moves beyond the initial boundary or, alternatively, narrows the boundary because it is clear that the causal mechanisms are more narrowly focused than previously thought. The next issue is the nature of the research question. Here critical realism has a rather clear answer. The question must be of the form “What caused the events associated with the phenomenon to occur”. As suggested previously, it is only possible to understand social phenomenon by recording and analysing the associated events that take place as a result of the actors acting, whether they are human or non human (e.g. machines such as computers). The events can be recorded live or exist in records of the past including the memories of those human actors who can attest to the events. The next and crucial task is that of identifying the entities/objects that characterise the phenomena being studied. Sayer points out that this process is usually given less attention than it requires. “Much rests on the nature of our abstractions, that is, our conceptions of particular one-sided components of the concrete object; if they divide what is indivisible, or if they con?ate what are different and separable components, then problems are likely to result” (Sayer, 2000, p.19). For example would it be sensible to consider a B2B relationship as a single entity or as a separate feature of each of the organisations involved? The Actors–Resources elements of the Actors–Resources– Activities model provide a good example of abstractions that have proved to be useful in B2B research. Activities, of course, correspond to events that involve them both. Moreover the addition of bonds,
124
G. Easton / Industrial Marketing Management 39 (2010) 118–128
links and ties as ways in which each has its own structures and relationships, integrates but does not con?ate separable components. Of course initial conceptualisations may change as the research progresses and clarifying what they are may be a key aspect of any research project. However, if theory is to progress then the basic entities and their powers and liabilities need to powerful enough in order to have some continuity of existence in terms of extant theory. The next step is collection of data. Case research is essentially eclectic with respect to the kinds of data that might be collected. Too often case research is equated with qualitative data collected by semi structured interviews. The strength of this method is that it is highly ?exible. However other forms of data collection, for example experimentation, can also work well in particular situations and provide insights not obvious in the more traditional modes of research. In any case the choice will be governed by what is thought to be required to establish a plausible causal mechanism, constrained by what data can actually be collected in the research context. Having collected the data the issue of its interpretation arises. Critical realists accept that there are differences between the empirical, the actual and the real, and that data are collected from people as well as from, and about, material things. As a result they accept that any explanations are necessarily fundamentally interpretivist in character. In particular when analysing respondent based data the researcher faces the problem of the double hermeneutic (Woodside et al., 2005). This adds another layer of complexity to the process since researchers are then required to include our understanding of the subjects' understandings. But while discourse is important and its interpretation crucial, reference to the referents of the discourse need to be made. In other words critical realists do not accept that studying the content and ways of expressing talk is enough in itself to provide explanation except of a very speci?c linguistic form. Seeking an explanation requires that the researcher goes back to the research site collecting more data until epistemological closure, however ?awed and temporary, is obtained. Since retroduction is the key epistemological process that critical realists recognise, understanding how it might work is important. Retroduction is a metaprocess the outcome of which is the identi?cation of mechanisms that explain what caused particular events to occur. Its adoption does not imply that the mechanisms are postulated then data collected or that they are “induced” from the event data. In practice the process is likely to be an iterative one (Dubois & Gadde, 2002). Case studies may employ both deductive and inductive cycles of data collection. Deduction helps to identify the phenomenon of interest, suggests what mechanism may be at play and provide links with previous research and literature. Induction provides event data to be explained and tests the explanations. Finally explanations invoke causal language and the identi?cation of mechanisms and offer the data collected as evidence. It differs from other research process cycles only in that its goals are different. The cutting edge of this method is to continue to ask the question why? Sayer points out that it is easy to make causal misattributions given the complexities of the systems we study and the possibility that different mechanisms can cause the same events. In addition there may be one or more mechanisms at work and the research design must be adapted to address both of these issues. The case method with its “cut and come again” disposition is ideally suited to the task. That a number of different causal explanations are put forward and researched is also not only possible but pragmatically desirable. Pawson and Tilley (1997), for example, suggest 8 different mechanisms that might account for the ?nding that placing CCTV cameras in car parks reduces car thefts and damage. Deciding which of these mechanisms appears to be working in the particular case then becomes necessary. Finally there is the issue of deciding whether your explanation is “good” or not. Clearly we would all hope that it is, and if not good then
simply the best that we can do in the circumstances. Another way of making this decision is by interpreting “good” as “acceptable”. For some critical realists the answer lies in the concept of “judgemental rationality”. “Judgemental rationality means that we can publically discuss our claims about reality as we think it is, and marshal better or worse arguments on behalf of those claims. By comparatively evaluating existing arguments, we can arrive at reasoned, though provisional, judgements about what reality is objectively like; about what belongs to that reality and what does not. ” (Archer, Collier, & Porpora, 2004, p2). 5. An example of a critical realist industrial marketing case analysis In the section that follows a necessarily truncated analysis of a case situation involving the purchase and implementation of a CRM system is described. 5.1. Black box case study Black Box Ltd is a small UK manufacturer and retailer of an extensive line of top end consumer electrical and electronics products. Until the late 90s they catered to the “technically competent” segment of the market but when it was realised that this was no longer viable they changed strategy and targeted the larger, less expert “plug and play” segment. At that time there was a small amount of ad hoc advertising, a once a year customer mailing was sent out and the company marketing information systems were mainly paper based. To implement their new strategy they extending their product line, redesigned their mail order catalogue, created a static website and began direct marketing. As a result from 2000 to 2003 sales grew steadily and Black Box struggled to keep up with the extra work that was involved. They invested in extra staff and more space assuming that sales would continue to rise. However, contrary to expectations, during 2004, sales plateaued and have remained steady since. In the mid 1990s, the main MIS was a standard PC accounting package which did the invoicing; everything else was on paper. In the late 90s they acquired their ?rst computer based mail order system, Mailstrom, and a transactional website to replace their simple brochure website. Then in early 2005 Black Box decided that they needed to purchase a CRM system. Their reasons for the purchase were that the current system was unreliable, slow in terms of customer response and didn't give them the customer information which would allow them to target customers more precisely. In addition there was the option to replace labour by capital and so reduce staff overheads. A decision was made to buy the CREOLE CRM system on the basis of its functionality and price offering rather than on its technical attributes. CREOLE allows Black Box to target customers and design particular promotions for particular market segments by means of promotional emails (e-blasts) and a twice monthly soft sell e-newsletter with new product information, tips and advice. In addition ordering on line is now possible. However the implementation of the CREOLE system was not without problems, which appeared to be of two kinds. The ?rst were software problems, for example two parts of the programme were supposed to provide the same information but didn't. The second and more important general problem involved the level and quality of service support from CREOLE. A key reason why Black Box had chosen CREOLE was that they had a good reputation for training and after sales service. This perception was probably due to CREOLE's policy of offering just enough training to get a client started and then providing tailored manuals to allow them to train themselves afterwards. In effect they were substituting cheaper tailored manuals for more expensive upfront training. In practice the training turned out to be largely perfunctory. The initial training lasted only a few hours with hand written notes as the sole record. In one case the onsite training was called off because of a
G. Easton / Industrial Marketing Management 39 (2010) 118–128
125
system failure and was never completed. Worse still the tailor made manuals arrived months after the system went live. Black Box staff wondered whether it would have been better to have a general manual to start with until the customised version was available. Even when the manuals turned up they didn't necessarily solve all the problems since they were incomplete. Another service option was CREOLE's help desk but this didn't always operate satisfactorily. Sometime after the installation, and clearly under pressure, CREOLE requested customers to email their help requests which would be acknowledged within 4 hours, prioritised and dealt with appropriately. Unfortunately the responses were often very late; the longest wait was 7 months. Black Box also acknowledged that there were problems with their personnel being unable or unwilling to use the system properly or to ?ag up its faults. In particular, the person responsible for co-ordinating the project didn't perform well. She continued to try to work in parallel with the old system and tried to solve all the problems herself. As a result other staff stopped complaining about speci?c issues and simply grumbled about the system and developed a generally negative attitude towards the change. The person concerned ultimately left the company. At this point the staff were told to report speci?c problems to their managers for forwarding to CREOLE. During this period CREOLE were involved in a UK government funded Knowledge Transfer Programme which was designed to help them improve their marketing performance. One of the elements of the programme was the institution of a Key Customer Forum to which senior staff from Black Box was invited. The forum provided an opportunity for customer staff to raise problems and issues and for CREOLE to listen and explain their situation and policies. A particular request for more self help information was made at the ?rst forum but this CREOLE seemed reluctant to offer. They argued that much of what they were being asked to provide was, in effect, internal systems redesign which was not what they were contracted to do. However more speci?c help was forthcoming when the Black Box staff met their CREOLE account manager, who they hadn't known of before. She was given a list of problems and questions, visited Black Box and successfully sorted out the former and answered the latter. Many of these queries were straightforward and could have been dealt with by the help desk. At this point CREOLE confessed to their failures, an admission that Black Box appreciated, and this served to foster a better relationship between the two companies. However it is clear that Black Box changed as a company as a result of the acquisition of the CREOLE system and staff now recognise how vulnerable that might make them due to their lack of IS expertise. “I think we all probably have a bit of a love hate relationship with computers.” This attitude has not preventing them from planning to get the most out of the CREOLE system and to extend and modify it to provide even better customer service in the future. 6. Case analysis 6.1. Research question The research topic is the creation and development of a business service relationship. One of the key research questions could be; What caused the CREOLE service problems as perceived and/or experienced by Black Box and what caused their subsequent solution? 6.2. Events The case describes a series of service interaction events that occurred between an MIS supplier and customer and the internal events within both parties that were linked to those interactions. The key interaction events/sequences involved training, delivery of the tailored manuals, help desk exchanges, the forum meeting between
Black Box staff and their CREOLE key account manager and her subsequent visit. The key internal events within Black Box were the attempts (not always unsuccessful) to make the CREOLE system do what it was supposed to do and the activities of their erstwhile project coordinator. The key internal events within CREOLE were the creation of the tailored manuals, the operations of the help desk, the management and enactment of the tailored manual policy and the appointment of the account manager. These events are partly sequential in nature (forum meeting then visit of the CREOLE account manager) and partly cyclical and reproductive (accessing the help desk). In the former case it is tempting to suggest that events cause events. However a critical realist account would be that as a result of an event the entities involved, and their powers and liabilities changed and this, in turn, caused the new event. 6.3. Identi?cation of entities Entities only attain meaning if they are embedded in a theoretical framework. Of course in exploratory research entities may be chosen then accepted, modi?ed or discarded as useful ways of try to understand data. However in this case a previously developed theoretical framework can, for the sake of an example, be employed; an economic exchange model (Easton, 2002). The entities involved are seller, buyer and exchange entities. There are necessary relationships among them all. An economic exchange cannot occur without a seller and a buyer and something that is exchanged. A seller has the power to sell and the buyer to buy and what is exchanged has power to affect both the buyer and seller. In this particular case the primary exchange events have already taken place but the extended exchange continues as after sales service. The service element means that what are exchanged are partly joint activities undertaken by the actors (calling the help desk) and partly physical entities (manuals). In this case the exchanges have continued for some time and an exchange relationship can be said to have come into existence. This relationship then becomes, of itself, an entity which has powers and liabilities which interact with those of both CREOLE and Black Box. For example the current atmosphere of the relationship is created by either or both and affects the relationship of each of them. The fourth entity is the Knowledge Transfer Programme (KTP) which had the power to in?uence the actions of CREOLE, not least in creating an event which had not only a major impact on the company but also in its relationship with Black Box. The relationship between the KTP and CREOLE is a contingent one since there is nothing necessary about its relationship to the buyer–seller relationships. The latter can exist without the former. However in this case the contingent powers of the KTP i.e. to bring into close, personal proximity CREOLE and Black Box personnel were activated to great effect. All three entities have important internal structures that affect events or, more properly, form part of a mechanism that cause those events to occur. For Black Box, internal structures include the CREOLE system as well as the other systems and procedures and the culture and personal characteristics of its employees. For CREOLE, the internal structures include the CREOLE CRM system, the help line resources, the number and expertise of its employees, its reputation for good service and the policy of tailored manuals and little training. The internal structure of the relationship includes the mutual expectations of the two actors, its service elements, the history of previous joint events, adaptations that each company made to each other and its atmosphere. Entities can be tangible or intangible, social or physical, dormant or active. The physical entities include, for example, the manuals and the output of the CREOLE code and data. The intangible entities include the reputation of CREOLE and the atmosphere of the relationship. In the case of the manuals this turns out to be a crucial aspect of the exchange since their form, either as general or tailor made manuals, had strong causal powers to affect events.
126
G. Easton / Industrial Marketing Management 39 (2010) 118–128
6.4. Possible causal mechanisms Three interconnected mechanisms can be described as being at work in this case. The one that is primary is the “perceived balance of responsibility” mechanism. There is always likely to be complex interaction between the nature and characteristics of a MIS system and the internal structures and processes of the recipient organisation. If both were simple and easily understood it might be possible to write a contract specifying exactly what the responsibilities of both parties would be. However this is rarely the case so that there is always likely to be uncertainty and ambiguity about who is responsible for what parts of the implementation. However each actor made assumptions about what responsibilities the other had in making the CREOLE system work as well as what the criteria for working actually were. Black Box had high expectations that CREOLE would be largely responsible for the whole process. This was in turn caused by their perceptions of CREOLE's reputation and their own inexperience in MIS implementation. On the other hand CREOLE was attempting to enact a strategy that systematised their service through the use of tailored manuals which they hoped would allow Black Box to solve more of their own implementation problems. This strategy was most probably caused by a lack of service resources at the time but was also an attempt to provide an acceptable mid point in terms of balance of responsibility for both parties to the exchange. The second mechanism was the “enacted balance of responsibility”. CREOLE completed neither the basic level of training nor the delivery of the training manuals at the required time. In addition the help line was unable to ?ll the resulting gap, again most likely as a result of inadequate resources and using a response system designed to minimise the use of those resources. And the software had some remaining bugs. Black Box, on its part, had a project coordinator who was trying to continue use of existing systems and rejecting help, uncoordinated activities in terms of managerial control of the process and naïve and unrealistic expectations. The timely resolution of the problems was, in part, a chance contingent event. If CREOLE and Black Box personnel had not met at the forum then any number of possible scenarios could have been envisaged. The crucial point about these events was that they suggest a third related “key in lock” mechanism. In practice the help required was relatively trivial in nature, solved many of the remaining problems and restored the relationship between CREOLE and Black Box. The key in lock metaphor describes the situation where a particular con?guration of information allows a door to be “opened” and a system to operate as expected. This metaphor depends on the knowledge that by the time the forum happened, Black Box had changed their coordinator and taken steps to control the implementation and so were ready to go. What remained were technical issues that CREOLE could show them how to solve relatively easily. Fig. 3 below gives a partial summary of one of the sets of events which took place and which are described in the case. It gives examples of some
of the powers, liabilities and structures of the entities which operate through one of the proposed mechanisms to create another rather different set of events than those experienced previously. It suggests an answer to the question “Why did the exchanges between CREOLE and Black Box become more helpful and productive after a particular point in time?” More complex explanations of the previous situations could also be constructed using the concept of perceived and actual balance of responsibility as key components. The returns from carrying out the above analyses are therefore three fold. Firstly they have, hopefully, provided a means of improving our conceptualisation of some of the key entities in B2B research, buyer, seller, exchange entities, in terms of how they relate to one another. Secondly they suggest three possible mechanisms that might occur in other buyer–seller situations. Thirdly they provide an example of the way in which a critical realist approach might be employed when analysing case studies. 7. Theory development and generalisability The knowledge claims of case study research are often attacked on the grounds of lack of generalisability. Most cases reported are single cases. The argument runs thus: we can accept that you have made a very convincing job of explaining the rather complex situation you have been researching, but how do we know your case situation was representative?. How can you generalise? Yin, in the ?rst edition of his book on case research, suggests, “The short answer is that case studies, like experiments, are generalisable to theoretical propositions and not to populations or universes. In this sense, the case study, like the experiment, does not represent a ‘sample’, and the investigator's goal is to expand and generalise theories (analytical generalization) and not to enumerate frequencies (statistical generalization)” (Yin, 1989, p.21). How this analysis should take place or in what sense generalisation is possible, Yin did not make clear. From a critical realist perspective, expansion and generalisation come from identifying the deep processes at work under contingent conditions via particular mechanisms. A causal explanation in a single case must be based upon a theory structured in terms of what comprises a critical realist causal explanation. The best explanation, that is the one most consistent with the data, is what is being sought. However generalisation of any kind is not possible unless there is some invariance in the world. If all events and their causes are unique then there could never be theories that work. And by unique I mean substantially and not trivially unique since in some sense every event in the world is unique. Clearly it is not possible to explain every event in all its detail. It would be suf?cient if it could be said what, in general, caused this MIS to be implemented in this particular form in these circumstances and to ignore apparently unimportant details. Invariance for a critical realist relates to elements of what they claim are the constituents of a causal explanation. Sayer puts it this way “Structures can therefore be said to be ‘invariant under certain
Fig. 3. Outline summary of the operation of a possible mechanism.
G. Easton / Industrial Marketing Management 39 (2010) 118–128
127
transformations’, that is, they can continue to exist while their constituents undergo changes in attributes which are not relevant to their reproduction” (Sayer, 1992 p.94). If a defensible causal explanation has been produced in one case then the constituents of that explanation provide a basis for developing theory beyond that case. This means, ?rstly, that the identi?cation of the entities concerned should been useful in terms of further theory development. Here the entities which are of most interest in the long term for industrial marketing research are the selling organisation and the buying organisation and their relationship. In other situations the net or network might be central. However these entities will usually be at too general a level to explain interorganisational events and the entities that form their internal structures, for example organisational form and culture, will be of more use. The structures of the bodies have been elucidated to some extent and relations among them established. In the Black Box case the relationship between the project coordinator and the rest of the organisation was a crucial internal structure. Secondly, the powers and liabilities of entities need to be studied empirically. For example these can, even in this relatively simple case, be hypothesised for an MIS and might include the power to transform customer relationships and the liability to destroy the organisation into which it has been placed, the power of CREOLE to withhold training resources and the liability of Black Box to accept CREOLE's training system are examples in the Black Box case. Thirdly the necessary relations have to be con?rmed, as was the case between CREOLE and Black Box, since events could not have been explained unless the relationship between the two ?rms is implicated. Also contingent relations have been shown to operate such as the impact of the KTP forum. It might be expected that contingent relations would be the most dif?cult to deal with since they are do not always affect events. For example in the Black Box case the KTP was a crucial element in explaining the events but since it was a time limited organisation it is hardly likely to be of much use in furthering theory. However if the focus of the research was change agencies and whether they were successful then it would have a necessary relationships with those organisations and might form the basis for a theory about organisational change. Fourthly the evidence appears to support the notion that certain mechanisms, the central components of a critical realist theory, have been shown to operate; the perceived and enacted balances of responsibility between CREOLE and Black Box. It would certainly be possible to think of other mechanisms that might apply here and these could be induced from the data or deduced from other theories or by means of creative thought. Some phenomena can be explained by building from unique mechanisms and contingent variables. Others occur (perhaps closer to the natural than the social world) that are explicable in terms of rather uniform causal mechanisms. However most phenomena can probably be explained by a mixture of the general and the speci?c. In this situation the case method has the task of uncovering the structure of underlying reality in each case, investigated to see what can be added to theory in the way of con?rmation and articulation. A second form of case development is when more than one case is used. A key decision in case research is what number of cases to research. The following conditions suggest when a small, number, perhaps one, rather than many would be wise. Where there is heterogeneity, or even uniqueness, of events then any kind of constant conjunction is unlikely to be detectable and seeking a credible casual mechanism is the only way to advance theory. Such heterogeneity may however be caused by the impact of a large number of powerful and active contingent relations and so working through the particular ways in which they operate in a single case will be not be as rewarding as using several cases. Explanations can be more or less “deep” and “wide”. They can incorporate a number of different emergent levels (individual, group and organisation) and a number of different entities. In general deeper and wider explanations are to be preferred in order to promote
theory development in the ways described above but there is always likely to be a trade-off with available research resources, access and theoretical complexity. Relationship with existing theory is important. Where little exists then one case can be enough to begin the process of theory creation. Where there exists well-articulated theory, particular aspects of that theory, the entities, their powers, the nature of the relationships and the overall mechanisms can be targeted and attempts made to elucidate one or all of them in a single case. More generally, a pragmatic approach can be adopted. If the objective is to advance theory then one should be able to say which theoretical aspects one wishes to address and how will that be served by undertaking one case study compared with many. 8. Managerial implications Research into management decision making suggests that managers often rely on gut feel or in academic terms, intuition. Since many, if not most, decisions have to be made quickly they have little time to think through, from ?rst principles, what the real issues are, what options are available and what value to be placed on each of those options. They rely on experience which, in practice, means recognition primed decision making (Klein, 1998). They recognise a particular pattern in the data, usually events, and so take the action that has worked for them, most of time in the past. For example if a new buyer asks for price reduction from a marketing manager she has usually found that it is best to agree, after vigorously arguing about it. Then, some months later, she asks for a price increase on “rational” grounds and it will be accepted most of the time. Of course intuition also means drawing on other experiences when the pattern isn't quite normal. The managerial implications of a critical realist approach to decision making is that managers should be encouraged to think about why certain decisions lead to certain outcomes; to try to discover what causes them or in critical realist terms what the causal mechanisms are. They may be forced into doing this anyway when a novel situation arises (selling into a poor African country) and they have to carry out a real time case analysis, collecting data and trying to ?gure out if the usual strategies will or won't work. However there is also a case for managers thinking about their intuitive powers. Experience has taught them that certain actions have good outcomes. But why? If they knew that, then they might be more sensitive in detecting possible rogue situations where normal actions probably won't work. They might also be able improve on existing decision making if they understood the causal mechanisms that were in play. For example in the new buyer situation mentioned earlier it might be that the buyer is trying to impress new colleagues by getting a price reduction. A few months later they are either busier or have less need to impress and so don't worry about a price increase. But which is it most likely to be the case and what are the appropriate tactics? That would depend on a number of factors that could be investigated, albeit rather subtly. It would be naïve to argue that managers should build critical realist causal models. However even if they simply try to employ causal language and, at the same time, become more inquisitive about the situations they face there could be real bene?ts to be gained. 9. Conclusions Case study research is the prevalent research method in B2B research. However it is often justi?ed, when it is justi?ed at all, in terms of the interesting results or unusual phenomena it reveals. What it lacks is philosophical validation, i.e. ontological and epistemological underpinnings. I argue in this paper that critical realism provides such underpinnings and seems ideally matched to case research. Certainly case research cannot be justi?ed in terms of positivism since case research is almost always small numbers
128
G. Easton / Industrial Marketing Management 39 (2010) 118–128 Dickens, P. (2003). Changing our environment, changing ourselves: Critical realism and transdisciplinary research. Interdisciplinary Science Reviews, 28(2), 95?105. Dubois, A., & Gadde, L. -E. (2002). Systematic combining: An abductive approach to case research. Journal of Business Research, 55, 553?560. Easton, G. (1995). In D. T. Wilson & K. Moller (Eds.), Methodology and industrial networks in relationships and networks theory and application Norwell, Mass: Kluwer Academic Publishers. Easton, G. (1998). Case research as a methodology for industrial networks: A realist apologia. In P. Naude & P. W. Turnbull (Eds.), Network dynamics in international marketing Oxford: Elsevier Science. Easton, G. (2002). Marketing: A critical realist approach. Journal of Business Research, 55(2), 103?109. Easton, G. and Araujo, L. Language, metaphors and networks in advances in international marketing, (5), 67–85. ed., Deo Sharma, New York, JAI Press, 1993. Eisenhardt, K. M. (1989). Building theories from case study research. Academy of Management Review, 14(4), 532?551. Gerring, J. (2004). What is a case study and what is it good for. American Political Science Review, 98(2), 341?354. Hanson, J., & Yosifon, D. (2004). The situational character: A critical realist perspective on the human animal. Georgetown Law Journal, 93(1), 1?179. Hanley, C. (1995). On facts and ideas in psychoanalysis. International Journal of Psychoanalysis, 76(5), 105?136. Houston, S. (2001). Beyond social constructionism: Critical realism and social work. British Journal of Social Work, 31(6), 845?861. Keat, R., & Urry, J. (1975). Social theory as science (pp. 9?12). London: Routledge and Keagan Paul. Hunt, S. D. (2003). Controversy in marketing theory: For reason, realism, truth, and objectivity. Armonk, N.Y: M.E. Sharpe. Kendall, J. E., & Kendall, K. E. (1993). Metaphors and methodologies: Living beyond the systems machine. MIS Quarterly, 17(2), 149?162. Klein, G. (1998). Sources of power: How people make decisions. Boston: The MIT Press. Lau, R. W. K. (2004). Critical realism and news production. Media, Culture and Society, 26(5), 693?711. Lawson, T. (1997). Economics and reality. London: Routledge. Layder, R. (1990). The realist image in social science. London: Macmillan. Lewis, P. A. (1996). Metaphor and critical realism. Review of Social Economy, 54(4), 487?506. Mitchell, J. C. (1983). Case and situation analysis. The Sociological Review, 3(1), 187?211. Nellhaus, T. (1998). Signs, social ontology, and critical realism. Journal for the Theory of Social Behaviour, 28(1), 1?25. Pawson, R., & Tilley, N. (1997). Realistic evaluation. Sage: London. Proctor, J. D. (1992). The social construction of nature: Relativist accusations, pragmatist and critical realist responses. Annals of the Association of American Geographers, 88(3), 352?376. Robbins, J. W. (1999). Pragmatism, critical realism, and the cognitive value of religion and science. Zygon, 34(4), 655?666. Sayer, A. (1984). Method in social science. London: Routledge. Sayer, A. (1992). Method in social science: A realist approach, (2nd ed.) London: Routledge. Sayer, A. (2000). Realism and social science. London: Sage. Steinmetz, G. (1998). Critical realism and historical sociology; A review article. Comparative studies in society and history, 40(1), 170?186. Verschuren, P. J. M. (2003). Case study as a research strategy: Some ambiguities and opportunities. International Journal of Social Science Research Methodology, 6(2), 121?139. Trosper, R. L. (2005). Emergence unites ecology and society. Ecology and Society, 10(1) Article 14. Tsoukas, H. (1989). The validity of idiographic research explanations. Academy of Management Review, 14(4), 551?561. Wikgren, M. (2005). Critical realism as a philosophy and theory in information science? Journal of Documentation, 6(1), 11?22. Woodside, A. G., & Wilson, E. J. (2003). Case study research methods for theory building. Journal of Business and Industrial Marketing, 18(6/7), 493?508. Woodside, A. G., Pattinson, H. M., & Miller, K. E. (2005). Advancing hermeneutic research for interpreting inter?rm new product development. Journal of Business and Industrial Marketing, 20(7), 493?508. Yeung, H. W. C. (1997). Critical realism and realist research in human geography: A method or a philosophy in search of a method? Progress in Human Geography, 21(1), 51?74. Yin, R. K. (1989). Case study research: Design and methods. Newbury Park: Sage. Yin, R. K. (2003). Case study research: Design and methods. Newbury Park: Sage. Geoff Easton is a professor in the Department of Marketing at Lancaster University. His main research interests are B2B markets and marketing, organisational and managerial behaviour, strategic practices, critical realism and case teaching methods. He has published in Industrial Marketing Management, the Journal of Management Studies, OMEGA, British Journal of Management, European Journal of Marketing, Organization, Journal of Business Research, Journal of World Business and Complex Systems Advances.
research. Interpretivism is more relevant but is largely epistemological in its objectives. Critical realism however provides not only a basis for justi?cation but also guidelines as to how case research might be done and how theory can be fashioned. Critical realism ?rst of all makes the ontological assumption that there is a reality but that it is usually dif?cult to apprehend. It distinguishes between the real world, the actual events that are created by the real world and the empirical events which we can actually capture and record. Thus we will always be surmising about the nature of the real. Critical realists argue that in the real world there are entities, such as organisations, which have powers to act and are liable to be acted upon by others. These entities can also have internal structures, such as departments and individuals which in their turn, have their own powers. Entities have relationships among themselves which are de?ned as necessary in terms of our understanding of their powers, for example buyers and sellers have necessary relationships by virtue of the designations that are given to those who act by buying or selling. They may also have a contingent relationship, that is, they may or may not have such relationships. For example buyers and sellers may be connected to the same EDI system or they may not and so the EDI system is in a contingent relationship to the buyer and seller. The relationship among entities that is thought to cause the events is called a mechanism. A critical realist approach to case research involves developing a research question that identi?es a research phenomenon of interest, in terms of discernible events, and asks what causes them to happen. The key entities involved, their powers, liabilities, necessary and contingent relationships are then provisionally identi?ed. Research then proceeds by capturing data with respect to ongoing or past events asking at all times why they happened or are happening and taking into account the problems and issues associated with interpreting the empirical data back to the real entities and their actions. The research process is one of continuous cycles of research and re?ection. The ?nal result is the identi?cation of one or more mechanisms that can be regarded as having caused the events. This process was exempli?ed via a real life case study, the research for which was carried out in this way, and 3 possible mechanisms were identi?ed at work in a post purchase service relationship between an MIS provider and a buyer. Finally it was suggested that generalisation to theory via case research carried out under critical realist conventions occurs by virtue of clarifying the theoretical nature of the entities involved, the ways in which they act and the nature and variety of mechanisms through which they exert their powers or acted upon by other entities. As with all philosophical approaches critical realism cannot be proved to be the “right answer”. Acceptance of a critical realist approach depends on whether you agree with its basic assumptions and that acceptance can be based on any number of things. In my own case it I accept it because (a) I think that this is how the world is, (b) even if it isn't like that, I behave as if it is like that, (c) I think that critical realism is better than the alternatives and (d) it is a well thought through and relatively coherent perspective on the world.
References
Ackroyd, S., & Fleetwood, S. (2004). Critical realist applications in organisation and management studies. London: Sage. Archer, M., Collier, A., & Porpora, D. V. (2004). Transcendence. London: Routledge. Bania, R. (1995). Critical realism and urban and regional studies. Environment and Planning B—Planning and Design, 22(5), 563?580. Bhaskar, R. (1978). A realist theory of science. Hemel Hempstead: Harvester Press.
doc_295619369.pdf
Marketing management is a business discipline which is focused on the practical application of marketing techniques and the management of a firm's marketing resources and activities. Globalization has led firms to market beyond the borders of their home countries, making international marketing highly significant and an integral part of a firm's marketing strategy.
Industrial Marketing Management 39 (2010) 118–128
Contents lists available at ScienceDirect
Industrial Marketing Management
Critical realism in case study research
Geoff Easton ?
Department of Marketing, Lancaster University Management School, Lancaster University, Lancaster, LA1 4YX, United Kingdom
a r t i c l e
i n f o
a b s t r a c t
Case study research is the most popular research method for researchers in industrial marketing. However despite a number of attempts the problem of satisfactorily justifying the use of case research remains. This paper proposes critical realism as a coherent, rigorous and novel philosophical position that not only substantiates case research as a research method but also provides helpful implications for both theoretical development and research process. The article describes the critical realist approach due to Sayer and develops a general application of a critical realist approach to case research. An example of its use in practice is presented using a case study of the development of a buyer–seller relationship after the installation of a new MIS system. © 2009 Elsevier Inc. All rights reserved.
Article history: Received 1 May 2007 Received in revised form 12 June 2008 Accepted 20 June 2008 Available online 24 September 2009 Keywords: Case study research Critical realism MIS Causality
1. Introduction Case study research is, probably, the most popular research method used by industrial marketing researchers. This may be, in part, because of the nature of the subject. The main units of analysis are organisations and relationships, which are dif?cult to access, and complex in structure in comparison with, for example, consumer markets. As a result a case study of a single, or a small number, of such entities can provide a great deal of, largely qualitative, data which can be written up as a case study, offering insights into the nature of the phenomena. But how do we know that what is written as a case analysis represents the “truth”? How can we justly claim that we know, in some fundamental sense, what it is that we have researched? Often the use of the case method is not justi?ed at all in the resulting published work. When it is warranted, the justi?cation is made on the grounds of the interesting and novel nature of the results, exempli?cations of particular phenomena or applications of speci?c concepts or models. Few authors of case based papers offer a defence of their choice of the case method on formal epistemological grounds. Such reticence is hardly surprising. Making truth claims based upon such seemingly limited data is clearly a daunting prospect. This discretion is at least partly due to the dominance of the epistemologically positivistic underpinnings of most academic research in marketing. Clearly the sample size in any case study research project is never going to be large enough to qualify for the use of statistical inference. However a number of writers on case research as a research method seem to take a positivist position without perhaps realising it. For example, Eisenhardt in a widely cited paper offered
? Tel.: +44 1524 592917. E-mail address: [email protected]. 0019-8501/$ – see front matter © 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.indmarman.2008.06.004
the following advice on the number of cases to be used. “Finally, while there is no ideal number of cases, a number between 4 and 10 cases will usually work out well” (Eisenhardt (1989). The justi?cation for this statement is based on her experience with case research and is implicitly about increasing the number of cases as a way of ?nding the same results in each case. However there are a number of reasons why positivism would not be the answer anyway. Its de?ning feature is its nomothetic epistemological stance which implies that there exist regularities or law-like generalisations in material or social settings that provide the basis for both explanation and prediction. This regularity allows positivists to believe that they can make causal statements. If two events occur in sequence regularly then one is said to explain the other. However this simple and elegant formulation has any number of problems which makes its use in any research situation problematic. The most crucial problem is that constant conjunction of elements or variables is not a causal explanation or indeed an explanation of any kind. It is simply an atheoretical statement about the world. It doesn't answer the question why? Sometimes interpretivism, in one of its many forms, is used as a way of defending the use of case studies. Researchers simply interpret cases placing the weight of the research on authentic ways of data capture and sensitive and detailed data analysis. In general interpretivists deny the possibility of knowing what is real and reject the possibility of discerning causality. They can only provide their own interpretation. What is not clear in the interpretivist approach is by what standards one interpretation is judged to be better than another. It is even more problematic when the interpretations are particularistic since this would appear to rule out not just regularity as a criterion but also any form of comparison. In complete contrast the original pragmatists, such as Dewey, James and Pierce argued that it is the uses to which truth is put that
G. Easton / Industrial Marketing Management 39 (2010) 118–128
119
are important. Pragmatism espouses usefulness but only speci?cally and in context. Truth is what is useful to people researching in a ?eld, what helps the research project, what can be accepted and defended, what is open to criticism and renewal. It is a linguistic convention, a sort of shorthand that helps us to achieve our various objectives when researching and theorising. Clearly pragmatism can provide a very powerful justi?cation for the use of case studies since case studies as a research method offers the possibility of studying a problem de?ned situation in great detail. Indeed many action research and illuminative evaluation studies are, in effect, based on single cases of, usually, organisations. However there have been very few papers published in industrial marketing where the data describe problem solving situations and none that use pragmatism as a justi?cation for the validity of the interpretation. For a more extensive discussion of philosophical orientations and case research see Easton (1995), Easton (1998). In this paper I argue that critical realism offers a way forward. Critical realism is, by philosophical standards, a relatively new approach to ontological, epistemological and axiological issues. In a search of the ISI Web of Science database there were 334 papers which included critical realism in the title or abstract. Of those only 42 were in journals that could be said to be in management or organisation studies (and mostly in the latter). Again, of those 42 only 4 papers were in marketing journals and I was author or coauthor of 3 of them. Only 2 of the 334 dealt with case research and critical realism and I was a co-author of both of them. The fundamental tenet of critical realism is that we can use causal language to describe the world. Since all philosophical positions rely on assumptions they can only be ultimately judged pragmatically, not in the limited sense used by pragmatists but in terms of our beliefs that they result in better explanations. One powerful pragmatic argument in favour of critical realism is that it is performative. Critical realists assume that there is a real world out there. However there is no way that such an assumption can ever be proved or disproved, as social constructivists, pragmatists and even positivists are ready to argue. But this assumption is surely performative. In other words we behave as if it was true, as if the world was real. In general this supposition works, especially for the physical world. For example no constructivist would dare to say any longer that the world is totally socially constructed since that is in itself a realist statement. “In both everyday life and social science, we frequently explain things by reference to causal powers” (Sayer, 2000, p.14). Critical realism mirrors the language and procedures we routinely adopt and the explanations that we create. We use causal language without thinking. Critical realists argue for the use of causal language with thinking. Critical realism is particularly well suited as a companion to case research. It justi?es the study of any situation, regardless of the numbers of research units involved, but only if the process involves thoughtful in depth research with the objective of understanding why things are as they are. The paper is structured as follows. It begins with a discussion of the nature of case research and proceeds to a description of critical realism. It then offers an examination of the implications of adopting a critical realist justi?cation of case research and continues with an example of a critical realist case analysis involving the creation of a buyer–seller relationship through the (problematic) implementation of a new Management Information System (MIS). The issue of the generalisability of case research is then discussed, managerial implications are suggested and the paper concludes with a summary of the issues covered. 2. What is case research? This is a very dif?cult question to answer since, in practice, the label is often attached to quite disparate forms of research method. “The term ‘case study’ may refer to several different epistemological entities” (Mitchell (1983) in Verschuren (2003 p.122).
“Indeed the case study is probably best understood as an ideal type rather than a method with hard and fast rules. Yet the fact that the case study is fuzzy round the edges does not mean that it doesn't have distinctive characteristics” (Gerring, 2004 p.346). However sampling mode surely de?nes case research. A case is a single instance; a sample of one. Once the decision to use case research has been made a set of constraints and opportunities are realised. The key constraint is its low (statistical) representativeness. Although it is possible to research several cases this is not done in order to increase the sample size in the conventional sense. The logic of generalisability is totally different for case research and this will be dealt with later in the paper. A single case study must be able to stand on its own. The key opportunity it has to offer is to understand a phenomenon in depth and comprehensively. Research questions are de?nable in terms of the questions; who, what, where, how and why (Yin, 1989, p.18; Yin, 2003. p.5). Case studies are more suited to how and why questions which can be explanatory in nature. “This is because such questions deal with operational links needing to be traced over time, rather than mere frequency or incidence” (Yin, 1989 p.18; Yin, 2003, p.6). Case research allows the researcher the opportunity to tease out and disentangle a complex set of factors and relationships, albeit in one or a small number of instances. This is a process of iterative–parallel research which “…implies a continuous moving back and forth between the diverse stages of the research project” (Verschuren, 2003). The ?exibility that case research allows in this respect is one of its major advantages and one that is not shared by, for example, survey based methods. Case research can therefore be de?ned as a research method that involves investigating one or a small number of social entities or situations about which data are collected using multiple sources of data and developing a holistic description through an iterative research process. 3. Critical realism 3.1. Provenance Critical realism assumes a transcendental realist ontology, an eclectic realist/interpretivist epistemology and a generally emancipatory axiology. While critical realism is a relatively new orientation it is being taken up in many disciplines including economics (Lawson, 1997), sociology (Sayer, 2000; Layder, 1990), criminology (Pawson & Tilley, 1997), geography (Proctor, 1992; Yeung, 1997), linguistics (Nellhaus, 1998), religious studies (Robbins, 1999), history (Steinmetz, 1998), psychiatry (Hanley, 1995), social work (Houston, 2001), ecology (Trosper, 2005), environmental studies (Bania, 1995), law (Hanson & Yosifon, 2004), information studies (Wikgren, 2005 ), media studies (Lau, 2004), interdisciplinary science studies (Dickens, 2003) and management (Ackroyd & Fleetwood, 2004). There are a number of differing views and approaches to realism (Hunt, 2003). This paper espouses the version due to Sayer since his account of critical realist ontology is the most detailed and comprehensive and so makes it easier to demonstrate how it can provide both a philosophical justi?cation for case research and a guide to its use in practice (Sayer, 1992). In addition Hunt recognises Sayer as the key ?gure in the critical realism movement (Hunt, 2000 p286). Therefore in what follows the phrase critical realism means Sayer's version of the ontological position. 3.2. Basic assumptions of critical realism Sayer sets out what he regards as the 8 key assumptions of critical realism in the following extract from his book. 1. “The world exists independently of our knowledge of it. 2. Our knowledge of the world is fallible and theory-laden. Concepts of truth and falsity fail to provide a coherent view of the
120
G. Easton / Industrial Marketing Management 39 (2010) 118–128
3.
4.
5.
6.
7.
8.
relationship between knowledge and its object. Nevertheless knowledge is not immune to empirical check and its effectiveness in informing and explaining successful material practice is not mere accident. Knowledge develops neither wholly continuously, as the steady accumulation of facts within a stable conceptual framework, nor discontinuously, through simultaneous and universal changes in concepts. There is necessity in the world; objects—whether natural or social— necessarily have particular powers or ways of acting and particular susceptibilities. The world is differentiated and strati?ed, consisting not only of events, but objects, including structures, which have powers and liabilities capable of generating events. These structures may be present even where, as in the social world and much of the natural world, they do not generate regular patterns of events. Social phenomena such as actions, texts and institutions are concept dependent. We not only have to explain their production and material effects but to understand, read or interpret what they mean. Although they have to be interpreted by starting from the researcher's own frames of meaning, by and large they exist regardless of researchers' interpretation of them. A quali?ed version of 1 therefore applies to the social world. In view of 4–6, the methods of social science and natural science have both differences and similarities. Science or the production of any kind of knowledge is a social practice. For better or worse (not just worse) the conditions and social relations of the production of knowledge in?uence its content. Knowledge is also largely—though not exclusively— linguistic, and the nature of language and the way we communicate are not incidental to what is known and communicated. Awareness of these relationships is vital in evaluating knowledge. Social science must be critical of its object. In order to be able to explain and understand social phenomena we have to evaluate them critically” (Sayer, 1992, p.5).
fundamental nature and capabilities of the things we research rather than simply their measurable properties. It requires a shift from epistemology and methodology to ontology. 3.4. Causal powers and liabilities Entities have causal powers and liabilities. Causality is, of course, a subtle and disputed concept which Sayer attempts to capture by a process of interpolation using what he describes as an “ordinary” (arguably pragmatic) account of causality. “To ask for the cause of something is to ask ‘what makes it happen’, what ‘produces’, ‘generates’, ‘creates’ or ‘determines’ it, or, more weakly, what ‘enables’ or ‘leads to’ it” (Sayer, 1992, p.104.) Sayer also argues, “…..particular interpretations (of causality) can only be justi?ed in terms of their compatibility with our most reliable beliefs…” Put another way, they rely on an assumption which, together with other assumptions, create a system of thinking about the world that we ?nd acceptable. We must have reason to believe that bodies that we study have powers or liabilities to cause events to occur. They make things happen. Using an example from the case study used in the latter part of the paper, MIS have the powers to change organisations in any number of ways. Similarly organisations have the powers to modify new MIS that are in the process of being implemented or in the ways in which they operate. A liability may be regarded as a susceptibility to the action of other entities, for example particular kinds of organisations may be liable to have particular MIS implementation problems. The bene?t of this conceptualisation is that it focuses attention on three key questions. What are the entities that de?ne our research ?eld, what are their relationships and what are their powers and liabilities? 3.5. Events Events or outcomes are what critical realists investigate, that is the external and visible behaviours of people, systems and things as they occur, or as they have happened. However it is important to acknowledge that most social science research methods create data that are reported rather than directly observed. Descriptions of the events that occur during the implementation of an MIS are rarely experienced at ?rst hand or recorded in a way that is close to the event. Particular attention is paid to processes in critical realist accounts, especially those that produce and reproduce the ordering of events and social institutions. Again MIS are systems that reproduce themselves and are reproduced by other systems in place in organisations. Critical realists also believe that the non occurrence of an event when one is expected not only requires explanation but may also provide very useful insights. Failure of an MIS to provide some of its designed outcomes would be an example of such a non occurrence. Again this is a subtly different approach to that normally adopted in social research. It places the focus, at least in the beginning, on that which we can mostly clearly discover. 3.6. Structure of entities Entities will usually be structured. Structure is “...a set of internally related objects or practices” (Sayer, 1992, p.92). For example an organisation may be considered to comprise a series of other entities (departments, people, processes, resources) all of which can affect one another. Structures are nested within structures. For example entities can be organisations that have departmental structures and relations and, within them, individuals who have particular characteristics such as gender and psychological structures. Gender forms part of the internal relations of a person if gender is to be regarded as a necessary part of the structure that is being built.
Points 1, 4 and 5 set out the key critical realist assumptions about ontology. Critical realists propose an ontology that assumes that there exists a reality “out there” independent of observers. A naïve realist epistemology would assume that this reality can be readily accessed. This is a view often espoused by researchers in the natural sciences because of their ability to measure accurately and their access to controllable and / or closed systems. However these conditions rarely occur in social systems. As a result critical realists accept that reality is socially constructed and points 2, 3, 6 and 7 spell out this proposition thus creating a tension between these apparently contradictory views. However critical realists resolve the tension by arguing that the world is socially constructed but not entirely so. The “real” world breaks through and sometime destroys the complex stories that we create in order to understand and explain the situations we research. 3.3. Objects/entities Objects, or more generally entities, provide the basic theoretical building blocks for critical realist explanation and can be such things as organisations, people, relationships, attitudes, resources, Management Information Systems (MIS), inventions, ideas and so on. They can be human, social or material, complex or simple, structured or unstructured. Entities stand in contrast to the idea of variables that dominates most social research traditions. Variables are measures of things and not the things themselves. “Similarly, the concept of variable that is used in quantitative analysis is an indifferent one as regards causal explanation: variables can only register (quanti?able) change, not its cause” (Sayer, 1992, p.180). This modi?cation radically alters the way that we need to think about theory. It directs our attention to the
G. Easton / Industrial Marketing Management 39 (2010) 118–128
121
This conceptualisation suggests that there are two alternative ways in which MIS can be treated theoretically. The ?rst is to include MIS as a part of the structure of an organisation. The implication would then be that the study of an MIS might consider it to be primarily concerned with organisational behaviour as it relates to one particular part of its structure having interesting and important effects on its behaviour. A second approach would be to see the relationship between an organisation and an MIS as external so that one does not encompass the other but is related to it. This might be a way to deal with changes in MIS. 3.7. Emergence In the language of critical realism, entities may be analysed at a number of different levels of aggregation. A crucial critical realist assumption concerns the existence of emergence in such situations. The properties of entities at a higher level of aggregation are not necessarily understood through a summative process or, working from the top down, a reductionist approach. They emerge from those of the lower level but are not easily derived from them. For example MIS systems have emergent properties that are more than, and different from, the sum of their constituent parts. Similarly entities at a higher level such as organisations cannot simply be reduced to the summation of their components such as MIS. The implication is that in choosing a level of analysis one accepts that ready access to other levels is not necessarily easy. The current inability of biochemists and physiologists to ?nd a biochemical model of consciousness provides a case in point. Emergence must always involve some element of connectedness. For example biology could be regarded as an emergent from chemistry in the sense that it is the combinations and connections among atoms and molecules that create and sustain biological processes. Similarly the social world is only understood through the connections between the people that comprise a society not by studying the individuals in isolation. Closer to home the properties of organisations stem, in part, from the connections among the individuals and groups they contain. 3.8. Necessary relations Critical realists argue that there are two kinds of relationships among entities; necessary and contingent. As Sayer writes “…the relation between a slave and a master is necessary, in that what the object is dependent on its relation to the other; a person cannot be a slave without a master and vice versa” (Sayer. 1992, p.89). In terms of MIS, for example, organisations and MIS have a necessary causal relation since one cannot exist without the other. Organisations are managed and involve information systems, though not necessarily ones that are electronically based. MIS must exist within organisations. Necessary relations are not tautologies but derive directly from the nature of the bodies involved. The relation between entities and the events they cause will usually be a rich and varied one. Thus the elements of the relation are not simply given by the mutual de?nition. MIS and organisations have many ways of relating to and affecting one another. Entities are, however, de?ned in terms of their necessary relations. A person must be able to use or ignore an MIS. For complex entities there may be any number of relations that de?ne both it and the other entities to which it is related. Necessary relations are not inevitably rigid though changes in one entity will, of necessity, lead to changes in the other. Necessary relations do not have to be particularly important since this depends on the nature and objectives of the research being carried out. Of course that does not mean they do not exist. None of these ideas are likely to be particularly surprising. Research builds up theories that comprise a number of concepts
that refer to particular entities and specify the relations among them to create a theory or theoretical framework. But it is clear that in doing so we are relying on the relations among entities to hold the whole edi?ce together. In other words the de?nitions become referential and interdependent. Change in one body leads to change in another body with which it has necessary relations. That does not mean that there have to be regular changes though there may be. The change may also lead to the body changing its nature (e.g. from a national to a multinational corporation). Internal relations do not have to be symmetric. One body may be not be able to exist without the other, for example an organisation cannot exist without individuals but individuals can exist without, at least, formal organisations. While bodies may de?ne relations it is equally likely that there will also be an element of mutual de?nition. A Chief Information Of?cer will be de?ned in terms of the necessary relations he or she has with other board members, other institutions, certain employees, etc. We rely on these referential and interdependent relations to underpin our theoretical understandings. 3.9. Contingent relations A contingent relation occurs when “It is neither necessary nor impossible that they stand in any particular relation” (Sayer, 1992, p.89). Put at its simplest this distinction recognises that entities can have some relations (necessary) that will affect one another and some (contingent) that may affect one another. The implementation of a new MIS system may be affected by the home country of the system supplier or it may not. There is no necessity because the relation is a contingent one. However contingent relations are different from necessary relations only in the nature of those relations. “……the contingently related conditions are never inert, but are themselves the product of causal processes and have their own causal powers and liabilities” (Sayer, 1992, p.140). The theoretical framework chosen governs the difference between necessary and contingent. All events need to be explained by a combination of necessary and contingent relations. If all relations were contingent then each explanation would be unique and incapable of contributing towards anything by way of generalisation. Sayer puts it this way. “…Structures can therefore be said to be ‘invariant under certain transformations’, that is, they can continue to exist while their constituents undergo changes in attributes which are not relevant to their reproduction” (Sayer, 1992, p.94). This is crucial because it recognises that there will and indeed must be invariance in the system. Building theory and progress then become possible. 3.10. Context Distinguishing between context and contingency is possible. The former offers a simpler, less well-articulated version of the latter. Context is simply “relevant circumstances”. It is a very general concept and says little about the relationship between the focal entities and the environment except that it is (possibly) relevant. In the case of critical realism the entity should not only be de?ned but the form of the causal relationship clearly set out. The question should be “in what ways may the external contingency affect the events that have occurred?” 3.11. The structure of causal explanation The most fundamental aim of critical realism is explanation; answers to the question “what caused those events to happen?” In Fig. 1 causal explanation is structured in terms of the relationships among the concepts that have so far been discussed. This is a formal statement of the critical realist structure of explanation. Less formally a very simple example demonstrates the
122
G. Easton / Industrial Marketing Management 39 (2010) 118–128
Fig. 1. The structure of casual explanation (Sayer, 1984).
most basic form that such a formal explanation can take. Objects (a salesperson) having structures (knowledge and personality traits etc) and necessarily possessing causal powers (to persuade a buyer, who is another object) and liabilities (to be rejected by technical buyers, to get tired towards the end of the day) will, under speci?c condition c1 (the buyer has a need for the product and the offering is suitable) result in an event e1 (a sale), or alternatively under speci?c condition c2 (the buyer has a need for the product but the offering is not suitable) will result in an event e2 (no sale). In practice such formal explanations will not normally be possible because of the complexity of real world behaviour but they do provide a logical framework to guide case researchers. However critical realists argue that there should always be competing explanations since different interpretations of the data are necessary to ensure that the “best” current interpretation is made. But there should always be room for revision through the normal processes of academic presentation and critique. 3.12. Mechanisms Mechanisms are “nothing other than the ways of acting of things” (Bhaskar, 1978. p. 14). A causal explanation is one that identi?es entities and the mechanisms that connect them and combine to cause events to occur. Other writers (Tsoukas, 1989; Keat & Urry, 1975; Layder, 1990) use somewhat different terms such as “deep structures”, “deep processes” or “generative mechanisms”. The term mechanism has problematic connotations since it implies clear structure and invariance in operation, something that critical realists would reject. A better portmanteau term would be deep generative processes and structures. However mechanism has become embedded in the language community and so will be used hereafter but with the caution that it should not be taken literally. Mechanisms are at the heart of causal explanation but also generate a constant source of debate. Perhaps the simplest way of regarding mechanisms is that they are ways in which structured entities by means of their powers and liabilities act and cause particular events. “When activated, particular mechanisms produce effects in “conjunctures”, which may be unique. According to conditions, the same mechanism may sometimes produce different events, and conversely the same type of event may have different causes” (Sayer, 1992, p.116).
A simple example of a mechanism would be a log jam or tipping point. When a buyer and seller have agreed on almost everything (structure and conditions) there is always the possibility of a deal breaker emerging. The crucial deal breaker condition then solely determines whether the event will take place. Recognising that the deal breaker is actually a deal breaker rather than just a part of the players' negotiation strategy then becomes crucial. An important aspect of mechanisms in the critical realist tradition is that they offer a rich source of explanatory devices. They do not need to be linear additive as required by statistical models or logicorational as in box and arrow diagrams. Instead they can be linguistic in nature and metaphorical (Easton & Araujo, 1993). A brief survey of research in the social sciences produced this sample of source metaphors in use; marriage, garbage can, contagion, room and key, photography, war, dance, fortress, jazz, life cycle, tourist, theatre, cancer and computer. However, “A generative metaphor is not merely an ornate expression of similarities and analogies its author was already aware of, but is the source of new perceptions of similarity and analogy, picking out similarities and analogies that were unknown until the metaphor pointed them out and thereby brought them to the author's attention” (Lewis, 1996 p.493). At a more formal level, various models or less well speci?ed theories such chaos, catastrophe and complexity theory could be adopted. For example Kendall and Kendall (1993) elicit 9 metaphors that were used by information systems users and in systems development methodologies offering the potential to use one or more of them in developing explanations for MIS events.
3.13. Epistemology Critical realists accept that our world is, of course, socially constructed but argue that this is not entirely the case. They construe rather than construct the world. Reality kicks in at some point. “Critical realism acknowledges that social phenomena are intrinsically meaningful, and hence that meaning is not only externally descriptive of them but constitutive of them (though of course there are usually material constituents too). Meaning has to be understood, it cannot be measured or counted, and hence there is always an
G. Easton / Industrial Marketing Management 39 (2010) 118–128
123
interpretative or hermeneutic element in social science” (Sayer, 2000, p.17). The difference between critical realists and social constructionists lies in the acceptance of the possibility of knowing reality in the former case and its rejection in the latter who, in general, concentrate instead on uncovering the constructions that social actors make. Bhaskar (1978) has suggested that critical realism has a strati?ed rather than ?at ontology and this has major epistemological implications. The strata are the empirical, the actual and the real. The empirical domain is where observations are made and experienced by observers. However events occur in the actual domain and may be not observed at all or may be understood quite differently by observers. There is a process of interpretation that intervenes between the two domains. Events occur as a result of mechanisms that operate in the real domain. It is not the case that the real or actual cannot be observed but simply that it may not always be capable of being observed. We see just the tip of an iceberg but that doesn't mean that the invisible three-quarters is not there or is unconnected to what we see. Observation is fallible. It is unlikely to reveal completely and lead to a full understanding of any social situation. Since there can be no de?nitive criteria to judge the “truth” of a particular version, critical realism relies on the researcher to collect further data that helps to distinguish among alternative explanations and on the community of researchers to debate them thoroughly. As a result criticality within a discipline becomes essential since only by seeing the same data through the different theoretical lenses employed by different researchers can understanding of some of the features of the real world occur (Woodside & Wilson, 2003; Woodside, Pattinson, & Miller, 2005). 3.14. Research process Critical realists further distinguish themselves from other schools of thought in terms of the way that they believe that the process of research should be conducted. Pierce, ultimately a pragmatist, ?rst suggested in 1911 what the alternatives might be; induction, deduction and retroduction. Retroduction is a “…mode of inference in which events are explained by postulating (and identifying) mechanisms which are capable of producing them…” (Sayer, 1992, p.107). Put another way “...there is a concern in realist accounts, with a different form of inference to the more common induction and deduction. Whereas the latter are concerned with movements at the level of events from the particular to the general and vice versa, retroduction involves moving from a conception of some phenomenon of interest to a conception of a different kind of thing (power, mechanism) that could have generated the given phenomenon” (Lawson, 1997, p.236). Retroduction means “moving backwards” and that is what the process involves. It asks “What must be true in order to make this event possible?” Abduction is a closely related process which has already been suggested as appropriate for case research (Dubois & Gadde, 2002). 4. Critical realism and case research Sayer argues that critical realism is relatively tolerant with respect to different research methods. “Compared to positivism and interpretivism, critical realism endorses or is compatible with a relatively wide range of research methods, but it implies that the particular choices should depend on the nature of the object of study and what one wants to learn about it” (Sayer, 2000, p.19) (Fig. 2). He suggests that there are two broad types of research method; extensive and intensive. The former employs large scale surveys, formal questionnaires and statistical analyses, looks for regularities, patterns and similarities, accepts given taxonomic categories, privi-
Fig. 2. Critical realist view of causation (Sayer, 2000, p15).
leges replication and has restricted ability to generalise to other populations and limited explanatory power. The latter focuses on individual agents in context using interviews, ethnography and qualitative analysis, asks the question “what produces change?”, employs causal groups, produces causal explanations which are, however, limited to the situation studied so that testing is by corroboration. However he points out that, “the extensive/intensive distinction is not identical to the survey-analysis / case study or ethnography distinction; extensive methods might be used within a single case study; intensive approaches do not have to be limited to single cases and can use other methods besides ethnographic ones.” These descriptions af?rm that case research is an intensive research method according to Sayer's description and is therefore entirely consistent with a critical realist ontology. However it is important to consider how this might work in practice.
4.1. Critical realist case method The ?rst task in any research project is to decide on the phenomenon to be studied. A critical realist case approach is particularly well suited to relatively clearly bounded, but complex, phenomena such as organisations, interorganisational relationships or nets of connected organisations. It is less well suited, for example, to the study of individual behaviour, although that can be done, or to situations which are largely characterised in quantitative terms such as perceptions and attitudes tracking data. The boundaries of the phenomena, for example organisation or net, must be determined although it is not uncommon for those boundaries to be changed during the course of the research. For example, determining causality may require that the researcher moves beyond the initial boundary or, alternatively, narrows the boundary because it is clear that the causal mechanisms are more narrowly focused than previously thought. The next issue is the nature of the research question. Here critical realism has a rather clear answer. The question must be of the form “What caused the events associated with the phenomenon to occur”. As suggested previously, it is only possible to understand social phenomenon by recording and analysing the associated events that take place as a result of the actors acting, whether they are human or non human (e.g. machines such as computers). The events can be recorded live or exist in records of the past including the memories of those human actors who can attest to the events. The next and crucial task is that of identifying the entities/objects that characterise the phenomena being studied. Sayer points out that this process is usually given less attention than it requires. “Much rests on the nature of our abstractions, that is, our conceptions of particular one-sided components of the concrete object; if they divide what is indivisible, or if they con?ate what are different and separable components, then problems are likely to result” (Sayer, 2000, p.19). For example would it be sensible to consider a B2B relationship as a single entity or as a separate feature of each of the organisations involved? The Actors–Resources elements of the Actors–Resources– Activities model provide a good example of abstractions that have proved to be useful in B2B research. Activities, of course, correspond to events that involve them both. Moreover the addition of bonds,
124
G. Easton / Industrial Marketing Management 39 (2010) 118–128
links and ties as ways in which each has its own structures and relationships, integrates but does not con?ate separable components. Of course initial conceptualisations may change as the research progresses and clarifying what they are may be a key aspect of any research project. However, if theory is to progress then the basic entities and their powers and liabilities need to powerful enough in order to have some continuity of existence in terms of extant theory. The next step is collection of data. Case research is essentially eclectic with respect to the kinds of data that might be collected. Too often case research is equated with qualitative data collected by semi structured interviews. The strength of this method is that it is highly ?exible. However other forms of data collection, for example experimentation, can also work well in particular situations and provide insights not obvious in the more traditional modes of research. In any case the choice will be governed by what is thought to be required to establish a plausible causal mechanism, constrained by what data can actually be collected in the research context. Having collected the data the issue of its interpretation arises. Critical realists accept that there are differences between the empirical, the actual and the real, and that data are collected from people as well as from, and about, material things. As a result they accept that any explanations are necessarily fundamentally interpretivist in character. In particular when analysing respondent based data the researcher faces the problem of the double hermeneutic (Woodside et al., 2005). This adds another layer of complexity to the process since researchers are then required to include our understanding of the subjects' understandings. But while discourse is important and its interpretation crucial, reference to the referents of the discourse need to be made. In other words critical realists do not accept that studying the content and ways of expressing talk is enough in itself to provide explanation except of a very speci?c linguistic form. Seeking an explanation requires that the researcher goes back to the research site collecting more data until epistemological closure, however ?awed and temporary, is obtained. Since retroduction is the key epistemological process that critical realists recognise, understanding how it might work is important. Retroduction is a metaprocess the outcome of which is the identi?cation of mechanisms that explain what caused particular events to occur. Its adoption does not imply that the mechanisms are postulated then data collected or that they are “induced” from the event data. In practice the process is likely to be an iterative one (Dubois & Gadde, 2002). Case studies may employ both deductive and inductive cycles of data collection. Deduction helps to identify the phenomenon of interest, suggests what mechanism may be at play and provide links with previous research and literature. Induction provides event data to be explained and tests the explanations. Finally explanations invoke causal language and the identi?cation of mechanisms and offer the data collected as evidence. It differs from other research process cycles only in that its goals are different. The cutting edge of this method is to continue to ask the question why? Sayer points out that it is easy to make causal misattributions given the complexities of the systems we study and the possibility that different mechanisms can cause the same events. In addition there may be one or more mechanisms at work and the research design must be adapted to address both of these issues. The case method with its “cut and come again” disposition is ideally suited to the task. That a number of different causal explanations are put forward and researched is also not only possible but pragmatically desirable. Pawson and Tilley (1997), for example, suggest 8 different mechanisms that might account for the ?nding that placing CCTV cameras in car parks reduces car thefts and damage. Deciding which of these mechanisms appears to be working in the particular case then becomes necessary. Finally there is the issue of deciding whether your explanation is “good” or not. Clearly we would all hope that it is, and if not good then
simply the best that we can do in the circumstances. Another way of making this decision is by interpreting “good” as “acceptable”. For some critical realists the answer lies in the concept of “judgemental rationality”. “Judgemental rationality means that we can publically discuss our claims about reality as we think it is, and marshal better or worse arguments on behalf of those claims. By comparatively evaluating existing arguments, we can arrive at reasoned, though provisional, judgements about what reality is objectively like; about what belongs to that reality and what does not. ” (Archer, Collier, & Porpora, 2004, p2). 5. An example of a critical realist industrial marketing case analysis In the section that follows a necessarily truncated analysis of a case situation involving the purchase and implementation of a CRM system is described. 5.1. Black box case study Black Box Ltd is a small UK manufacturer and retailer of an extensive line of top end consumer electrical and electronics products. Until the late 90s they catered to the “technically competent” segment of the market but when it was realised that this was no longer viable they changed strategy and targeted the larger, less expert “plug and play” segment. At that time there was a small amount of ad hoc advertising, a once a year customer mailing was sent out and the company marketing information systems were mainly paper based. To implement their new strategy they extending their product line, redesigned their mail order catalogue, created a static website and began direct marketing. As a result from 2000 to 2003 sales grew steadily and Black Box struggled to keep up with the extra work that was involved. They invested in extra staff and more space assuming that sales would continue to rise. However, contrary to expectations, during 2004, sales plateaued and have remained steady since. In the mid 1990s, the main MIS was a standard PC accounting package which did the invoicing; everything else was on paper. In the late 90s they acquired their ?rst computer based mail order system, Mailstrom, and a transactional website to replace their simple brochure website. Then in early 2005 Black Box decided that they needed to purchase a CRM system. Their reasons for the purchase were that the current system was unreliable, slow in terms of customer response and didn't give them the customer information which would allow them to target customers more precisely. In addition there was the option to replace labour by capital and so reduce staff overheads. A decision was made to buy the CREOLE CRM system on the basis of its functionality and price offering rather than on its technical attributes. CREOLE allows Black Box to target customers and design particular promotions for particular market segments by means of promotional emails (e-blasts) and a twice monthly soft sell e-newsletter with new product information, tips and advice. In addition ordering on line is now possible. However the implementation of the CREOLE system was not without problems, which appeared to be of two kinds. The ?rst were software problems, for example two parts of the programme were supposed to provide the same information but didn't. The second and more important general problem involved the level and quality of service support from CREOLE. A key reason why Black Box had chosen CREOLE was that they had a good reputation for training and after sales service. This perception was probably due to CREOLE's policy of offering just enough training to get a client started and then providing tailored manuals to allow them to train themselves afterwards. In effect they were substituting cheaper tailored manuals for more expensive upfront training. In practice the training turned out to be largely perfunctory. The initial training lasted only a few hours with hand written notes as the sole record. In one case the onsite training was called off because of a
G. Easton / Industrial Marketing Management 39 (2010) 118–128
125
system failure and was never completed. Worse still the tailor made manuals arrived months after the system went live. Black Box staff wondered whether it would have been better to have a general manual to start with until the customised version was available. Even when the manuals turned up they didn't necessarily solve all the problems since they were incomplete. Another service option was CREOLE's help desk but this didn't always operate satisfactorily. Sometime after the installation, and clearly under pressure, CREOLE requested customers to email their help requests which would be acknowledged within 4 hours, prioritised and dealt with appropriately. Unfortunately the responses were often very late; the longest wait was 7 months. Black Box also acknowledged that there were problems with their personnel being unable or unwilling to use the system properly or to ?ag up its faults. In particular, the person responsible for co-ordinating the project didn't perform well. She continued to try to work in parallel with the old system and tried to solve all the problems herself. As a result other staff stopped complaining about speci?c issues and simply grumbled about the system and developed a generally negative attitude towards the change. The person concerned ultimately left the company. At this point the staff were told to report speci?c problems to their managers for forwarding to CREOLE. During this period CREOLE were involved in a UK government funded Knowledge Transfer Programme which was designed to help them improve their marketing performance. One of the elements of the programme was the institution of a Key Customer Forum to which senior staff from Black Box was invited. The forum provided an opportunity for customer staff to raise problems and issues and for CREOLE to listen and explain their situation and policies. A particular request for more self help information was made at the ?rst forum but this CREOLE seemed reluctant to offer. They argued that much of what they were being asked to provide was, in effect, internal systems redesign which was not what they were contracted to do. However more speci?c help was forthcoming when the Black Box staff met their CREOLE account manager, who they hadn't known of before. She was given a list of problems and questions, visited Black Box and successfully sorted out the former and answered the latter. Many of these queries were straightforward and could have been dealt with by the help desk. At this point CREOLE confessed to their failures, an admission that Black Box appreciated, and this served to foster a better relationship between the two companies. However it is clear that Black Box changed as a company as a result of the acquisition of the CREOLE system and staff now recognise how vulnerable that might make them due to their lack of IS expertise. “I think we all probably have a bit of a love hate relationship with computers.” This attitude has not preventing them from planning to get the most out of the CREOLE system and to extend and modify it to provide even better customer service in the future. 6. Case analysis 6.1. Research question The research topic is the creation and development of a business service relationship. One of the key research questions could be; What caused the CREOLE service problems as perceived and/or experienced by Black Box and what caused their subsequent solution? 6.2. Events The case describes a series of service interaction events that occurred between an MIS supplier and customer and the internal events within both parties that were linked to those interactions. The key interaction events/sequences involved training, delivery of the tailored manuals, help desk exchanges, the forum meeting between
Black Box staff and their CREOLE key account manager and her subsequent visit. The key internal events within Black Box were the attempts (not always unsuccessful) to make the CREOLE system do what it was supposed to do and the activities of their erstwhile project coordinator. The key internal events within CREOLE were the creation of the tailored manuals, the operations of the help desk, the management and enactment of the tailored manual policy and the appointment of the account manager. These events are partly sequential in nature (forum meeting then visit of the CREOLE account manager) and partly cyclical and reproductive (accessing the help desk). In the former case it is tempting to suggest that events cause events. However a critical realist account would be that as a result of an event the entities involved, and their powers and liabilities changed and this, in turn, caused the new event. 6.3. Identi?cation of entities Entities only attain meaning if they are embedded in a theoretical framework. Of course in exploratory research entities may be chosen then accepted, modi?ed or discarded as useful ways of try to understand data. However in this case a previously developed theoretical framework can, for the sake of an example, be employed; an economic exchange model (Easton, 2002). The entities involved are seller, buyer and exchange entities. There are necessary relationships among them all. An economic exchange cannot occur without a seller and a buyer and something that is exchanged. A seller has the power to sell and the buyer to buy and what is exchanged has power to affect both the buyer and seller. In this particular case the primary exchange events have already taken place but the extended exchange continues as after sales service. The service element means that what are exchanged are partly joint activities undertaken by the actors (calling the help desk) and partly physical entities (manuals). In this case the exchanges have continued for some time and an exchange relationship can be said to have come into existence. This relationship then becomes, of itself, an entity which has powers and liabilities which interact with those of both CREOLE and Black Box. For example the current atmosphere of the relationship is created by either or both and affects the relationship of each of them. The fourth entity is the Knowledge Transfer Programme (KTP) which had the power to in?uence the actions of CREOLE, not least in creating an event which had not only a major impact on the company but also in its relationship with Black Box. The relationship between the KTP and CREOLE is a contingent one since there is nothing necessary about its relationship to the buyer–seller relationships. The latter can exist without the former. However in this case the contingent powers of the KTP i.e. to bring into close, personal proximity CREOLE and Black Box personnel were activated to great effect. All three entities have important internal structures that affect events or, more properly, form part of a mechanism that cause those events to occur. For Black Box, internal structures include the CREOLE system as well as the other systems and procedures and the culture and personal characteristics of its employees. For CREOLE, the internal structures include the CREOLE CRM system, the help line resources, the number and expertise of its employees, its reputation for good service and the policy of tailored manuals and little training. The internal structure of the relationship includes the mutual expectations of the two actors, its service elements, the history of previous joint events, adaptations that each company made to each other and its atmosphere. Entities can be tangible or intangible, social or physical, dormant or active. The physical entities include, for example, the manuals and the output of the CREOLE code and data. The intangible entities include the reputation of CREOLE and the atmosphere of the relationship. In the case of the manuals this turns out to be a crucial aspect of the exchange since their form, either as general or tailor made manuals, had strong causal powers to affect events.
126
G. Easton / Industrial Marketing Management 39 (2010) 118–128
6.4. Possible causal mechanisms Three interconnected mechanisms can be described as being at work in this case. The one that is primary is the “perceived balance of responsibility” mechanism. There is always likely to be complex interaction between the nature and characteristics of a MIS system and the internal structures and processes of the recipient organisation. If both were simple and easily understood it might be possible to write a contract specifying exactly what the responsibilities of both parties would be. However this is rarely the case so that there is always likely to be uncertainty and ambiguity about who is responsible for what parts of the implementation. However each actor made assumptions about what responsibilities the other had in making the CREOLE system work as well as what the criteria for working actually were. Black Box had high expectations that CREOLE would be largely responsible for the whole process. This was in turn caused by their perceptions of CREOLE's reputation and their own inexperience in MIS implementation. On the other hand CREOLE was attempting to enact a strategy that systematised their service through the use of tailored manuals which they hoped would allow Black Box to solve more of their own implementation problems. This strategy was most probably caused by a lack of service resources at the time but was also an attempt to provide an acceptable mid point in terms of balance of responsibility for both parties to the exchange. The second mechanism was the “enacted balance of responsibility”. CREOLE completed neither the basic level of training nor the delivery of the training manuals at the required time. In addition the help line was unable to ?ll the resulting gap, again most likely as a result of inadequate resources and using a response system designed to minimise the use of those resources. And the software had some remaining bugs. Black Box, on its part, had a project coordinator who was trying to continue use of existing systems and rejecting help, uncoordinated activities in terms of managerial control of the process and naïve and unrealistic expectations. The timely resolution of the problems was, in part, a chance contingent event. If CREOLE and Black Box personnel had not met at the forum then any number of possible scenarios could have been envisaged. The crucial point about these events was that they suggest a third related “key in lock” mechanism. In practice the help required was relatively trivial in nature, solved many of the remaining problems and restored the relationship between CREOLE and Black Box. The key in lock metaphor describes the situation where a particular con?guration of information allows a door to be “opened” and a system to operate as expected. This metaphor depends on the knowledge that by the time the forum happened, Black Box had changed their coordinator and taken steps to control the implementation and so were ready to go. What remained were technical issues that CREOLE could show them how to solve relatively easily. Fig. 3 below gives a partial summary of one of the sets of events which took place and which are described in the case. It gives examples of some
of the powers, liabilities and structures of the entities which operate through one of the proposed mechanisms to create another rather different set of events than those experienced previously. It suggests an answer to the question “Why did the exchanges between CREOLE and Black Box become more helpful and productive after a particular point in time?” More complex explanations of the previous situations could also be constructed using the concept of perceived and actual balance of responsibility as key components. The returns from carrying out the above analyses are therefore three fold. Firstly they have, hopefully, provided a means of improving our conceptualisation of some of the key entities in B2B research, buyer, seller, exchange entities, in terms of how they relate to one another. Secondly they suggest three possible mechanisms that might occur in other buyer–seller situations. Thirdly they provide an example of the way in which a critical realist approach might be employed when analysing case studies. 7. Theory development and generalisability The knowledge claims of case study research are often attacked on the grounds of lack of generalisability. Most cases reported are single cases. The argument runs thus: we can accept that you have made a very convincing job of explaining the rather complex situation you have been researching, but how do we know your case situation was representative?. How can you generalise? Yin, in the ?rst edition of his book on case research, suggests, “The short answer is that case studies, like experiments, are generalisable to theoretical propositions and not to populations or universes. In this sense, the case study, like the experiment, does not represent a ‘sample’, and the investigator's goal is to expand and generalise theories (analytical generalization) and not to enumerate frequencies (statistical generalization)” (Yin, 1989, p.21). How this analysis should take place or in what sense generalisation is possible, Yin did not make clear. From a critical realist perspective, expansion and generalisation come from identifying the deep processes at work under contingent conditions via particular mechanisms. A causal explanation in a single case must be based upon a theory structured in terms of what comprises a critical realist causal explanation. The best explanation, that is the one most consistent with the data, is what is being sought. However generalisation of any kind is not possible unless there is some invariance in the world. If all events and their causes are unique then there could never be theories that work. And by unique I mean substantially and not trivially unique since in some sense every event in the world is unique. Clearly it is not possible to explain every event in all its detail. It would be suf?cient if it could be said what, in general, caused this MIS to be implemented in this particular form in these circumstances and to ignore apparently unimportant details. Invariance for a critical realist relates to elements of what they claim are the constituents of a causal explanation. Sayer puts it this way “Structures can therefore be said to be ‘invariant under certain
Fig. 3. Outline summary of the operation of a possible mechanism.
G. Easton / Industrial Marketing Management 39 (2010) 118–128
127
transformations’, that is, they can continue to exist while their constituents undergo changes in attributes which are not relevant to their reproduction” (Sayer, 1992 p.94). If a defensible causal explanation has been produced in one case then the constituents of that explanation provide a basis for developing theory beyond that case. This means, ?rstly, that the identi?cation of the entities concerned should been useful in terms of further theory development. Here the entities which are of most interest in the long term for industrial marketing research are the selling organisation and the buying organisation and their relationship. In other situations the net or network might be central. However these entities will usually be at too general a level to explain interorganisational events and the entities that form their internal structures, for example organisational form and culture, will be of more use. The structures of the bodies have been elucidated to some extent and relations among them established. In the Black Box case the relationship between the project coordinator and the rest of the organisation was a crucial internal structure. Secondly, the powers and liabilities of entities need to be studied empirically. For example these can, even in this relatively simple case, be hypothesised for an MIS and might include the power to transform customer relationships and the liability to destroy the organisation into which it has been placed, the power of CREOLE to withhold training resources and the liability of Black Box to accept CREOLE's training system are examples in the Black Box case. Thirdly the necessary relations have to be con?rmed, as was the case between CREOLE and Black Box, since events could not have been explained unless the relationship between the two ?rms is implicated. Also contingent relations have been shown to operate such as the impact of the KTP forum. It might be expected that contingent relations would be the most dif?cult to deal with since they are do not always affect events. For example in the Black Box case the KTP was a crucial element in explaining the events but since it was a time limited organisation it is hardly likely to be of much use in furthering theory. However if the focus of the research was change agencies and whether they were successful then it would have a necessary relationships with those organisations and might form the basis for a theory about organisational change. Fourthly the evidence appears to support the notion that certain mechanisms, the central components of a critical realist theory, have been shown to operate; the perceived and enacted balances of responsibility between CREOLE and Black Box. It would certainly be possible to think of other mechanisms that might apply here and these could be induced from the data or deduced from other theories or by means of creative thought. Some phenomena can be explained by building from unique mechanisms and contingent variables. Others occur (perhaps closer to the natural than the social world) that are explicable in terms of rather uniform causal mechanisms. However most phenomena can probably be explained by a mixture of the general and the speci?c. In this situation the case method has the task of uncovering the structure of underlying reality in each case, investigated to see what can be added to theory in the way of con?rmation and articulation. A second form of case development is when more than one case is used. A key decision in case research is what number of cases to research. The following conditions suggest when a small, number, perhaps one, rather than many would be wise. Where there is heterogeneity, or even uniqueness, of events then any kind of constant conjunction is unlikely to be detectable and seeking a credible casual mechanism is the only way to advance theory. Such heterogeneity may however be caused by the impact of a large number of powerful and active contingent relations and so working through the particular ways in which they operate in a single case will be not be as rewarding as using several cases. Explanations can be more or less “deep” and “wide”. They can incorporate a number of different emergent levels (individual, group and organisation) and a number of different entities. In general deeper and wider explanations are to be preferred in order to promote
theory development in the ways described above but there is always likely to be a trade-off with available research resources, access and theoretical complexity. Relationship with existing theory is important. Where little exists then one case can be enough to begin the process of theory creation. Where there exists well-articulated theory, particular aspects of that theory, the entities, their powers, the nature of the relationships and the overall mechanisms can be targeted and attempts made to elucidate one or all of them in a single case. More generally, a pragmatic approach can be adopted. If the objective is to advance theory then one should be able to say which theoretical aspects one wishes to address and how will that be served by undertaking one case study compared with many. 8. Managerial implications Research into management decision making suggests that managers often rely on gut feel or in academic terms, intuition. Since many, if not most, decisions have to be made quickly they have little time to think through, from ?rst principles, what the real issues are, what options are available and what value to be placed on each of those options. They rely on experience which, in practice, means recognition primed decision making (Klein, 1998). They recognise a particular pattern in the data, usually events, and so take the action that has worked for them, most of time in the past. For example if a new buyer asks for price reduction from a marketing manager she has usually found that it is best to agree, after vigorously arguing about it. Then, some months later, she asks for a price increase on “rational” grounds and it will be accepted most of the time. Of course intuition also means drawing on other experiences when the pattern isn't quite normal. The managerial implications of a critical realist approach to decision making is that managers should be encouraged to think about why certain decisions lead to certain outcomes; to try to discover what causes them or in critical realist terms what the causal mechanisms are. They may be forced into doing this anyway when a novel situation arises (selling into a poor African country) and they have to carry out a real time case analysis, collecting data and trying to ?gure out if the usual strategies will or won't work. However there is also a case for managers thinking about their intuitive powers. Experience has taught them that certain actions have good outcomes. But why? If they knew that, then they might be more sensitive in detecting possible rogue situations where normal actions probably won't work. They might also be able improve on existing decision making if they understood the causal mechanisms that were in play. For example in the new buyer situation mentioned earlier it might be that the buyer is trying to impress new colleagues by getting a price reduction. A few months later they are either busier or have less need to impress and so don't worry about a price increase. But which is it most likely to be the case and what are the appropriate tactics? That would depend on a number of factors that could be investigated, albeit rather subtly. It would be naïve to argue that managers should build critical realist causal models. However even if they simply try to employ causal language and, at the same time, become more inquisitive about the situations they face there could be real bene?ts to be gained. 9. Conclusions Case study research is the prevalent research method in B2B research. However it is often justi?ed, when it is justi?ed at all, in terms of the interesting results or unusual phenomena it reveals. What it lacks is philosophical validation, i.e. ontological and epistemological underpinnings. I argue in this paper that critical realism provides such underpinnings and seems ideally matched to case research. Certainly case research cannot be justi?ed in terms of positivism since case research is almost always small numbers
128
G. Easton / Industrial Marketing Management 39 (2010) 118–128 Dickens, P. (2003). Changing our environment, changing ourselves: Critical realism and transdisciplinary research. Interdisciplinary Science Reviews, 28(2), 95?105. Dubois, A., & Gadde, L. -E. (2002). Systematic combining: An abductive approach to case research. Journal of Business Research, 55, 553?560. Easton, G. (1995). In D. T. Wilson & K. Moller (Eds.), Methodology and industrial networks in relationships and networks theory and application Norwell, Mass: Kluwer Academic Publishers. Easton, G. (1998). Case research as a methodology for industrial networks: A realist apologia. In P. Naude & P. W. Turnbull (Eds.), Network dynamics in international marketing Oxford: Elsevier Science. Easton, G. (2002). Marketing: A critical realist approach. Journal of Business Research, 55(2), 103?109. Easton, G. and Araujo, L. Language, metaphors and networks in advances in international marketing, (5), 67–85. ed., Deo Sharma, New York, JAI Press, 1993. Eisenhardt, K. M. (1989). Building theories from case study research. Academy of Management Review, 14(4), 532?551. Gerring, J. (2004). What is a case study and what is it good for. American Political Science Review, 98(2), 341?354. Hanson, J., & Yosifon, D. (2004). The situational character: A critical realist perspective on the human animal. Georgetown Law Journal, 93(1), 1?179. Hanley, C. (1995). On facts and ideas in psychoanalysis. International Journal of Psychoanalysis, 76(5), 105?136. Houston, S. (2001). Beyond social constructionism: Critical realism and social work. British Journal of Social Work, 31(6), 845?861. Keat, R., & Urry, J. (1975). Social theory as science (pp. 9?12). London: Routledge and Keagan Paul. Hunt, S. D. (2003). Controversy in marketing theory: For reason, realism, truth, and objectivity. Armonk, N.Y: M.E. Sharpe. Kendall, J. E., & Kendall, K. E. (1993). Metaphors and methodologies: Living beyond the systems machine. MIS Quarterly, 17(2), 149?162. Klein, G. (1998). Sources of power: How people make decisions. Boston: The MIT Press. Lau, R. W. K. (2004). Critical realism and news production. Media, Culture and Society, 26(5), 693?711. Lawson, T. (1997). Economics and reality. London: Routledge. Layder, R. (1990). The realist image in social science. London: Macmillan. Lewis, P. A. (1996). Metaphor and critical realism. Review of Social Economy, 54(4), 487?506. Mitchell, J. C. (1983). Case and situation analysis. The Sociological Review, 3(1), 187?211. Nellhaus, T. (1998). Signs, social ontology, and critical realism. Journal for the Theory of Social Behaviour, 28(1), 1?25. Pawson, R., & Tilley, N. (1997). Realistic evaluation. Sage: London. Proctor, J. D. (1992). The social construction of nature: Relativist accusations, pragmatist and critical realist responses. Annals of the Association of American Geographers, 88(3), 352?376. Robbins, J. W. (1999). Pragmatism, critical realism, and the cognitive value of religion and science. Zygon, 34(4), 655?666. Sayer, A. (1984). Method in social science. London: Routledge. Sayer, A. (1992). Method in social science: A realist approach, (2nd ed.) London: Routledge. Sayer, A. (2000). Realism and social science. London: Sage. Steinmetz, G. (1998). Critical realism and historical sociology; A review article. Comparative studies in society and history, 40(1), 170?186. Verschuren, P. J. M. (2003). Case study as a research strategy: Some ambiguities and opportunities. International Journal of Social Science Research Methodology, 6(2), 121?139. Trosper, R. L. (2005). Emergence unites ecology and society. Ecology and Society, 10(1) Article 14. Tsoukas, H. (1989). The validity of idiographic research explanations. Academy of Management Review, 14(4), 551?561. Wikgren, M. (2005). Critical realism as a philosophy and theory in information science? Journal of Documentation, 6(1), 11?22. Woodside, A. G., & Wilson, E. J. (2003). Case study research methods for theory building. Journal of Business and Industrial Marketing, 18(6/7), 493?508. Woodside, A. G., Pattinson, H. M., & Miller, K. E. (2005). Advancing hermeneutic research for interpreting inter?rm new product development. Journal of Business and Industrial Marketing, 20(7), 493?508. Yeung, H. W. C. (1997). Critical realism and realist research in human geography: A method or a philosophy in search of a method? Progress in Human Geography, 21(1), 51?74. Yin, R. K. (1989). Case study research: Design and methods. Newbury Park: Sage. Yin, R. K. (2003). Case study research: Design and methods. Newbury Park: Sage. Geoff Easton is a professor in the Department of Marketing at Lancaster University. His main research interests are B2B markets and marketing, organisational and managerial behaviour, strategic practices, critical realism and case teaching methods. He has published in Industrial Marketing Management, the Journal of Management Studies, OMEGA, British Journal of Management, European Journal of Marketing, Organization, Journal of Business Research, Journal of World Business and Complex Systems Advances.
research. Interpretivism is more relevant but is largely epistemological in its objectives. Critical realism however provides not only a basis for justi?cation but also guidelines as to how case research might be done and how theory can be fashioned. Critical realism ?rst of all makes the ontological assumption that there is a reality but that it is usually dif?cult to apprehend. It distinguishes between the real world, the actual events that are created by the real world and the empirical events which we can actually capture and record. Thus we will always be surmising about the nature of the real. Critical realists argue that in the real world there are entities, such as organisations, which have powers to act and are liable to be acted upon by others. These entities can also have internal structures, such as departments and individuals which in their turn, have their own powers. Entities have relationships among themselves which are de?ned as necessary in terms of our understanding of their powers, for example buyers and sellers have necessary relationships by virtue of the designations that are given to those who act by buying or selling. They may also have a contingent relationship, that is, they may or may not have such relationships. For example buyers and sellers may be connected to the same EDI system or they may not and so the EDI system is in a contingent relationship to the buyer and seller. The relationship among entities that is thought to cause the events is called a mechanism. A critical realist approach to case research involves developing a research question that identi?es a research phenomenon of interest, in terms of discernible events, and asks what causes them to happen. The key entities involved, their powers, liabilities, necessary and contingent relationships are then provisionally identi?ed. Research then proceeds by capturing data with respect to ongoing or past events asking at all times why they happened or are happening and taking into account the problems and issues associated with interpreting the empirical data back to the real entities and their actions. The research process is one of continuous cycles of research and re?ection. The ?nal result is the identi?cation of one or more mechanisms that can be regarded as having caused the events. This process was exempli?ed via a real life case study, the research for which was carried out in this way, and 3 possible mechanisms were identi?ed at work in a post purchase service relationship between an MIS provider and a buyer. Finally it was suggested that generalisation to theory via case research carried out under critical realist conventions occurs by virtue of clarifying the theoretical nature of the entities involved, the ways in which they act and the nature and variety of mechanisms through which they exert their powers or acted upon by other entities. As with all philosophical approaches critical realism cannot be proved to be the “right answer”. Acceptance of a critical realist approach depends on whether you agree with its basic assumptions and that acceptance can be based on any number of things. In my own case it I accept it because (a) I think that this is how the world is, (b) even if it isn't like that, I behave as if it is like that, (c) I think that critical realism is better than the alternatives and (d) it is a well thought through and relatively coherent perspective on the world.
References
Ackroyd, S., & Fleetwood, S. (2004). Critical realist applications in organisation and management studies. London: Sage. Archer, M., Collier, A., & Porpora, D. V. (2004). Transcendence. London: Routledge. Bania, R. (1995). Critical realism and urban and regional studies. Environment and Planning B—Planning and Design, 22(5), 563?580. Bhaskar, R. (1978). A realist theory of science. Hemel Hempstead: Harvester Press.
doc_295619369.pdf