اعتبار سنجی در تحقیقات تفسیری حسابداری مدیریت

This paper extends and contributes to emerging debates on the validation of interpretive research (IR) in management accounting. We argue that IR has the potential to produce not only subjectivist, emic understandings of actors’ meanings, but also explanations, characterised by a certain degree of “thickness”. Mobilising the key tenets of the modern philosophical theory of explanation and the notion of abduction, grounded in pragmatist epistemology, we explicate how explanations may be developed and validated, yet remaining true to the core premises of IR. We focus on the intricate relationship between two arguably central aspects of validation in IR, namely authenticity and plausibility. Working on the assumption that validation is an important, but potentially problematic concern in all serious scholarly research, we explore whether and how validation efforts are manifest in IR using two case studies as illustrative examples. Validation is seen as an issue of convincing readers of the authenticity of research findings whilst simultaneously ensuring that explanations are deemed plausible. Whilst the former is largely a matter of preserving the emic qualities of research accounts, the latter is intimately linked to the process of abductive reasoning, whereby different theories are applied to advance thick explanations. This underscores the view of validation as a process, not easily separated from the ongoing efforts of researchers to develop explanations as research projects unfold and far from reducible to mere technicalities of following pre-specified criteria presumably minimising various biases. These properties detract from a view of validation as conforming to pre-specified, stable, and uniform criteria and allow IR to move beyond the “crisis of validity” arguably prevailing in the social sciences.

The nature and status of interpretive research (IR) in management accounting have recently been the subject of considerable debate (e.g., Ahrens, 2008, Ahrens et al., 2008, Kakkuri-Knuuttila et al., 2008a and Kakkuri-Knuuttila et al., 2008b). This debate suggests that this research genre has truly come of age and affirms the broad scope of what currently counts as “good” IR (see also Baxter & Chua, 2003). However, little attention has been paid to the potentially critical issue of how such research should be validated and how interpretive management accounting researchers go about this.1 This relative neglect of the notion of validation is perhaps not surprising given the initial reluctance of interpretive researchers to debate the issue (Bloor, 1978) and the tendency of extreme post-modernists to dismiss it as just another indication of the “chains of modernism” (see Koro-Ljungberg, 2004). However, the starting point for our analysis is the assumption that validation of some kind is a necessary condition for any scholarly research endeavour to be taken seriously (see e.g., Koro-Ljungberg, 2004, Polkinghorne, 2007, Sandberg, 2005, Smith, 2006, Smith and Deemer, 2000 and Tsoukas, 1989). In a very broad sense, validation refers to the ways through which the credibility of a piece of research is developed and legitimised in front of relevant audiences ( Lather, 1993, Lincoln and Guba, 2000, Polkinghorne, 2007 and Silverman, 2000). The relevance of addressing the issue of validation in IR is underscored by critical debates on this topic in the social sciences, which have not yet been fully recognised in the management accounting literature. It has been argued that IR suffers from a “crisis of validity” (e.g., Gergen and Gergen, 2000 and Smith and Deemer, 2000). The core of this argument is that although there is some consensus that IR cannot be validated with traditional validation methods, it is still unclear how it should be validated. This crisis is arguably exacerbated by the continuing quest for alternative and presumably stable validity criteria (see, e.g., Creswell and Miller, 2000, Maxwell, 1992 and Onwuegbuzie and Leech, 2007), sometimes under other overriding labels such as “trustworthiness” (Lincoln & Guba, 1985). A number of critics have argued that this signifies a failure to emancipate the larger IR project from a foundationalist (or positivist) epistemology (Garratt and Hodkinson, 1998, Schwandt, 1996, Smith and Deemer, 2000 and Smith and Hodkinson, 2005). This is particularly problematic if one subscribes to a strongly subjectivist position, such as that characterising Burrell and Morgan’s (1979) widely cited portrayal of IR. Such a position is inconsistent with traditional validity criteria such as construct, internal and external validity as well as reliability, which assume that there is a reasonably stable reality “out there” to be captured independently of the perceiving subject (e.g., Gergen and Gergen, 2000, Koro-Ljungberg, 2004, Koro-Ljungberg, 2008, Lather, 1993 and Smith and Deemer, 2000). How could we, and why should we, even try to capture the singular essence of the world, if we first assume there is none? Whilst the debate surrounding the “crisis of validity” in IR has not been explicitly recognised in the management accounting literature, concerns have recently been raised that researchers propounding to work within this genre are not always consistent and rely on validation procedures typically considered, in principle, problematic in such research (Ahrens and Chapman, 2006 and Elharidy et al., 2008).2 This is an indication of a “crisis of validity” looming in interpretive management accounting research as well and suggests that the issue of how such research may be validated requires more focused attention. Although validation is potentially problematic in all types of research, including the positivist “mainstream”, management accounting scholars need to be alert to the implications of recent debates about how IR may be validated. Whilst the present institutional and political environment in parts of the accounting academe exhibits strong pressures for conformity with this “mainstream” position (Hopwood, 2007, Hopwood, 2008, Khalifa and Quattrone, 2008 and Tuttle and Dillard, 2007), it is worth noting that attention to the “crisis of validity” has emerged within IR circles and should be viewed as critical self-reflection within this paradigm. This should be regarded as a healthy phenomenon as a capacity for self-reflection, even critical at times, indicates a certain maturity of the genre of research in question. It is certainly not our ambition to undermine the larger IR project in management accounting by just reiterating this critical debate. We argue instead that the position of IR may be strengthened if researchers were more explicit about the notion (and potential problems) of validation and how it may be extended in light of recent debates. But what kind of position can be justified in interpretive management accounting research regarding validation? This is the major concern in this paper. Most importantly, the debate following the “crisis of validity” in IR has challenged the notion of validity as little more than a technical issue of reducing bias in data collection and analysis and directs attention to the relationship between authors/researchers and their audience (cf. Baxter and Chua, 2008 and Golden-Biddle and Locke, 1993). Hence, rather than carrying on the quest for tightly specified and stable validity criteria, we will explore how IR in management accounting is currently understood and practised and how the notion of validity can be inter-subjectively negotiated within a particular research community. Following Kakkuri-Knuuttila et al. (2008a), we extend the somewhat caricatured portrayal of IR as being mainly about describing social phenomena in order to convey a rich, in-depth understanding of the meanings attached to them by researched individuals to also encompass an important explanatory element. Even though some commentators suggest that such a position is already widely accepted in contemporary IR in management accounting (see Ahrens, 2008), the extent to which this is actually recognised is still unclear (Kakkuri-Knuuttila et al., 2008b). In particular, the discussion above suggests that it is not yet entirely clear how the validity of explanations can be established in IR, whilst staying reasonably true to the core premises of such research. The purpose of this paper is to shed further light on this issue by revisiting the critical debate on validation in IR and explicating how it can be extended by mobilising central tenets of the modern philosophical theory of explanation. We will argue that, contrary to what is typically stated within IR circles, IR can produce – whilst being by definition anchored in them – much more than just subjectivist emic understandings of actors’ meanings. By focusing on careful analyses of the sequences of actual events and actions in specific, local contexts, it is in a good position to trace the dependencies (causal linkages) between examined phenomena. As these dependencies tend to be more than just individuals’ subjective mental states, IR can produce explanations from a more external viewpoint. Since these explanations are developed on the basis of profound emic understandings, we call these explanations thick. Our basic position is similar to that of Kakkuri-Knuuttila et al. (2008a) as we argue for an integration of social constructionism with a moderate form of realism. Consistent with social constructionism, we affirm the inescapable role of judgement and inter-subjectively inferred “truths” on the part of researchers in making sense of situated meanings. This allows for the acknowledgement of emic understandings of people’s meanings. However, following the leads of moderate realism, we take issue with a view of the world as simply reducible to ever-changing subjective experiences and argue that IR may also entail an important etic element aimed at generating explanations. We extend the work of Kakkuri-Knuuttila et al. (2008a) in two ways. Firstly, ideas drawn from pragmatism inform us to further clarify the possibility to integrate social constructionism with a certain kind of realism. Secondly, we apply these background understandings by examining the question as to how IR may be validated. More specifically, we discuss the intricate relationship between two central aspects of validation in IR, namely those of authenticity and plausibility. Authenticity lies at the core of validating the defining elements of any IR research, namely rich descriptions, whilst plausibility is relevant for assessing the credibility of the explanations being developed. The paper is structured as follows: as there seem to be differing notions of IR within management accounting academe, we will first identify some variations within this research genre with particular reference to the notion of explanatory IR. Thereafter, we will deal conceptually and philosophically with the problems and possibilities of validating the various forms of IR. After that we will illustrate our analysis by examples from published management accounting studies viewed or claimed to be interpretive by nature. The concluding section summarises our discussion and its implications for future IR in management accounting.

This paper contributes to and extends emerging debates seeking to clarify the nature of IR in management accounting (Ahrens, 2008, Ahrens et al., 2008, Kakkuri-Knuuttila et al., 2008a and Kakkuri-Knuuttila et al., 2008b). In addressing the relatively neglected issue of how such research can be validated, we draw attention to its capacity to include both rich, emic accounts, grounded in profound understandings of the meanings of the researched, and explanations entailing an important etic and theoretically informed element. Whilst preserving the former property is a pre-requisite for IR to be deemed authentic, we have paid particular attention to how thick explanations, emerging from emic accounts, are rendered plausible to particular research communities. Consistent with recent advances portraying validation as a more or less ongoing process (e.g., Koro-Ljungberg, 2004 and Koro-Ljungberg, 2008), we argue that the impression of authenticity can be enhanced through the construction of relatively holistic accounts giving voice to the “Other”. Whilst this may entail some trade-offs with the explanatory element of IR, the process of generating and validating explanations from emic accounts may be facilitated by the mobilisation of key tenets of the modern philosophical theory of explanation (counterfactual conditionals and contrast class) in combination with abductive reasoning, grounded in pragmatist epistemology. However, it is important to bear in mind that such pragmatist premises ultimately lead to a view of validity as embedded in the belief systems guiding particular research communities. Such belief systems may themselves be in a state of flux (Greckhamer et al., 2008). This cautions us against notions of validity as a question of establishing stable epistemological guarantees (Koro-Ljungberg, 2004, Koro-Ljungberg, 2008 and Lather, 1993). Although pragmatist thinkers reject the notion of “truth” as infinitely open to negotiation, we are reluctant to conceive of the concepts of authenticity and plausibility as immutable validity criteria, but note that also they are open to the constant possibility of re-assessment. By way of implications, a useful starting point for researchers applying the ideas advanced in this paper is to adopt a mindful and reflexive approach recognising how the belief systems dominating the research communities of which they are part condition the scope for advancing particular knowledge claims. The empirical examples reviewed in this paper offer some clues to how we may go about validating research findings based on this premise. One strategy, similar to that adopted by Covaleski and Dirsmith (1986), is to mobilise theoretically informed explanations which may not yet be fully accepted by the wider research community by appealing to the need to give voice to the “Others” who remain under-represented by established theories. By framing such explanations in an abductive mode of reasoning, contrasting them with more entrenched explanations, researchers may allow for multiple “truths” to emerge whilst clearly demonstrating how such “truths” are theory-related and not reducible to a single way of representing the world. Combining such a strategy with systematic mobilisation of the notions of contrast class and counterfactual conditionals, researchers may convince readers of the plausibility of particular explanations. However, this needs to be balanced with the emic element of research to avoid sacrificing too much of the authenticity of representations. One way of doing this is to recognise the partiality of any account generated through the mobilisation of a particular set of theories. This may compel researchers to remain open to the potentially paradoxical aspects of organisational life and the possibilities this offers to further the impression of authenticity. An alternative strategy, which is perhaps more applicable when theories are firmly entrenched within particular research communities, is to take these as a rather general starting point and rely heavily on rich, emic accounts whilst leaving explanations somewhat open-ended and thus affirming the messy and often paradoxical nature of life in complex organisations. This is essentially the approach adopted by Vaivio (2006) and it places a premium on the authenticity of representations. There is a risk that accounts emanating from such an approach amount to little more than rich descriptions, without offering much by way of plausible explanations. As demonstrated by Vaivio (2006), however, it is possible to combine rich, emic accounts with an explanatory element although explanations are perhaps mobilised as more seamless parts of broader case narratives. Although this may include an abductive element, it tends to be less systematically linked to the notions of contrast class and counterfactual conditionals to convince readers of the plausibility of explanations. This suggests that the process of validation is not always easily separable from that of developing explanations in case narratives. As should be clear, the key difference between these two approaches lies in the use of theories in convincing readers of the validity of representations. The first approach relies on less well-established theories as sensitising devices for alerting research communities to the risk of particular “truths” being under-represented as well as fostering researcher reflexivity. By contrast, the second approach operates in a more “given” theoretical space within which validity is mainly negotiated through emic accounts. However, both approaches recognise the theory-relatedness of knowledge claims emerging from empirical studies and underline the notion of validation as a process of negotiating the truth of such claims within particular research communities. In contrast to positivist approaches to validating qualitative management accounting research (e.g., Atkinson and Shaffir, 1998, Lillis, 1999 and McKinnon, 1988), this implies a view of validation as far from reducible to mere technicalities of following pre-specified criteria presumably minimising various observational biases. By emphasising the notion of validation as a theory-related and genuinely relativist exercise, we have started to address emerging concerns suggesting that a “crisis of validity”, similar to that observed in the wider social sciences, may be looming in interpretive management accounting research. Adding to the debate following this observation, we extend the notion of validation in IR by describing how it may entail an important explanatory element. This is consistent with the view that much IR in management accounting effectively straddles between the functionalist and interpretive paradigms (Kakkuri-Knuuttila et al., 2008a), We have thus taken a first step in moving interpretive management accounting research beyond the “crisis of validity” whilst recognising recent attempts to re-conceptualise this research genre and challenge others to follow.