در دفاع از مثلث بندی: یک رویکرد واقع گرای انتقادی به روش مختلط پژوهشی در حسابداری مدیریت

The notion of triangulation constitutes a key component of mixed methods research but has been contested on ontological and epistemological grounds, especially where this entails integration of theories and/or methods rooted in different philosophical assumptions (or paradigms). Drawing on critical realism, this paper addresses two criticisms of the use of triangulation in mixed methods research straddling between the functionalist and interpretive paradigms, namely (1) its propensity to suppress variations in situated meanings and (2) its treatment of empirical observations as objectively verifiable rather than inherently theory-related. The modified notion of triangulation advanced in this paper counters these criticisms by re-conceptualizing it as firmly grounded in abductive reasoning. This provides a foundation for maintaining researchers’ sensitivity to context-specific variations in meanings in efforts to derive theory-related explanations. The possibilities of using such a modified notion of triangulation in management accounting research are illustrated through a review of two empirical studies straddling between the functionalist and interpretive paradigms.

Mixed methods research, combining qualitative and quantitative approaches, has a long history in management accounting (e.g., Hopwood, 1973 and Otley, 1978) and has recently attracted increasing attention (Anderson and Widener, 2007, Brown and Brignall, 2007, Hopper and Hoque, 2006, Lillis and Mundy, 2005, Modell, 2005 and Modell, 2007). Although the majority of this research tends to be classified into the functionalist paradigm, there is evidence of growing interest in combining theories and methods generally associated with different paradigms (Brown and Brignall, 2007, Hopper and Hoque, 2006 and Modell, 2005). Recent debates on qualitative management accounting research also suggest that such paradigm-straddling approaches, combining elements of the functionalist and interpretive paradigms, are indeed more widespread and perhaps accepted than previously recognized (Ahrens, 2008, Kakkuri-Knuuttila et al., 2008a and Kakkuri-Knuuttila et al., 2008b). At the same time, however, there are growing concerns with the lack of recognition and communication across these paradigms resulting from the fragmentation and polarization of accounting academia (e.g., Ahrens et al., 2008, Hopwood, 2007, Hopwood, 2008, Khalifa and Quattrone, 2008 and Reiter and Williams, 2002). This suggests that, however desirable from the point of view of furthering our understanding of management accounting (cf. Gioia and Pitre, 1990), there may be considerable barriers to combining different paradigms in actual research practice. The tensions associated with straddling between paradigms are readily observable when one considers the issue of triangulation between theories and methods rooted in different paradigms. While triangulation is generally seen as a vital validation technique in mixed methods research ( Brewer and Hunter, 1989, Erzberger and Kelle, 2003 and Erzberger and Prein, 1997), some scepticism of such practices has emerged in the management accounting literature. The notion of triangulation originates in positioning geometry, where one's position in space is determined in relation to some stable and objectively verifiable reference points. Several authors have questioned the usefulness of this metaphor as a means of more ‘accurate’ positioning of research findings where reality is seen as changeable and socially constructed ( Ahrens and Chapman, 2006, Llewellyn, 2007, Tomkins and Groves, 1983a and Tomkins and Groves, 1983b). Similar criticisms have been raised in the wider mixed methods literature in the social sciences and suggest that triangulation is not a legitimate validation technique where research straddles between the functionalist and interpretive paradigms (e.g., Blaikie, 1991 and Blaikie, 2000). These philosophically tuned criticisms of triangulation have not yet been satisfactorily addressed by advocates of mixed methods research in management accounting. Discussions of such research have either tended to be mute about its philosophical underpinnings (but see Brown and Brignall (2007) for an exception) or premised on the assumption that the choice of theories and methods may be emancipated from their ontological and epistemological roots (Modell, 2005). This neglect of the philosophical foundations of triangulation is potentially detrimental to the status of mixed methods research in management accounting as it may lead opponents of such research to reject it as excessively eclectic or resting on inconsistent philosophical assumptions (Modell, 2007). This paper addresses this problem by advancing a critical realist approach to mixed methods research in management accounting. Critical realism has recently been advocated as a potential way of bridging the polarized positions of the functionalist and interpretive paradigms in organization and management studies (e.g., Ackroyd and Fleetwood, 2000, Fleetwood, 2004, Fleetwood, 2005, Reed, 2005a and Sayer, 2004) and has gradually found its way into the mixed methods literature in management accounting (e.g., Brown and Brignall, 2007). A key premise in this regard is that critical realism, unlike more ‘naïve’ forms of empirical realism,1 accepts the existence of some reasonably stable and mind-independent reality but rejects the possibility of verifying research findings in any absolute or ‘objective’ sense. This position partly mirrors the criticisms of triangulation outlined above and underlines the potential usefulness of critical realism for dealing with these. However, critical realism has not yet been systematically applied to examine how a modified notion of triangulation may be advanced in mixed methods research in management accounting. The remainder of the paper and my over-riding line of argument are structured as follows. Given the often ill-specified meanings attributed to triangulation, I start by briefly reviewing the use of this concept in functionalist management accounting research and elaborate on the target of criticism in Section 2. I also examine how mixed methods researchers subscribing to pragmatist lines of thought have responded to these criticisms before advancing a critical realist approach to triangulation in Section 3. While there is some resemblance between these pragmatist and critical realist approaches, I argue that the latter provides a more clearly articulated foundation for modifying the notion of triangulation in response to objections to its use in mixed methods research straddling between the functionalist and interpretive paradigms. This modified notion of triangulation rests heavily on abductive reasoning as a means of deriving theoretically informed explanations while preserving researchers’ sensitivity to variations in situated meanings. To illustrate how such a notion of triangulation may be applied in practice, I review two management accounting studies in Section 4 (Covaleski and Dirsmith, 1983, Hoque and Hopper, 1994 and Hoque and Hopper, 1997). My analysis of these studies raises the issue of whether some mixed methods research in management accounting is indeed more consistent with critical realism than hitherto recognized. This forms the stepping-stone for discussing the conclusions and implications for future research in Section 5.

This paper has explicated how the notion of triangulation may be refined in response ontological and epistemological criticisms of it in management accounting research, as well as in the wider social sciences. Particular attention has been paid to the alleged propensity of method triangulation to treat empirical observations as objectively verifiable rather than inherently theory-related and to suppress variations in situated meanings. By mobilizing critical realism, I illustrate how these criticisms may be addressed and re-conceptualize triangulation as a theory-related and context-bound validation technique grounded in abductive reasoning. The two empirical examples illuminate how this modified notion of triangulation may be applied in mixed methods research straddling between the functionalist and interpretive paradigms. The use of multiple theories in accordance with an abductive mode of inquiry played a vital role in researchers’ efforts to explain context-specific tendencies. Indeed, difficulties in explaining some empirical observations arose mainly from a lack of well-developed theories. Evidence of this can be found in Covaleski and Dirsmith's (1983) initial problems in interpreting survey findings guided by a functionalist perspective and in Hoque and Hopper, 1994 and Hoque and Hopper, 1997 inability to fully explain the role of aid agencies. This cautions us against the assumption of ‘theory-free’ observations directly corresponding to some empirical reality as a basis for explaining complex social phenomena (Ackroyd, 2004, Bhaskar, 1978, Sayer, 2000 and Tsoukas, 1989). At the same time, both studies entailed thorough attempts to situate theoretical insights through close examination of variations in context-specific meanings and an explicit awareness of the risks of pre-mature theoretical closure. While it would be presumptuous to argue that the two studies were based on ‘hidden’ critical realist assumptions, my analysis thus reveals many features that are consistent with the modified notion of triangulation advanced in this paper. This observation is consistent with the view that actual research practices do not always correspond strictly to the philosophical assumptions embedded in the functionalist and interpretive paradigms, but may in fact be located in the ‘transition zone’ between these two paradigms (Gioia and Pitre, 1990, Kakkuri-Knuuttila et al., 2008a and Kakkuri-Knuuttila et al., 2008b). This implies a need to modify the assumptions embedded in these paradigms as a basis for triangulation. As explicated below, critical realism provides a useful point of departure for doing so, as it enables researchers to address the ontological and epistemological criticisms of triangulation discussed in this paper. Following the assumption of reality as an at least partly mind-independent entity, which is made knowable through abductive reasoning, critical realism affirms key realist premises while relaxing the strict ontological commitments of theories and methods. It thus opens up the possibility of mobilizing a wide range of theories and methods in producing valid knowledge claims. The rationale for this can be traced to a profound scepticism of the possibilities of achieving a final state of theoretical closure, since theories are human constructions that can rarely, if ever, be conclusively verified. Hence, a critical realist mode of theorizing is far removed from the ‘grand’, totalizing claims about the prospects of some theories to exhaustively represent the world which sometimes emerge in functionalist (economics-based) accounting research (e.g., Zimmerman, 2001). Instead, it sees all theories as potentially fallible and in need of refinement based on complementary frameworks. It also recognizes that the theoretical frameworks available to researchers at any point in time are conditioned by the values and background assumptions of larger research communities (Niiniluoto, 1999). As illustrated by the empirical examples reviewed in this paper, the availability of adequate theories has an important influence on researchers’ ability to make sense of empirical observations, but it cannot be taken for granted even if theories have reached a seemingly ‘mature’ state. The discussion above counters the critique that triangulation ignores the theory-relatedness of empirical observations (Ahrens and Chapman, 2006 and Fielding and Fielding, 1986). As such, it is important to maintain researchers’ sensitivity to variations in situated meanings, so that these meanings are not unwittingly suppressed by the mobilization of particular theories and methods. The empirical examples reviewed in this paper illustrate how mixed methods studies may address this problem. A deeper understanding of context-specific meanings, reflecting a much richer understanding of budgeting than that embedded in functionalist theories, was achieved through the immersion of researchers in the lived experiences of the researched. Instead of trying to ‘reconcile’ context-specific variations in meanings diverging from narrow conceptualizations of budgeting, these variations were taken as the starting point for further theorizing. Hence, at least as far as these studies are concerned, the critique that triangulation invariably suppresses variations in meanings and distances researchers from the phenomena under examination (see e.g., Blaikie, 1991) seems exaggerated. More generally, critical realists would be highly sceptical of such distancing of researchers since it is likely to compromise their possibilities of analysing the contingent conditions under which particular mechanisms and causal powers give rise to certain empirical experiences (cf. Ackroyd, 2004, Sayer, 2000 and Tsoukas, 1989). Critical realism thus offers an alternative to the entrenched paradigm thinking informing criticisms of triangulation in mixed methods research (see e.g., Blaikie, 1991, Blaikie, 2000, Tomkins and Groves, 1983a and Tomkins and Groves, 1983b). As demonstrated in this paper, it is possible to accommodate much of this criticism within a critical realist framework which is largely consistent with the broader pragmatist tradition to which mixed methods research typically appeals. However, critical realism provides more clearly articulated ontological and epistemological premises than are found in many pragmatist approaches to mixed methods research (see e.g., Greene and Caracelli, 2003, Smaling, 1994, Tashakkori and Teddlie, 1998 and Teddlie and Tashakkori, 2003). In particular, it offers a sophisticated ontology which does not force researchers to slavishly subscribe to particular theories and methods. Similar to pragmatist approaches, such choices are seen as contingent on the values and background assumptions of wider research communities. While this epistemological position is no guarantee for theoretical and methodological pluralism, the emphasis on abduction as a key mode of inquiry can provide a strong impetus in this direction (Downward and Mearman, 2006). Nevertheless, we need to recognize the potential barriers to a critical realist approach to mixed methods inquiries in contemporary management accounting research. The institutional and political barriers to more pluralistic approaches, evident in research training, publication patterns and career dynamics, in some functionalist (especially North American) accounting research are well-documented (e.g., Panozzo, 1997, Reiter and Williams, 2002 and Tuttle and Dillard, 2007). Similarly, some interpretive researchers reject critical realism as an attempt to salvage some remnants of empirical realism, such as the belief in a mind-independent reality, which cannot be reconciled with a more ‘extreme’ subjectivist understanding of the world (e.g., Smith and Deemer, 2000 and Smith and Hodkinson, 2005). Even though recent debates suggest that few interpretive management accounting scholars actually subscribe to stronger forms of subjectivism, grounded in a belief that there will be little, if any, consensus about the nature of social phenomena (Ahrens, 2008, Kakkuri-Knuuttila et al., 2008a and Kakkuri-Knuuttila et al., 2008b), there may be resistance to research approaches based on a realist ontology (cf. Symon et al., 2008). Given such barriers, the approach to triangulation advanced in this paper should be viewed as a preliminary guide to what may constitute a philosophically informed foundation for mixed methods research in management accounting. I do not wish to supplant entrenched paradigm thinking with another dominant paradigm claiming to once and for all resolve philosophical tensions emerging in such research.