Print
Non-replicable?
expand article infoWillem Buijink
‡ Open Universiteit, Heerlen, Netherlands
Open Access

In my introductory remarks to the FAR Conference, I thought it would be useful to point out to the practitioners in the audience that there are different strands of auditing research. Practitioners at the conference were auditors, regulators and oversight body members. We are lucky to have so many of them attending the FAR conferences. Auditing academics in the audience know about these different strands. It is the existence of strands of research in a research area that generates part of the discussions within that area. But we do not often reflect on the impression this must make on practitioners. Practitioners will easily hear a cacophony of opinions. So as academics we should explain that this cacophony is productive. I think it would be helpful to tell an audience at each conference where auditing research is presented, ‘where a researcher carrying out a given auditing research project is coming from’; as Wim Gijselaers put it later during the FAR conference. Regrettably, that does not often happen.

What these strand in auditing research are, is what I attempted to explain in my introductory remarks: very briefly and in a neutral fashion. In this column, I take this opportunity to explain the issue, the importance of seeing the existence of different strands of auditing research, in a bit more detail. Also, it occurred to me afterwards that I had been too neutral. That needs a correction. Not all strands are praiseworthy.

For auditing students the existence of different strands of auditing research must often also be confusing. This column, many students have to read this journal, may be helpful for them as well.

So, how can research strands in auditing research be distinguished?

Auditing is human behaviour. Here are a few examples. Humans create corporations and nonprofit entities; they create corporate law. Stakeholders in corporations create demand for auditing, to control the behaviour of people managing the corporations and nonprofit entities, and for regulation of auditing. The auditing process uses a collection of techniques that is designed and used by humans. The entities created trade with each other and with consumers over markets in an economy. Within these entities management devise sets of management controls. All of this is humans behaving.

So, the first way in which auditing researchers can tell us where they ‘come from’, is by saying which behavioural science, anthropology, economics, psychology or sociology, they use to understand auditing behaviour. Philosophers also reflect on human behaviour, and indeed in auditing research different kinds of philosophy are also used. I mentioned a number of philosophers ‘used’ in auditing research in my opening remarks: Marx, Foucault and Habermas. Marx is also an economist of course. Using one of the behavioural sciences, or philosophical insights, auditing researchers are attempting to understand auditing behaviour: how does the auditing-related world work? How can it be made to work better? Doing this, they in effect formulate causal claims to explain and to improve the auditing-related world.

In practice, I often observe auditing researchers mixing behavioural sciences. Behavioural economics inspired auditing researchers mix psychology and economics. I actually prefer competing tests. Say, sociology based explanations against economics based ones. Why do corporations have auditors? Microeconomics explains this; sociological institutional theory does as well. Let us test which is the better explanation. I prefer this approach, because sociology is really different from microeconomics. I am on the losing side of the argument I fear. Mixing of behavioural science based explanation these days also occurs because data driven analysis is gaining ground.

The second way to characterize auditing research is to determine which ‘testing’ methodology, research design, an auditing researcher uses. Empirical behavioural science and philosophy based auditing research uses various research designs to test its causal claims: case studies, field studies, lab experiments, field experiments, natural experiments, archival data based observational studies. The outcome of these research designs, speaking to the claims made, is then analyzed using statistical methods and other data analytical techniques. Note that theoretical auditing research, mostly economics based, uses logic and mathematics to test its claims. But it does test.

Therefore, combining the above, auditing researchers can say: this is how carefully I develop my claims about the auditing world and this is how I carefully test them; can see them being replicated. Auditing researchers could say for instance: “I am a psychology based auditing researcher and I use mostly lab experiments to test my claims”. It is the careful development and testing of these claims, and the replicability of the tests, that makes auditing research valuable in public debates about auditing and also auditing policy making.

However, and this is the concern that I did not voice during my introductory remarks, there is now a group of auditing researchers that uses behavioural science or philosophy, but argues that testability of their claims is unimportant. This sub-community has been receiving more attention recently. An example of this is a recent overview paper in AJPT (Auditing; a Journal of Theory and Practice) that sympathetically explains the ‘principles’ of this sub-community (Power and Gendron, 2015; hereafter (PG). The paper is easy to find using Google Scholar. AJPT is the leading auditing research field journal. Table 1 in PG lists the assumptions of this sub-community in accountancy (auditing) research. I highlight three of these. (1) reality is socially constructed; (2) subjectivity in research is inevitable; (3) research into cause and effect relations is not necessary: human behaviour is too complex. PG use the label ‘constructivist’ for this strand of auditing research.

These assumptions though, make claims of the constructivist sub-community about how the auditing-world work non-replicable. Subjectivity will reign. Critiquing constructivist claims about auditing is made impossible. This poses a threat to evidence-based, evidence-informed, policy making related to the auditing-world. That is not good. PG offer some critical reflection on the, of necessity, subjective nature of constructivist auditing research outcomes, but they are sympathetic to constructivist auditing research. The important issue of the lack of testability, of replicability, of constructivist research is not addressed by them.

The fact that PG appeared in AJPT, not a constructivist auditing research journal is telling. It highlights the growing attractiveness of constructivist auditing research. Constructivist auditing research is often published under the more familiar label of ‘critical (perspectives) (auditing) research’. The constructivist/critical sub-community is also gaining influence among auditing researchers in the Netherlands. That can be easily observed on the Accountant.nl website. It appears to have an increasing impact on auditing policy makers, regulators, and oversight bodies.

Note the following. Critical perspectives auditing research, this is where the name ‘critical’ comes from, also aims to contribute to emancipation in a society, by studying how auditing helps or limits marginalization of members of a society. This aim is also voiced by the Dutch critical community of auditing researchers. It can be a laudable aim. My criticism does not target their emancipation aim. It targets their methodological approach.

Not mentioning this sub-community of auditing researcher in my FAR Conference was unwise. It prevented me from alerting the practitioners and students, and perhaps researchers as well, to its existence, and from criticizing the constructivist research strand in auditing research. This column is a correction.

Professor dr. W.F.J. (Willem) Buijink is Hoogleraar Financial Accounting aan de Open Universiteit Heerlen. Hij is ook emeritus Hoogleraar Bedrijfseconomie aan de Universiteit Tilburg. Daarnaast is hij academisch bestuurslid van de FAR.

References

  • Power MK, Gendron Y (2015) Qualitative research in auditing: a methodological roadmap. Auditing. A Journal of Practice & Theory 34(2): 147–165. https://doi.org/10.2308/ajpt-10423
login to comment