3089-prelims.qxd 4/2/03 4:42 PM Page i RESEARCH METHODS IN ACCOUNTING MALCOLM SMITH SAGE Publications London • Thousand Oaks • New Delhi 3089-prelims.qxd 4/2/03 4:42 PM Page ii © Malcolm Smith 2003 First published 2003 Apart from any fair dealing for the purposes of research or private study, or criticism or review, as permitted under the Copyright, Designs and Patents Act, 1988, this publication may be reproduced, stored or transmitted in any form, or by any means, only with the prior permission in writing of the publishers, or in the case of reprographic reproduction, in accordance with the terms of licences issued by the Copyright Licensing Agency. Enquiries concerning reproduction outside those terms should be sent to the publishers. SAGE Publications Ltd 6 Bonhill Street London EC2A 4PU SAGE Publications Inc 2455 Teller Road Thousand Oaks, California 91320 SAGE Publications India Pvt Ltd B-42, Panchsheel Enclave Post Box 4109 New Delhi 110 017 British Library Cataloguing in Publication data A catalogue record for this book is available from the British Library ISBN 0 7619 7146 7 ISBN 0 7619 7147 5 (pbk) Library of Congress Control Number available Typeset by C&M Digitals (P) Ltd., Chennai, India Printed in Great Britain by The Cromwell Press Ltd, Trowbridge, Wiltshire 3089-prelims.qxd 4/2/03 4:42 PM Page iii This book is dedicated to Beth, Cedric and Alice 3089-prelims.qxd 4/2/03 4:42 PM Page iv 3089-prelims.qxd 4/2/03 4:42 PM Page v Contents List of Figures xi List of Tables xii Acknowledgements xiii Preface xiv 1 Introduction and Overview Theory as testable explanation A critical approach to accounting research 1 6 8 2 Developing the Research Idea The research sequence Emergence of the research topic Conceptual frameworks The structure of DNA: the development of new theory The Bradman problem: the development of new strategies The longitude problem: implementing solutions Strategic management accounting 16 16 20 24 30 31 34 36 3 Theory, Literature and Hypotheses Sources of theory Searching the literature Modelling the relationship Developing the hypotheses Validity concerns 39 39 47 47 52 53 4 Data Collection and Analysis Sample selection Measurement issues Data management Descriptive statistics Differences in sample means Measures of association Analysis of variance Multivariate model building 55 56 56 59 60 64 69 73 75 3089-prelims.qxd 4/2/03 4:42 PM vi Page vi Contents 5 Research Ethics in Accounting The ethics quiz Informed consent Ethical guidelines 91 92 95 97 6 Experimental Research The problem statement Theory and context Experimental design The validity trade-off Quasi-experimental research 100 101 101 104 108 113 7 Survey Research Mail surveys Design and planning issues Pilot testing Data collection Measurement error Interview methods 117 117 120 122 123 126 128 8 Fieldwork Case study methods The qualitative analysis protocol Grounded theory Verbal protocol analysis 131 134 136 139 140 9 Archival Research Cross-section data Time-series data The validity trade-off in archival research Content analysis Critical analysis 142 143 144 146 147 150 10 Supervision and Examination Processes The role of the supervisor Examiner profiles The examination process 152 153 159 160 11 Turning Research into Publications Why publish? Where to publish? What to publish? How to publish? Concluding remarks 165 165 166 172 176 179 3089-prelims.qxd 4/2/03 4:42 PM Page vii Contents vii Appendix 1: Ranking of Accounting Journals 180 Appendix 2: Sample Paper (1) 182 Appendix 3: Sample Paper (2) 202 References 209 Index 230 3089-prelims.qxd 4/2/03 4:42 PM Page viii 3089-prelims.qxd 4/2/03 4:42 PM Page ix List of Figures Figure Figure Figure Figure Figure Figure Figure Figure Figure 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 The research sequence The positivist approach Alternative research methods Kolb’s Learning Cycle The deductive process The conceptual schema Measurement issues The Harvey-Jones approach to problem-solving Generalised process–improvement sequence 17 19 20 25 26 26 28 37 37 Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 3.10 Searching for construct validity Voluntary disclosure and reciprocal causality Intervening variables and causality Non-financial focus as an intervening variable Moderated causal relationship Size, industry and culture moderating performance Influence of extraneous variables Economic conditions as extraneous variables Multiple independent variables Modelling the recruitment process 41 48 49 49 49 50 50 51 51 52 Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure 4.1 4.2 4.3 4.4 4.5 4.6 4.7 4.8 4.9 4.10 4.11 4.12 Significance of test of proportion χ2-test for difference in frequencies Test of significance of correlation coefficient t-test for difference in means t-test for paired-case difference in means Mann–Whitney U-test for difference in means Product–moment correlation coefficient (Pearson’s r) Coefficient of rank correlation (Spearman’s rho) Measure of association within contingency tables One-way analysis of variance (ANOVA) The Kruskal–Wallis multiple sample test Summary of regression results for charity shops case study 62 63 65 66 67 68 70 71 72 74 76 82 3089-prelims.qxd 4/2/03 4:42 PM Page x List of Tables Table 1.1 Three alternative approaches Table Table Table Table Summary of significance tests by measurement level Measures of association by measurement level ANOVA summary table Summary of multivariate model-building methods by measurement level 4.1 4.2 4.3 4.4 Table 10.1 Expectations of supervisor and candidate 5 62 70 75 77 155 3089-prelims.qxd 4/2/03 4:42 PM Page xi Acknowledgements This book could not have been written without the help of numerous colleagues, most notably Bev Schutt and Paul Martin from the University of South Australia. Thanks are also owed to Glen Lehman, Linley Hartmann, Bruce Brown and Bruce Gurd at UniSA, and to David Russell, Ashok Patel and Elaine Harris at the Leicester Business School. However, as is normal, all errors and omissions remain the responsibility of the author. 3089-prelims.qxd 4/2/03 4:42 PM Page xii Preface This book aims to provide an insider’s view of the research process, by focusing on actual choices made in the conduct of accounting research projects, together with a realistic perception of what might go wrong, even with careful planning. We must, however, acknowledge that no single author can be an expert in all research methods; this author is no exception. My own publications will readily reveal a preponderance of studies concerning experimental methods and the use of archival data; there are fewer instances of studies using survey and field study methods. It would be unwise of me to claim expertise in the implementation of all such methods, so this book must necessarily lean heavily on the work of others. For the same reasons, and because of pressure of space, this book does not address issues of finance, capital markets research, or stockprice-related accounting research on the fringes of finance. Most other texts in this area are long, over-theoretical and not particularly user-friendly. This book aims to address these issues by adopting a practical approach which takes the reader from the initiation of the research idea right through to the publication of the research findings. The intended readership is wide, embracing instructors, doctoral candidates, and academics starting, or re-starting, their research careers. Although the focus, and examples, are mainly accounting based, much of the material will also be relevant to more general business applications, of particular interest to those pursuing a Doctorate of Business Administration (DBA) qualification. The practical examples employed are usually UK or Australia-based, these being the two countries in which I have extensive, current experience of teaching, supervision and examining, but the principles should normally adapt easily to alternative environments. An early distinction between ‘methods’ and ‘methodologies’ in research is essential because the two are so often confused, or else used interchangeably. Research methods are concerned with the technical issues associated with the conduct of research; research methodology is concerned with the philosophies associated with the choice of research method. This book is almost exclusively concerned with the former and, after Chapter 1, deliberately neglects the philosophical foundations of research except where reference thereto is unavoidable. Chapter 2 examines the research idea, and the documentary sources which might aid their development. A number of examples, many from nonaccounting environments, are used to illustrate the research sequence, and to examine research that is seeking either to improve outcomes, to explain improved outcomes through new theory, or to examine the improvement 3089-prelims.qxd 4/2/03 4:42 PM Page xiii Preface xiii process itself. Theoretical frameworks and research models are used extensively here to help the reader to picture the key variables and relationships underlying their research. Theory is the focus of Chapter 3, on the basis that ‘good research is founded on good theory’. The chapter addresses the sources of the theory widely applied in accounting research, but drawn from other disciplines. In the space available it can only hope to give a flavour of the diversity which is available; indeed, it prompts us to suggest that ‘theory in accounting research’ might provide a suitable follow-up text in its own right! Recognition of the importance of theory, reliability and validity as desirable characteristics of accounting research lead, in Chapter 4, to the issues of data collection, management and analysis necessary to conduct hypothesis testing. This chapter is unashamedly quantitative in nature, but the relative strengths of qualitative analysis are addressed in subsequent chapters. Chapter 5 addresses the increasingly important ethical considerations which underpin the conduct of accounting research, and the subsequent publication of research findings. It highlights the confusion which is still apparent among many academics as to what constitutes unethical conduct, and specifies the necessary guidelines for good practice. Chapters 6 to 9 are devoted, respectively, to the core forms of accounting research: experimental, survey-based, fieldwork and archival. Numerous examples are used to demonstrate the relative advantages of alternative methods so that researchers can both make an informed choice and justify their preferred approach. Research can be based on quantitative or qualitative methods, and both should be equally acceptable as long as the most appropriate method has been chosen. Richardson (1996, p. 4) notes that ‘work on how science really gets done’ (such as that described in Chapter 2 based on Watson, 1968) shows that even though we are in an extreme ‘positivist’ domain, interpretive knowledge is still important in the development of new theory. The majority of the readers of this text will likely be doctoral candidates so Chapter 10 is devoted to supervisor–candidate relationships, highlighting the mutual responsibilities of both parties to the supervision process, from the outset right up to the examination process. Publication is the natural target output of the research process, and Chapter 11 addresses the complexity of the publication process. In doing so it recognises that we are working in a dynamic process; what was once acceptable in accounting research is no longer so because of a more appropriate emphasis on research ethics; what is publishable, at all or in specific journals, changes too, both with the passage of time and the passing of particular journal editors. Many journals remain very conservative in the type of research they will publish, often on the grounds that it is difficult to demonstrate that ‘new’ methods constitute ‘good’ research in the same way as the traditional methods. But this situation is changing gradually – the wider opportunities for publishing case-based research in recent years provides evidence of this. However, the renewed emphasis on journal and university rankings, and associated funding systems based on the quality of publications, provides fresh difficulties. 3089-prelims.qxd xiv 4/2/03 4:42 PM Page xiv Preface The provision of ‘acceptable journals’ listings by many universities, and the prohibition of publication elsewhere, perpetuates the position of the wellestablished journals, while making it extremely difficult for the editors of other journals to attract quality submissions. The opportunities for innovative new journals are also severely diminished in such circumstances. Contributions to the profession by academic accountants are generally not well regarded, either by one’s colleagues or by government bodies providing funding based on publications performance, even though, arguably, the education of the potential employers of our students might be seen as an important part of our jobs. So journalistic pieces in practitioner magazines and workshops to professional audiences count for close to nought – even though the individuals concerned would never read a refereed journal or attend an academic conference. We need to exploit the available media to get our message, and the power of research findings, over to those implementing change in an unbiased way, before the consultants get in on the act! This process must be of mutual benefit to all parties, but if the practitioners feel they are being short-changed, or even used, then future collaborative efforts will be threatened. It is just such attitudes which generate the ‘them and us’ cultures leading to accusations of academics being out of touch with reality. In this context it is interesting to note the changes taking place within professional journals: there were once two such journals called Management Accounting, but now there are none, the US version becoming Strategic Finance and the UK one Financial Management. With moves to term ‘accounting’ as ‘assurance services’ it will surely not be long before some of the professional bodies themselves follow their journals with the removal of the word ‘accounting’ from their titles. Communication problems also remain. The timeliness and relevance of much of the content of the refereed literature does little more than suggest that it is written by academics, only for the consumption of other academics! Most practitioners do not have an appreciation of research methods, nor do they read the refereed literature, so important findings and recommendations often do not reach the individuals who can make sure it has the greatest impact. A number of journals have emerged with the express intention of providing readable research for practitioner audiences, but even these have tended to become more academic and less readable over relatively short time periods. This book aims to provide a treatment of research methods that will be of use to both accounting practitioners and those contemplating the conduct of research projects. Space restrictions mean that this slim volume cannot hope to tackle all of the detail of the application of different research methods, or the associated intricacies of complex quantitative methods. But if its use causes one paper to be published that would otherwise have gone unpublished, then all will have been worthwhile. Malcolm Smith 1 Introduction and Overview CHAPTER CONTENTS • Theory as testable explanation • A critical approach to accounting research 6 8 A number of authors (e.g., Brownell, 1995, p. 2) describe accounting researchers as ‘parasites’ who prey on the work of others to generate their findings. The term may be an overstatement, but as with most rash generalisations it contains more than a germ of truth: accounting researchers have little theory of their own (they rely on economics, finance, psychology, sociology and organisational behaviour as their major sources); they have no methods of their own (they are all adapted from the natural and social sciences); and they have few instruments of their own (with many of these originating in or adapted from the organisational behaviour literature). Merchant (quoted in Brownell, 1995, p. 140) even suggests that organisational behaviourists are much better at developing survey instruments than their accounting counterparts. The overall aim of this book is to facilitate the conduct of applied research studies in accounting, and to do this we must recognise our reliance on work in other disciplines. To accomplish this aim, a number of subordinate objectives may be identified, all of which will contribute to the overall goal: • an understanding of contemporary research ideas in accounting, so that readers can identify and define research problems and prepare strategies for their solution; • an awareness of alternative research methods, to facilitate the selection of the most appropriate method for addressing particular research questions; • an ability to review existing research and to offer critiques of articles published in refereed journals; and • an appreciation of the ethical constraints on the conduct of accounting research. 2 Research Methods in Accounting Research in accounting is concerned with solving problems, investigating relationships and building a body of knowledge. Because we rely to such a great extent on prior research in the natural and social sciences to do so, this volume will take a similar approach in leaning on work in other disciplines where it helps to inform accounting research. Bennett (1991) identifies four basic levels of research: • Description – concerned with the collection and reporting of data related to what is, or was, the case. This would include means and standard deviations of individual variables, and correlations between pairs of variables. • Classification – still descriptive, but easing the reporting process, and highlighting similarities and clustering through grouping and classifying (e.g., through the familiar cross-tabulation facility in most basic statistical packages). • Explanation – an attempt to make sense of observations by explaining the relationships observed and attributing causality based on some appropriate theory. • Prediction – going beyond the understanding and explaining of the prior stage, to model observations in a way that allows testable predictions to be made of unknown events. We return to this structure in Chapter 4 when discussing alternative quantitative methods, but an early distinction between ‘explanation’ and ‘prediction’ is appropriate here, because, as in the natural sciences, we are able to make excellent predictions of accounting behaviour without the backing of a sound underpinning theory. Bankruptcy prediction modelling provides an excellent example. A number of researchers (e.g., Altman, 1968; Taffler, 1983) have developed models that have proved very successful in identifying ‘distressed’ companies – those companies that will fail in the short term. These models are statistically excellent but the theory underpinning their content, in terms of the ratios to be used and the variables they represent, is extremely weak; the essential problem is that such theories as we have (e.g., Blum, 1974; Myers, 1977; Scott, 1981) do not generate very good predictive models! Good research generates the sound evidence needed to overturn or revise existing theories. These assertions will, in turn, yield to revised theories based on better evidence, so that healthy competition between rival ideas will lead to better explanations and more reliable predictions. Two major processes of reasoning, ‘deductive’ (theory to observation) and ‘inductive’ (observation to theory), are important for theory construction and observation testing. Inductive reasoning starts with specific observations (data) from which theories can be generated; a generalisable pattern may emerge from further observations and repeated testing for compliance. The natural sciences, for example astronomy, provide numerous examples of inductive reasoning, thus Hawking (1998) provides a number of fascinating examples of theories revised, or still in question, with implications for the progress of accounting research. However, he notes that generalisations made on the basis of induction Introduction and Overview 3 can never be regarded as ‘certain’, since just one contrary instance can cause them to be overturned: • Big Bang versus Steady State. From the late 1940s to the mid-1960s two competing theories were prominent in offering alternative explanations of the origins of the universe. The ‘Big Bang’ theory recognised a singular event as causing an ever-expanding universe in which matter (notably galaxies) becomes continuously more widely dispersed. The ‘Steady State’ theory, attributed to Bondi, Gold and Hoyle, on the other hand, suggested that matter was continuously being created to fill the gaps between existing galaxies. They argued that the universe had no beginning, and had been forever expanding, with new matter being created out of apparently empty space. The Steady State theory importantly provided testable hypotheses in suggesting that the universe should look the same at all times and from wherever it was viewed. But surveys of radio waves in the early 1960s showed that sources were more numerous in the past, and that there were many more weak (distant) sources than strong (close) ones. Further, microwave radiation studies in 1965 demonstrated that the universe did not have a common density – it had been much denser in the past. These observations provided disconfirmations of the Steady State theory, causing its abandonment. • Newton’s Laws of Physics. New theory emerges when a new observation arises which does not correspond with existing theory. Once the technology permitted accurate observations of the planet Mercury to be made, it was clear that there were small differences between its observed motion and that expected under Newton’s Theory of Gravity. Einstein’s general theory of relativity matched the observed motions of the planet in a manner that Newton’s theory did not, providing confirmation for the new theory. • The Wave Theory of Light. We can attempt to explain the behaviour of light in terms of its being composed of either ‘waves’ or ‘particles’. Each view produces a plausible explanation of behaviour – both of which are needed to affirm existing properties – but they are incompatible explanations which cannot exist simultaneously. New theories are required (possibly those associated with parallel universes) for a complete understanding of the incompatibility. Deductive reasoning, on the other hand, starts with the theory and proceeds to generate specific predictions which follow from its application. The predictions can be verified, or otherwise, from subsequent observation. For example, in his seminal paper, Healy (1985) used agency theory to develop a bonus hypothesis which could be substantially verified through observations of how managers manipulated their accounting earnings to optimise their short-term bonus performance. 4 Research Methods in Accounting However, such a strict division of reasoning processes is not always helpful because interdependencies almost always exist: induction will usually imply some knowledge of theory in order to select the data to be observed (a common criticism of grounded theory advanced in Chapter 8); deduction will be dependent on the selection of the initial hypotheses for testing. Even without such problems, the scientific position of ‘objective measurement’ has come under repeated attack, in both natural and social sciences, because the act of observation is itself ‘theory-laden’ and influenced by the motives and preferences of the observer. For example, Hopwood (1987), in management accounting, and Hines (1988), in financial accounting, argue that accounting helps to create the ‘facts’ that it is supposedly reporting. More radical approaches (e.g., Tinker and Niemark, 1987) suggest that accounting distorts practice in a systematic manner. Such concerns have aided the development of new approaches: an interpretive perspective and a critical perspective. • An interpretive perspective – From an interpretive perspective, human actions are the result of external influences. These actions have both intentions and reflections, and take place within a structure of rules which binds the participants. The task of the researcher goes beyond measurement to developing an understanding of the situation. To do this effectively, active participation, rather than detached observation, may be required. Since the ‘action’ may be interpreted ambiguously when taken out of context, this perspective places the fundamental emphasis on the understanding of the process. In an accounting context, Arrington and Francis (1989) provide an example of this approach. • A critical perspective – The critical approach expands on the scope of the interpretive approach by focusing on the ownership of knowledge and the associated social, economic and political implications. An empirical approach is criticised on the grounds that the research process is value-laden, and that the acquisition of knowledge provides the opportunity to oppress those being researched. In an accounting context, Tinker (1980) provides an example of this approach. Table 1.1 summarises the differences in research assumptions, process and outcomes associated with each of these three major approaches. Kuhn (1970) suggests that researchers are concerned with problem-solving within a single framework of widely accepted beliefs, values, assumptions and techniques. This shared framework, or view of the world, he termed a paradigm, so that a ‘paradigm shift’ corresponds with some revolution where the existing framework and theories can no longer cope with the volume of disconfirming evidence. Kuhn neatly illustrates such a shift by reference to a simple psychology experiment: Subjects viewed cards from a deck. The deck included some unusual cards, including black hearts and red spades, but the subjects were not informed in Introduction and Overview 5 TA B L E 1 . 1 Three alternative approaches (adapted from Connole, 1993, p. 37) Positivist 1. 2. 3. 4. 5. 6. 7. What is the approach modelled on? Classical investigation founded in the physical sciences. What does it assume about reality? Reality is unitary and it can only be understood by empirical and analytic methods, i.e., the scientific approach. What is the foundation of data? Disciplined rules for observation. How is observation done? Through clear and unambiguous rules which are not modified by the setting and are totally independent of it. What is generated? Evidence and generalisable laws which are not affected by contexts and have nothing to do with the way in which they were discovered in the first place. Objectivity depends upon the removal of error and bias which is related specifically to the logic of observation and measurement. What interests are inherent? Prediction and control, technically exploitable knowledge, and explanation. What values are inherent? Science and scientific knowledge are inherently value-neutral. Interpretive Critical Historical, literary and existential studies in which the subjective understandings of subjects are significant. Marxist and interpretive studies which focus on the insights and judgements of the subjects. There are multiple realities which require multiple methods for understanding them. There are multiple realities which are made problematic through distorted communication. Meanings are the basis of data: meaning precedes logic and fact. Meanings are found in language and social behaviour and they precede logic and fact. Through the social, linguistic and cognitive skills of the researcher. Interpretive methods, plus critical self-reflection concerning the grounds of observation. Knowledge which is dependent on the process of discovery. The integrity of the findings depends upon the quality of the social, linguistic and cognitive skills of the researcher in the production of data analyses and conclusions. Knowledge which falls within the interpretive framework, but which also serves the purposes of assisting personal liberation and understanding, and emancipation from forces constraining the rational independence of individuals. Understanding at the level of ordinary language and action. Discovering the meanings and beliefs underlying the actions of others. Interpretive interests and those which underliey other forms of inquiry. Radically improving human existence. Practical and public involvement in knowledge formation and use. Science and scientific knowledge have both to be interpreted in terms of values they represent. Science and knowledge are never value-neutral: they always represent certain interests. 6 Research Methods in Accounting advance about their presence. Initially the subjects saw only ‘hearts’ and ‘spades’, because they believed that only ‘red hearts’ and ‘black spades’ existed; only with repeated viewing did they grasp that these cards were not typical of a normal deck. Then they could recognise the cards that existed rather than the ones they were expecting. In accounting research the parallels might be the paradigm shifts associated with the ideas introduced by Ball and Brown (1968) and the difficulty they had in getting a paper published which questioned the existing paradigm by showing a link between stock prices and accounting earnings, through the abnormal performance index. A similar, though perhaps less radical, movement is associated with Watts and Zimmerman (1978) and their popularisation of agency theory in an accounting environment. What is inescapable is that we are dealing with people, and in the research community that means individuals with their own agenda and with reputations to build and protect. The natural sciences are littered with character assassinations of individuals and their work, by others who have been less than willing to accept the impact of new findings on their own fiefdoms. Sir Humphrey Appleby, in Lynn and Jay (1987), outlines the four stages of the process necessary to discredit an unwelcome report. The parallels between the fictitious Department of Public Administration and academia are uncomfortable, where unwelcome findings might arise from academic competitors: 1 2 3 4 Refuse to accept the findings on the basis that they could be misinterpreted, and that a wider and more detailed study is required. Discredit the evidence on the basis that it is inconclusive and the figures are open to other interpretations, or that the findings are contradictory and leave important questions unanswered. Undermine the recommendations because they say nothing new, and provide insufficient information on which to draw valid conclusions. Discredit the researcher by questioning his or her integrity, competence and methods employed. We thus have doubts about the researchers, their research questions, their research methods, the means of data collection and analysis, and the validity of the interpretation and recommendations – all issues to which we will return. Theory as testable explanation Faced with a set of diverse observations, we can establish a set of tentative explanations which help to make sense of the diversity. Such explanations Introduction and Overview 7 constitute theory. In any set of circumstances there will usually be multiple theories available to explain the observations. The systematic collection of further data allows for the testing of the alternative theories so that we can establish which of the exiting theories best explains the facts. A layman’s perspective of ‘theory’ is cynically expressed in Michael Crichton’s The Lost World as: ‘A theory is nothing more than a substitute for experience put forward by someone who does not know what they are talking about’ (1995, p. 67). The data collection itself allows only a descriptive approach (e.g., means, standard deviations, ranges, correlations); we cannot attempt to attribute causation in any meaningful way without recourse to an explanatory theory. We are always looking for another theory which may fit better, so that, as Popper (1959, p. 104) suggests, a ‘genuine test of a theory is an attempt to falsify it or refute it’. We look for disconfirmations rather than confirmations. In the short term this may not be successful. In accounting, we witness the frequent and numerous ‘anomalies’ to which the Efficient Markets Hypothesis (EMH) is subject, but we have no other widely accepted theory of the manner in which stock prices react to the availability of relevant information. Popper’s suggestions are very attractive in providing a powerful empirical methodology for subjecting theories to attempts to refute them. However, this position is not always ideal because the process of ‘observation’ in itself may be fallible. Thus Hawking (1998) reports Heisenberg’s Uncertainty Principle: If we are to predict the future position and speed of any particle, then we require accurate measurement of both its present position and current speed. Heisenberg did this in 1926 by shining light on a particle, and observing the resultant scattering of light in order to reveal its position. However, to determine the position of the particle accurately an amount of light needed to be used which changed the speed of the particle in an unpredictable way: the more accurately he tried to measure position, the less accurate was the measurement of speed! The Uncertainty Principle has wide implications for research conducted in any environment, where it is impossible to measure the size and speed of a particle without altering all other characteristics in the process of measurement. We have a parallel situation in accounting research where the actions of the participants in ethnographic, experimental, survey or fieldwork impacts on the outcomes of the measurement process. Three fundamental criteria exist to judge whether theory fits observation: 1 Co-variation – even where no causality exists we would expect the two variables to move together so that a high degree of correlation exists between the two variables. Where there is no co-variation it will be difficult to establish a causal link. 8 2 3 Research Methods in Accounting Cause prior to effect – if a causal link is to be established then the ‘causal event’ should occur before the ‘effect event’. The sequence of events can therefore help to establish an explanatory direction. Absence of plausible rival hypotheses – the third rule seeks to eliminate alternative explanations of the events as being implausible. This may only be possible in the present, because future researchers may develop competing explanations of the events from a re-analysis of the data. Consider, for example, the voluntary disclosure of information in corporate reports and analyst following (i.e., the number of analysts examining the performance and reporting on the disclosures of large companies). There is a relationship between these two variables – they co-vary: the volume of voluntary disclosures and the number of analysts reporting move together. But which is causing which? Rival hypotheses suggest: a b companies are supplying more information voluntarily to the market to signal their intentions and reputation, attracting the attention of more investment analysis; investment analysts are focusing their attention on particular companies and demanding more information and more detailed disclosures. The existing empirical evidence is less than convincing: Lang and Lundholm (1996) find (a); but Walker and Tsalta (2001) provide only weak evidence for (a) but stronger evidence to support (b). Clearly more empirical work is required to clarify the direction of causation. A critical approach to accounting research Researchers must demonstrate a healthy scepticism towards both their own findings and those of other researchers. They must adopt a critical posture, questioning everything that they read until sufficient evidence has been provided for them to be satisfied with the quality of the outcomes. The development of critical appraisal skills is a fundamental requirement in researchers, so that they can distinguish between good and bad research, and clearly identify flaws of argument, methodology and analysis. Honest and transparent reporting of research practice is an ethical duty of those participating. Researchers should report everything that they did, why they did it, and how they did it. If they have doubts about any stage of the procedure, then these should be stated, along with their likely implications and what, if anything, has been done to overcome these doubts. Where researchers have been ‘economical with the truth’, this is usually apparent in their papers and is often an indicator of bad research. Introduction and Overview 9 Students frequently struggle initially when they are asked to critique published articles. They are often in awe of the reputation of the authors, or doubt whether they are able to offer sensible criticism of papers which, after all, have already undergone editorial scrutiny and double-blind review. Despite the above, some flawed papers do get published, and these are not always in lower-tier journals (see Hartmann and Moers, 1999, for their critique of 28 papers on contingency analysis in three top accounting journals – Accounting Organizations and Society (AOS), The Accounting Review (AR) and Journal of Accounting Research (JAR) – in which they identify problems in the design and analysis of 27 of the studies!). With appropriate guidelines as to the appropriate questions to ask, students can quickly develop some confidence in their ability to spot flaws and omissions. For example, Abernethy, et al. (1999) provide a stimulating critique of the three subsequent papers in the same, outstanding, edition of the journal Accounting & Finance. We would usually want to address the following: 1 Why is this article interesting/important? The paper must offer some new insights which constitute a contribution to knowledge. These insights should be non-trivial, so that they can be embraced in either further theory development or recommendations for improvement. 2 Are the outcomes important? Effectively, does the paper pass the ‘so what’ test? Will anyone be interested in the outcomes of this research, or will it have any implications for future practice? Would the scope of the research be well regarded by competitive grant authorities? This has important implications for those papers which produce ‘negative’ findings, that is, they test reasonable hypotheses based on the research literature, but their datasets fail to support any of the expectations. These findings still make a contribution in that they demonstrate that findings from elsewhere (often other disciplines) do not hold in accounting, but their negativity may restrict their publication opportunities. 3 What motivates the authors to write this article now? The paper may be clearly addressing issues of contemporary concern; on the other hand, it may be addressing more historical issues and/or be using ‘old’ data. If we have the latter, we may be dealing with an old paper recently recycled, or a paper which has been through many iterations at several different journals before being deemed ‘publishable’. 4 What is the research problem/question? We are looking for a clear statement of the problem very early on in the paper, so that its objectives are readily apparent. If we reach page 11, say, of the paper without a clear idea of its direction, or any sort of research model, then perhaps the authors need to readdress the fundamental purpose of the research. 10 5 Research Methods in Accounting What theory or theoretical framework underpins the research? Without some theoretical foundations we have a problem-solving exercise or a consultancy project, neither of which should be gracing the pages of a refereed journal. There must be some theoretical justification for the question being addressed and the research approach adopted. Theory will often not come first in the research process – it will frequently be preceded by an interesting idea or a perplexing observation. But we require some theoretical explanation for the relationships under investigation before we have the basics of a refereed journal article. Observed deficiencies in this area usually fall into one of four categories: • • • • the underlying theory is either non-existent or extremely thin; the theoretical context is there but appears to have been tacked on as an afterthought – usually at the beginning of the paper and often written by a co-author. Examination of writing styles suggests that we frequently do not have a seamless divide between ‘theory’ and ‘conduct of research’; the theoretical arguments are unconvincing, so that there are competing theories that may reasonably have been adopted in the paper but have been overlooked; a sound theoretical framework but findings which are totally at odds with theory. Apparently, a competing theory may be more appropriate, although this is unknown to the authors at the time. 6 What are the key motivating literatures on which the study depends? There will normally be a small number of seminal pieces of literature which are driving the research. If any of these are themselves unreliable, it may cast doubt on the state of the foundations on which the paper is based. If one of the papers is an unpublished conference or working paper from several years before, then alarm bells ring to question why that piece has not itself been published in the refereed literature. If key seminal pieces of literature have been overlooked, then again the integrity of the findings is reduced. 7 Which research method has been chosen? There should be a justification for the chosen method, and a clear preference over alternatives. The method should be consistent with both theory and literature and, ideally, prior empirical studies in the field will have adopted similar methods. Most importantly, we want to see a research method that has evolved rather than one that has been selected first, even before the research question has been fully developed. The use of survey methods should always be questioned in this way since frequently they seem to have been selected without explanation of the elimination of alternatives. Ideally, we should be able to trace through the emergence of abstract concepts, from theory, through their operationalisation and measurement, so that any hypotheses are entirely consistent with both theory and literature. Introduction and Overview 11 8 How has the sample been selected? Details on sample selection are often sketchy in many articles, perhaps because the authors feel vulnerable about the procedures adopted. Sometimes (see, for example, Young, 1996) the actual sample size employed is omitted, as is the response rate. Both omissions should be regarded as bad news. It is usually clear that scientific methods have not been adopted (unfortunately far too commonly in accounting research) where there is an over-reliance on convenience samples. What may be apparent is an attempt by the authors to obfuscate in this regard, to overlook detail and try to create an impression that the sample selection is more systematic than it has actually been. 9 How have questions of validity been addressed? Choice of research method should address issues of validity. Where experimental methods have been employed we would anticipate questions of internal validity to be paramount; where field studies are involved we would expect issues of external validity to be addressed. For survey methods we would anticipate the focus to be more on the reliability of the test instrument and the rigour of the subsequent statistical analysis, rather than on validity issues. 10 How have the results been analysed? We want to see the simplest analysis of the results consistent with the relationships being explored. We do not wish to see unnecessary complexity; this will make the paper less readable and tend to mask the findings and their significance. On the other hand, most academic accountants are only ‘amateur’ statisticians; if the level of their analysis is inadequate, then they may need to bring in a statistician as co-author (evidenced by the number of ‘quant jocks’ appearing as third or fourth authors on accounting papers to satisfy the reviewers). Importantly, we do not wish to see the method of analysis driving the study. In just the same way as the research method should not precede the research question, then neither should the method of analysis. For example, I recall a paper of my own (M. Smith, 1992) presented at a conference but never published. It attempted to show the advantages of using multidimensional scaling (MDS, then a little-used technique in the accounting literature) for problem-solving, but the journal referees rightly observed that the method was inappropriately sophisticated for a relatively simple research question. MDS was abandoned, simpler methods instituted and the revised paper eventually published as Smith (1996). 11 Are the conclusions and recommendations consistent with the findings? Effectively, does the paper hold together? Is the title appropriate? Do the abstract and introduction lead us to expect what we find at the end of the paper? In many papers the final sections are the weakest and may not do justice to the breadth of the research conducted. We look for explanations, limitations and a future research agenda. 12 Research Methods in Accounting Let us now consider how this framework may be applied to a critique of a published piece. Naturally, I chose one of my own publications (Smith et al., 2001) for the treatment because a knowledge of the history of the development of the paper, from an insider’s perspective, can be most instructive. Readers will be able to make the most of the subsequent discussion if they are first able to read a copy of the paper, and for this purpose the complete paper is reproduced as Appendix 2. 1 Interesting new insights: The paper posits an interesting connection between (1) audit firm; (2) manner of conduct of the audit; and (3) classification of audit firms based on their procedures and culture. The paper also attempts to impose a global perspective by employing findings from the USA, the UK and Australia. But neither the data nor the supporting literature is new, and it compromises the originality of the paper. 2 Importance: The paper is important if it makes a contribution to knowledge. This may be a contribution to theory development or implications for business practice. If the paper can demonstrate a relationship between ‘auditor’ and the manner in which the audit has been conducted, then this makes a contribution, even though it may only be of historical relevance. Such a relationship is shown for 1987/88 data, but evidence is also presented to suggest that this relationship no longer holds. The absence of a current relationship suggests that the paper has no implications for current auditing practice. The reasons why a relationship between the audit firm and its propensity towards tolerance of particular accounting policies among its clients is by no means clear. 3 Motivation: The timing of the paper is problematic. It is published in 2001 but uses data predominantly from 1987/88. There is a danger of its being regarded as a historical piece with little relevance to current practice. The authors justify the use of this dataset in that the Kinney classification, the target test of the paper, is based on data relating to the Big 8 group of accountants, with 1988 being the last year of existence of the Big 8 in Australia, prior to extensive merger activity in the sector. There is the suggestion, though, both from the paper itself and the references cited, that the data have been used primarily to generate failure prediction models for the Western Australian government (i.e., Houghton and Smith, 1991) and that the further use of this data in this paper may be incidental and opportunistic. 4 Problem statement: The problem statement is quite clearly stated as: Accounting Policy Changes = f {auditing firm}, where both sides of this equation are elaborated and measured for a large number of companies: • Accounting policy changes: discretionary/mandatory; income increasing, income reducing, neutral. Introduction and Overview • • 13 Auditing firm: by individual name, and by grouping according to classifications developed by Kinney (1986) and Moizer (1998). A number of extraneous variables (notably firm size, financial performance and industry) are also examined to determine their impact. 5 Theoretical framework: This remains something of a problem with this paper, despite strenuous efforts to overcome omissions. The literature demonstrates that there are differences between auditors, and in the procedures that they adopted for audit in 1988 (i.e., Cushing and Loebekke, 1986; Sullivan, 1984). However, why these procedural differences between auditors translate into differing tolerances towards income-impacting accounting policy changes is unclear, and is largely attributable to unpublished anecdotes from practising auditors and the discussion arising in a single paper (Dirsmith and Haskins, 1991). 6 Motivating literatures: Relatively few articles, noted above (i.e., Sullivan, Kinney, Cushing and Loebekke, Dirsmith and Haskins) motivate this paper, while Terry Smith (1992) and Peter Moizer (1998) provide the opportunity for UK comparisons. The pivotal paper is Dirsmith and Haskins (1991), published after the conduct of the data collection; there is thus a strong suspicion that interesting findings have arisen from data mining operations in 1988, for which Kinney (1986) provides a conceptual framework, but that publication must wait for a suitable theory. There is very little other supporting literature, though self-citation by the authors is also revealing: • Houghton and Smith (1991) relates to failure prediction models constructed with the same data and is employed here to provide a measure of overall financial performance; • Smith (1998a) reports current UK findings linking auditor with attitude to accounting policy change; • Smith and Kestel (1999) update the present study with a time series analysis, but the results are apparently insufficiently interesting to constitute publication in a refereed journal; • Brown (1988) reports on the most appropriate means of conducting statistical tests with contingency tables. 7 Research method: Archival methods are employed, since they are the only realistic alternative given the nature of the data: namely, historical, documentary, and covering many companies that are no longer in existence. The authors’ access to a dataset comprising the population of Western Australian public companies is a considerable strength of the paper. Data collection is meticulous and involves checks for consistency both between individual researchers and for temporal validity. 14 Research Methods in Accounting 8 Sample selection: The paper accesses the annual reports of all 463 publicly quoted companies in Western Australian (WA), so does not encounter any sampling issues other than a restriction on the nature of statistical tests that may be employed because of using a population rather than a sample from a normal distribution. 9 Validity issues: There are potential internal validity threats consequent upon the failure to consider competing theoretical explanations for the observations. The incidence of accounting policy change is apparently associated with auditing firm, but both the direction of causation for the relationship and alternative auditor motivations might be considered. The authors acknowledge the lack of external validity in the study – the applicability of the findings to other time periods and other datasets – in that conditions have changed so substantially since the data collection period that the procedures adopted by all auditors are now very similar. 10 Analysis: The fundamental analysis is relatively unsophisticated, involving the comparison of ‘observed’ and ‘expected’ frequencies through a chisquared test. A variation on the traditional approach is introduced to take account of an ordering effect in the contingency tables, the power of the tests being increased with the use of Kendall’s-tau. (A co-author with specialist statistical publications has been included to address testing issues, potentially in response to reviewer concerns on previous versions). A comparative fundamental analysis for UK data (alluded to in Smith, 1998a) is apparently not possible, and further analysis is restricted to tertiary sources. 11 Conclusions: There are no formal conclusions or recommendations, rather a discussion of other interesting findings in related fields which may impact on the integrity of the outcomes. The findings of this study are linked to merger activity in the Big 8, showing a pattern with considerable similarities to past successes. The paper suggests that future merger activity in the sector may be influenced by the organisational culture aspects of the Kinney classification and the clustering of companies generated by Moizer (1998); thus if we were looking at potential suitors for Arthur Andersen, say, then the analysis suggests that Ernst & Young would provide potentially the most successful alternative. Such a critique is revealing, giving glimpses of a less-than-optimum approach adopted in the development of this particular paper. Data were collected for the specific purpose of generating failure prediction models for the WA government, and corporate monitoring of distressed enterprises (i.e., Houghton and Smith, 1991). The interesting auditor findings were generated at the same time, but there was no substantive theory to justify the observed relationship – and consequently no research paper. Only with the emergence of new theories (e.g., Dirsmith and Haskins, 1991), which might motivate the study, could further development towards a publishable paper proceed. Introduction and Overview 15 Clearly research is not always simple, systematic and clean – despite the sanitised versions that we read in the published journals. The research process can be both chaotic and exciting, and very rarely proceeds exactly according to plan. Unfortunately, this impression is rarely created by what we read because published pieces usually have happy endings – positive findings and co-operative participants. For a more realistic version of events we must rely on books like this, conference presentations and research workshops! Armed with a critical and sceptical approach to the research of others, we can now start to develop the skills required to conduct competent research of our own, and commence a sequence which will eventually result in the publication of our research findings. 2 Developing the Research Idea CHAPTER CONTENTS • • • • • • • The research sequence Emergence of the research topic Conceptual frameworks The structure of DNA: the development of new theory The Bradman problem: the development of new strategies The longitude problem: implementing solutions Strategic management accounting 16 20 24 30 31 34 36 We recognise in Chapter 1 that research processes are usually neither simple, systematic nor clean because research rarely proceeds exactly to plan. However, this should not deter us from planning thoroughly in the first place to specify how, in an ideal world, we would like the research to be conducted. The research sequence Figure 2.1 specifies the typical research sequence described by Howard and Sharp (1983) as a series of stages we would expect to progress through in most forms of accounting research, while moving from original idea to eventual publication. The following chapters of this book address each of these stages, and detail the constraints we might anticipate. 1 Identify broad area: Narrow the focus from accounting in general to a stream associated with financial accounting, management accounting, auditing, accounting education or accounting information systems. Developing the Research Idea 17 FIGURE 2.1 Identify Broad Area Select Topic Decide Approach Formulate Plan Collect Information Analyse Data Present Findings The research sequence 2 Select topic: Specification of a sub-area to provide a tighter focus, and one for which supervision capacity is available, but one which may be modified in the light of subsequent developments. 3 Decide approach: Early thoughts regarding the approach to be adopted will revolve around the resources available, and in particular access to the necessary data sources. A detailed specification of research methods to be adopted must wait until the literature review has been conducted and theoretical foundations and outline hypotheses have been established. 4 Formulate plan: Milestones and targets should be established at the outset so that it is clear how the research will progress over an extended period. This is particularly important for part-time researchers who may be contemplating study over six or seven years. A Gantt chart, or similar, is very helpful in clarifying the extent of the programme of work and the mutual expectations of those involved, especially if this concerns the relationship between candidate and supervisor in doctoral work. This plan should include target conferences where preliminary findings may be presented, especially where deadlines are important to the candidate. The commencement period can often cause anxiety because of the 18 Research Methods in Accounting perceived need to make swift progress. It cannot be emphasised enough how important an extended period of reading is to ensure that effort is not wasted performing experiments or surveys which subsequently emerge as being unnecessary or fatally flawed. 5 Collect information: Data collection can safely proceed only when we recognise exactly what we want to know, and for what purpose. The planning stage should highlight the period over which we want to collect data; this usually effectively precludes most longitudinal studies, partly because it takes too long to collect data and partly because of the increased vulnerability associated with extended site access. We may require access to commercial databases; if these are an essential requirement then permissions should be sought immediately. 6 Analyse data: Methods of data analysis and software requirements should be apparent early in the research process. 7 Present findings: Preliminary findings will normally be presented at university workshops and seminars, and then at specialist conferences. These provide the precursor to publication in the refereed literature, which may take place before completion of any associated doctoral dissertation. Publication in the professional literature will bring important findings to the attention of interested practitioners. The research sequence above can easily be squeezed to provide the elements of the traditional positivist approach as devised by House (1970), and illustrated in Figure 2.2. Again this approach assumes the presence of specific conditions: • the specification of a priori hypotheses: formulated on the basis of theory and literature before any data is collected or fieldwork contemplated; • the specification of a priori criteria: to measure the acceptability of the hypotheses, most commonly in the form of standard statistical tests; • the isolation and control of the variables to be investigated: determination of which variable(s) will be treated as dependent, which will be independent (explanatory), and which will be held constant, matched or ignored; • the verification of the methods for measuring and the variables: specification of which variables can be measured directly, and how, and those which will require the use of proxy variables, or measurement instruments, of some form. However, we have to acknowledge that there is no single method which necessarily applies to research in all situations. Thus while the positivist tradition remains the most prominent in the accounting literature, non-positivist approaches have become increasingly acceptable. (Even so, some of the top Developing the Research Idea 19 FIGURE 2.2 Problem Literature Review Hypothesis Development Method Results The positivist approach US journals are still unmoved in their attitude to non-positivist approaches. Baker and Bettner (1997), among others, observe that most of the top journals are devoid of interpretive and critical research studies.) However, management-oriented investigations of change (e.g., the implementation of accounting innovations) may be particularly unsuitable to a scientific approach. Where people are involved and multiple variables are beyond the control of the researchers, including management’s own motivation and agenda, scientific approaches are of questionable validity. Checkland (1981, p. 316) observes: ‘attempts to apply scientific methodology to real world, essentially social problems, have been responsible for the limited success of management science’. Thus we can stay within our original ‘research sequence’ framework but extend out beyond the positivist approach. Figure 2.3 illustrates the range of possibilities. As we move from top to bottom in the figure, we move from the traditional positivist approaches (archival and experimental studies) through field studies towards a case-based approach typically associated with ethnographic studies. This movement corresponds with an increase in the number of uncontrolled variables, with our increasing inability to formulate testable hypotheses, and with the increasing prominence of the ‘human’ element. 20 Research Methods in Accounting FIGURE 2.3 Archival Research Laboratory Experiments Survey Methods Action Research • hypothesis testing using standard instruments/methods/controls • quantitative data • deductive positivism • quasi-research methods • field-based surveys/interviews Field Research Ethnographic Methods • subjective accounts ‘inside’ situations • no a priori hypotheses • qualitative data Alternative research methods Emergence of the research topic We should begin by choosing a research topic which is of interest both to the researcher and the supervisor, where the project is contributing to a doctoral qualification. The topic should generate enthusiasm in the researcher at the outset, otherwise he or she is unlikely to last the course of a protracted period of study in which motivation is bound to wane, even temporarily. The source of the topic can be from anywhere, but most commonly: • a problem at work with potentially wider implications; • a problem or application spotted in the newspaper or from television; • a conference presentation revealing the directions being explored by other researchers; • working papers and completed theses elsewhere – the contents of which are usually at least two years away from publication; • textbooks – particularly in management-related areas – which are a constant source of untested theories; Developing the Research Idea 21 • review articles, analysis of the literature in a particular area, to reveal the current boundaries of knowledge and a potential research agenda. The Journal of Accounting Literature is a particularly useful source in this regard (e.g., Bauman, 1996; Dunk and Nouri, 1998; Gramling and Stone, 2001; Jones and Shoemaker, 1994). Also good is Accounting Organizations and Society (e.g., Hartmann and Moers, 1999; Langfield-Smith, 1997; Libby and Luft, 1993). • Review monographs (e.g., Ashton and Ashton, 1995, information processing; Brownell, 1995, management accounting; Smith and Gurd, 2000, behavioural issues; Trotman, 1996, auditing) are also helpful, as is Foster’s (1986) excellent book, which unfortunately has never been updated since this second edition. • Refereed journal articles, particularly the final sections, revealing flaws in existing research, gaps in our knowledge and research opportunities. The ideas will rarely emerge, therefore, from a ‘spark’ of original thought. More likely, the thought development will have emerged elsewhere, with the originator having either discarded his or her ideas or not seen their full value. It may fit an ‘added value’ concept in the same way as a successful innovation may be remote from the inventor. This may be in the form of relating two concepts from different disciplines, in a manner which provides an application opportunity in the accounting environment. Here the ‘success’ is the publishing of research findings in a respectable journal. To do so we will inevitably be building on the work of others. The common element in each of the above approaches is ‘reading’ – hence the common advice given to doctoral candidates of ‘reading, reading and yet more reading’ to know an area and spot the opportunities. Candidates usually have a much greater commitment to a topic if they have developed it themselves, yet many find idea-generation an extremely difficult process. Thus it is not uncommon for the supervisor to be the source of the research idea, because active and experienced researchers usually have far more ideas than they are capable of exploring by themselves. As Gill and Johnson (1997, 2002) observe, topic selection can be risky if left entirely in the hands of the candidate; the chosen topic may prove to be too small, too large, or simply not feasible in the time frame (especially for longitudinal studies). There may need to be a trade-off between the ownership/commitment associated with a studentselected topic and the practicability/timeliness expected of a supervisordirected preference. Commonly, candidates will contemplate studies which involve the implementation of an accounting change, in order to monitor the change process and the resulting impact on financial outcomes. Such a model is rarely feasible because, apart from the access problems, involvement and data collection will be necessary over a period usually extending beyond that permitted within a standard candidature. Once the germ of an idea is forthcoming it must be worked over to see if it really constitutes ‘research’. For example, are we sure that it is more than a consultancy project? Is it more than a trivial problem with no wider 22 Research Methods in Accounting implications? Is it more than a replication of something someone else has done before, or done in a different industry or different country? If we are happy in this regard then several other questions emerge: • Is the project ‘doable’ in a reasonable time frame (e.g., the period of candidature)? • Will the project fit the NIRD acronym (usually attributed to Rashad Abdel-Khalik during his tenure as editor of The Accounting Review) in that it is new, interesting, replicable and defendable? Unsurprisingly, the acronym fits a positivist outlook since replicability may be impossible to guarantee in field study settings. • Will the data required be readily available? If site visits are required, will access be available over a sufficient time period and to a sufficient depth? This last scenario is of great practical concern and difficult to control. Young and Selto (1993) report on a study whose depth was seriously curtailed because management changed its mind and restricted the access to personnel due to be interviewed. Worse than that, there are numerous cases of change of company ownership during the data collection period, resulting in further site access being permanently denied. Once the general topic area has been determined, it may be refined by formal methods (e.g., brainstorming, attribute listing, etc.) to identify possible fruitful directions and potentially interesting relationships, and to eliminate blind alleys. Diagrammatic aids, particularly white-boards, are very helpful at this stage for mapping ideas, variables, relationships and processes. If you are in any doubt as to what constitutes ‘research’ and what constitutes a study that is published in a reputable refereed journal, then the reader is referred to recent examples from the professional literature. Often the professional journal will have interesting ideas and will convey a message with implications for practitioners (its essential purpose), but will not have the essential conceptual framework of a refereed research piece. Consider, for example, the contribution to the professional literature made by Smith and Briggs (1999). The paper is both amusing and interesting in making observations in a number of areas, and is reproduced in its entirety as Appendix 3. It addresses: • the traditional stereotype of the accountant as a boring individual with poor interpersonal skills; • the portrayal of the accountant in the media (particularly in films), initially as a comic incompetent, but latterly as an unethical, unprofessional manipulator with criminal tendencies; • the accountant, almost exclusively, seen as a white male occupying a subordinate position; • the concern in the accounting profession about its failure to attract the best candidates, who are lost to other professions; Developing the Research Idea 23 • the concern that the characteristics which the profession tries to convey about accountants (financial expertise, interpersonal skills, ethical conduct, openness to all) are at odds with those conveyed by the media. The paper speculates that there might be some link between these issues and suggests that recruitment to the profession is being damaged by the accountant’s image. However, it provides no evidence to support this relationship, nor does it provide any evidence for the manner in which this image may be being influenced by the popular media. Furthermore, there is as yet no theoretical explanation to justify these relationships. Without these foundations we have an interesting idea but no basis for a research project whose findings would interest the refereed journals. However, the idea could potentially yield rewarding research projects, were answers to fundamental questions to be found, and these are considered in more detail later in this chapter. Attention to the fundamental requirements of the refereed literature will allow the researcher to produce an outline research proposal, one that is continually revised during the reading period and may nevertheless have to be revised further during the conduct of the research itself due to unforeseen circumstances. A typical research proposal will include the following elements: • • • • • • • Title – should make it clear what you are trying to do. Abstract – should summarise the problem, objectives and expected outcomes. Issues – why they are interesting and important. Objectives – how the study relates to the problem. Literature – review of relevant, themed publications. Method – the how and why of the process. Benefits – the anticipated outcomes that make this study worthwhile. A working title is important to clarify the topic, especially if external grant income is being sought to support the project. However, the final title is rarely the original one and there are plenty of opportunities to make changes. The abstract is an important departure at this stage, because it allows the researcher to speculate on what the outcomes of the research might be should everything go according to plan. The abstract can be ambitious initially, but will require revision (perhaps radical revision) as problems and constraints emerge in the research process. The contribution of the paper and the way that it aims to address important issues in a systematic manner are fundamental to its success. Internal consistency and overall coherence should ensure that the objectives and the intended approach are appropriate. The outline literature review may be incomplete but it must none the less identify the key motivating literatures and theories. Research candidates must recognise that the literature review is a constantly moveable feast and something that will be added to right up to final presentation of the research findings. One of the major deficiencies of both papers and dissertations is 24 Research Methods in Accounting that they frequently overlook the most recent relevant publications: it can be a heart-stopping moment when one is about to submit and the latest issue of a journal appears to report the outcomes of a research project very similar to one’s own! At the very least this new paper must be cited. A common complaint from inexperienced researchers is that there is ‘no literature’ available. If this is true, it may mean that the projected topic may be too trivial to consider. More likely, however, is that the literature review should drill down further and search on different keywords. That there is a dearth of recent literature on a topic may foreshadow problems. For example, papers on ‘decision-making heuristics’ were common in the late 1970s and early 1980s, and papers on ‘group decision-making’ common in the mid-1980s, but progress in both of these research areas has slowed, and publications are relatively rare because the psychological theories underlying the research have not developed sufficiently to facilitate new approaches. With respect to apparently new projects, for example, research into ‘e-commerce’ related topics, students must recognise that e-commerce is just a new way of doing business, and that their review must address the implementation of prior business innovations. Discussion of method should address the alternatives available in order to demonstrate that the preferred choice is the most appropriate. The proposal should also echo the benefits of the research, in particular its contribution to knowledge and the potential implications for business practice. The research proposal would normally form a central feature of any application for ethics approval, and must therefore demonstrate both the value of the research, the integrity of the methods employed and the extent of the involvement of human participants. Conceptual frameworks A valuable part of the initial planning process is the development of a conceptual representation of the research project. This can help to clarify the important relationships (and the need for supporting theory), the explanatory and intervening variables, as well as the demonstration of causation. The inductive and deductive approaches identified in Chapter 1 provide an objective alternative to the conduct of research, but neither allows the opportunity for human interaction: the inductive approach, where new theory is developed on the basis of fresh observations (as is most commonly the case in hard sciences, like astronomy), and the deductive approach, where theory provides the basis for the testing of empirical observations (and which is the most common form of positivist accounting research). The deductive approach is suitable in a highly structured environment, involving the empirical testing of theoretical models, so that its reliability is dependent on the integrity of quantitative and statistical methods. However, the causal relationships Developing the Research Idea 25 FIGURE 2.4 Concrete Experiences Testing Implications of Concepts in New Situations Observations and Reflections Formation of Abstract Concepts and Generalisation Deduction Induction Kolb’s Learning Cycle explored rely on an internal logic and take no account of the human relationships present. The application of the inductive approach in the accounting environment necessitates a variation to the traditional model such as that provided by Kolb’s Experiential Learning Cycle (Kolb et al., 1979, p. 38), which is illustrated in Figure 2.4. Recognition of the importance of internal processes and human relationships to the inductive approach allows for the existence of human subjectivity without distorting research findings, even though they may be qualitative and not replicable. Where human relationships are central to an understanding of accounting behaviour the approach exploits the subjective environment. Although both models provide opportunities in accounting research, the deductive approach offers greater possibilities for the implementation of scientific methods, since it facilitates arguably more reliable measurement and control. Grounded theory (discussed in more detail in Chapter 8) illustrates the potential for inductive methods in accounting research. We can therefore develop the model of the deductive process (see Figure 2.5) so that it corresponds with Popper’s (1959) defining characteristics of scientific theory: • the theory is capable of empirical testing; • scientists (and researchers) should make rigorous attempts at falsifying theory; • science advances as falsified propositions are left behind, leaving a core of theory still to be disproved. 26 Research Methods in Accounting FIGURE 2.5 Theory Operationalisation of Concepts Test Theory Through Observation Falsified, So Discard Theory Not Falsified, So Predict Future Observations The deductive process FIGURE 2.6 Independent (Explanatory) Variables Dependent Variables Influential Variables The conceptual schema The basic conceptual schema (see Figure 2.6) provides a powerful tool for the examination of causal relationships in a positivist environment. By establishing the key variables of interest, and the other potentially influential factors, we can form a better impression of the breadth of the problem. If, for example, we return to our earlier article on ‘accounting stereotypes’, we can see how the content of a professional piece may be used to develop a conceptual schema for more systematic investigation. As a starting point we need to identify dependent variable(s) – one or more outcome variables whose value may be influenced by a number of (independent) explanatory factors. The scenario outlined extends over a long period of time – potentially from the final years of secondary school through undergraduate years to recruitment to the accounting profession; quite possibly in excess of five years even for straightforward cases. The length of this period, the process of maturation and the incidence of a multitude of uncontrolled variables makes it extremely unlikely that we will be able to predict behaviour in year five based on attributes in year zero. In trying to establish a causal relationship for investigation, it is much more realistic, therefore, to break up the process in Developing the Research Idea 27 order to be able to focus on events during more finite time periods. Such a reorientation allows us to focus on two ‘choices’ which will permit the suggestion of two measurable dependent variables: • the selection of commerce/accounting as the preferred undergraduate course by school leavers; and • the decision to enter the accounting profession by those nearing the end of their university course. The major dependent variable, in each case, as suggested by the professional article, is the ‘image’ of the accountant, though it is unclear how this concept is to be either defined or measured. In each case there is likely to be a potentially large list of influential variables that may impose conditional relationships: for the former, these include teachers, parents, schoolmates, perceptions of other courses, job prospects, etc.; and for the latter, these include income, career prospects, perceptions of competing professions, etc. Both of these ‘models’ in their simplest form ignore how the perceptions of the accounting profession have been formed (e.g., did film and media have any influence whatsoever, or are there other more important influences?), and whether entry into the profession can be regarded as a single decision, irrespective of the claims of the competing recruiters (e.g., individual members of the Big 5, the public sector, commerce). In terms of our example(s), we can only sample relatively small proportions of both school-leavers and potential recruits to the profession. In neither case will we have access to population statistics. In addition, there may be problems with interviewing school-leavers prior to university entrance, especially if we are contemplating the conduct of psychometric testing to examine personality issues and different cognitive styles. Such an investigation is better conducted when the decision has been made, early in the first year of university courses. This has the advantage of providing access to mature entrants and of looking at an actual decision rather than a choice intention, but it has the disadvantage of potential hindsight bias. Data collection at this time may determine the motives for choice of entry into competing professional courses (e.g., accounting, law, medicine). For the decision point prior to entry into the accounting profession, some collaboration with the major recruiters (e.g., the Big 5, state government) is advisable; they are all interested in attracting the ‘best’ students and make clear statements on what they are looking for among successful applicants. The researcher may gain access to student information, with appropriate permissions and participation, to identify those selected for interview and those subsequently offered positions. A comparison of recruiter decisions with student profiles should facilitate an examination of the transparency of recruitment policies. The design problems associated with the identification and measurement of variables means that the basic schema can be modified (see Figure 2.7). 28 Research Methods in Accounting FIGURE 2.7 Independent Variables Dependent Variables Moderating Variables Identify and Measure Identify and Control Identify and Measure Measurement issues In terms of our example, we have serious problems in defining and measuring the dependent and independent variables, and in controlling influential variables: • ‘Image’ cannot be measured directly. The literature will help in identifying those variables which contribute to the creation of the concept. This process may allow the identification of variable(s) that can be measured directly and used as a proxy for ‘image’. This is a far from satisfactory arrangement but may be the best we can do if the investigation of these relationships is to proceed. • Decision choices will have been influenced by many variables over long periods of time. We must restrict our focus to those variables cited in the literature as being potentially influential, and try to gauge their relative importance to decision-making in this context, possibly through pilot surveys or interviews after the event. The use of matching procedures (e.g., sampling candidates from the same schools, of the same age and gender, and with equivalent academic profiles at entry) will help to provide some control over the influence of extraneous variables. We clearly have some way to go before this interesting research idea becomes a feasible research project. Questions relating to theory and measurement remain problematic, but a search of the literature specifies a number of theories with precise implications for this scenario. Friedman and Lyne (2001) examine the development of the ‘beancounter’ stereotype and its implications for the future of the accounting profession. They identify a number of sociological theories which could assist the establishment of our conceptual framework: • Hamilton and Troiler (1986, p. 13) suggest that stereotyping is determined by cognitive, motivational and social learning processes. Social Developing the Research Idea 29 learning theory suggests that stereotypes are mostly acquired through socialisation, via variables such as parents, schools and mass media, but that they may be modified by subsequent experience. • Campbell (1965) suggests that realistic conflict theory may explain how stereotypes result from inter-group competition for scarce physical resources. • Tajfel and Turner (1985) develop a social identity theory which modifies conflict theory to take account of individual achievement motives so that conflict can also occur over scarce social resources like prestige and status. • Stroebe and Insko (1989, p. 14) recommend the simultaneous consideration of the three approaches: social learning theory, social identity theory and conflict theory. The implications of these theoretical approaches to our research idea are instructive: • The perception of a negative accounting stereotype is unlikely to fade among school students or accounting undergraduates while it continues to be widely perpetuated (social learning theory). • The negative associations are likely to impact on candidates’ choice of career profession and the associated income and prestige (social identity theory). • The inter-group conflict between the professions, when each is trying to attract the ‘best and brightest’ candidates, means that they may seek to perpetuate the negative stereotypes of other professions to their own resource advantage (realistic conflict theory). The inference here is that we may extend the research to consider the advertising material used by the competing professions (and by competitors within the accounting profession) and exposure of students to these. We may anticipate that recruitment at school level involves the denigration of the accounting profession by other professions, and at university level the denigration of some members of the profession by fellow members from rival organisations. What is also apparent is that the simple conceptual schema of Figure 2.6 has outlived its usefulness in this context. It was helpful to begin with but must now be modified to allow the further development of hypotheses and data collection methods. Chapter 3 focuses on these issues in the development of the important links between theory, literature, hypotheses and methods. In Chapter 1 we acknowledged our debt to researchers in the natural and social sciences. It is now helpful to turn to popular literature, which provides highly readable explanations of complex situations, for an insight into the emergence of research ideas, the development of research questions and the airing of potential solutions for testing, to illustrate the research sequence in practice. It is instructive to consider the different scope of original research in solving practical problems by examining the particular aspects of three welldocumented stories from non-accounting environments. Thus we explore the 30 Research Methods in Accounting development of new theory in a chemical environment in ‘The Structure of DNA’, the development and testing of alternative strategies to address a sporting issue in ‘The Bradman Problem’, and the solution of apparently insuperable implementation issues to a problem where the ‘answer’ was well known in ‘The Longitude Problem’. The structure of DNA: the development of new theory James D. Watson’s (1968) The Double Helix (subsequently filmed as Life Story) provides a brilliant description of the exciting process of discovery in scientific research, even if the approach adopted is rather unorthodox. The development of theory and conceptual modelling, from systematic deductions based on the empirical findings of others, is conducted in a competitive environment where the ultimate prize for winning the ‘race’ is the Nobel Prize. (Watson, together with Francis Crick, both of Cambridge University, and Maurice Wilkins, of King’s College, London, were awarded the 1962 Nobel Prize for Physiology of Medicine for their pioneering work during 1951–52.) Sir Lawrence Bragg reflects on these achievements by researchers in his Cavendish Laboratory through a revealing preface to Watson’s book, with implications for research ethics: He knows that a colleague has been working for years on a problem and has accumulated a mass of hard-won evidence, which has not yet been published because it is anticipated that success is just around the corner. He has seen this evidence and has good reason to believe that a method of attack which he can envisage, perhaps merely a new point of view, will lead straight to the solution. An offer of collaboration at this stage might well be regarded as trespass. Should he go ahead on his own? (Bragg, in Watson, 1968, p. vii) That Watson and Crick, at the Cavendish, did proceed with a belated relationship – though collaboration is too strong a word – with Wilkins and Rosalind Franklin is a matter of history. The course of their investigations, and the factors leading to their success in developing a new theoretical model, have implications for all research. Essentially, they proceed to develop a model which fits all of the evidence currently available to them, and await confirmation or disconfirmation of their framework from the empirical findings of others: • The work of Linus Pauling in the USA on a helix formation for polypeptide chains suggested that DNA (deoxyribose nucleic acid) too had a helical structure. Pauling’s early attempts at modelling though, without crystallographic evidence, had produced stereochemically impossible components. Early evidence from X-ray crystallographic diffraction presented by Wilkins also seemed to suggest a helical structure, but there was no Developing the Research Idea 31 evidence of whether a single, double or triple strand helical configuration was most appropriate. Crick and Watson apparently proceeded on the basis of an educated guess favouring the double helix because most things biological come in twos! • Ernst Chargaff had produced vital evidence on the ratios of constituent bases, and particularly the equalities existing between adenine (A) = thymine (T), and guanine (G) = cytosine (C). The A–T, G–C flat hydrogen-bonded base pairs formed the core of the Crick and Watson structure, rather like a spiral staircase in which the bases form the steps. • Important advice from a structural chemist colleague (who just happened to be sharing the same office with Crick and Watson) suggested that the normal textbook formulation of the A–T, G–C bases was incorrect and that they should work with an alternative ‘keto’ form. Without this important questioning of textbook content, and accepted knowledge, their structure would not have held together. The impact of this finding was that a given chain could contain both purines and pyrimidines (with the capacity to carry the genetic material for self-replication) and that the backbones of the chains should run in opposite directions. Thus Crick and Watson were able to construct a physical model comprising two intertwining helically coiled chains of nucleotides, right-handed and running in opposite directions, with complementary sequences of hydrogenbonded bases. The resulting structure was stereochemically possible, and subsequent X-ray evidence from Franklin confirmed that the sugar-phosphate backbone was indeed on the outside of the molecule. The more general implications for researchers are constant vigilance and a questioning attitude to the work of others and existing publications. The Bradman problem: the development of new strategies Test match cricket in the 1930s was dominated by a single, outstanding individual whose unique gifts of batsmanship threatened to change the way the game was played. Donald Bradman scored so many runs, and scored them so quickly, that he was likely to win a game almost single-handedly. Ashes test series, once so closely contested, threatened to be one-sided affairs, with Australia the perennial victor. In the 1930 England v. Australia test matches Bradman totalled a record 974 runs at an average of 139.14 per innings, and recorded separate scores of 334, 254 and 232. At the commencement of the 1932/33 series he averaged 112.29 in all test matches, and had already posted six scores in excess of 200 in only 24 completed innings. Bradman was a phenomenon, some would say a ‘freak’, and curtailing his dominance became a pressing question for successive England captains. 32 Research Methods in Accounting ‘The Bradman Problem’ is capably detailed by Lawrence Le Quesne in his book The Bodyline Controversy (1983), and presents an intriguing research question, investigating a number of alternatives that might provide successful solutions. 1. Changing the playing conditions Cricket was played on hard, fast wickets, largely true, though with the occasional unpredictable bounce. If bowlers did not take wickets in the first few overs, when the ball was still shiny and swinging, then they might have a lot of overs to bowl before they got a replacement ball, by which time 200 runs had been scored. Wickets were uncovered and exposed to the elements once a game had started and could become unplayable as a hot sun dried out a wet pitch; these ‘sticky’ wickets provided a possible solution to the problem because Bradman was nowhere near as prolific on bad wickets as many of his contemporaries were. However, captains could hardly rely on this occurrence to blunt Bradman’s genius on a regular basis. 2. Changing the rules of the game Test matches were timeless in the 1930s and played to a finish. Declarations were rare and slow play very common. Further, the leg-before-wicket (LBW) rule made it difficult for the batsman to get out in that way – he had to be caught in front of the stumps by a ball pitching in line, wicket-to-wicket. Life was difficult for bowlers, and a number of other batsmen (notably Ponsford for Australia and Hammond for England) regularly completed double and triple centuries, although none with the regularity, reliability or speed of Bradman. Reduction of the specified playing time to five (or four) days would mitigate slow play and dull the impact of less talented batsmen content to occupy the crease, but would not affect Bradman – witness his 309 in a day against England at Leeds in 1930. However, timeless tests were outlawed before the end of the inter-war period. Similarly, the LBW law was changed so that batsmen could be given out to a ball breaking back from outside the off stump and striking the pads in front of the wicket. 3. Changing the bowling Bowlers had often sought to restrict the batsman by bowling outside of the leg stump (e.g., Hirst and Foster in the early 1900s), ostensibly to cut out offside shots but also to restrict the on-drive through bowling just short of a length. An accurate bowler could therefore depress the scoring rate by keeping the batsman to deflected singles in the arc between the wicketkeeper and square leg. As a consequence the whole game is slowed to a snail’s pace – a Developing the Research Idea 33 strategy still used today to curb an aggressive batsman’s dominance. But Bradman was too good to fall for such tricks and used his agility, quick reflexes and nimble footwork to move to leg, outside the line of the ball, so giving himself room to hit through the off-side field. Such a strategy would have made less gifted batsmen vulnerable because it entailed the exposure of all three stumps to the bowler, but Bradman could execute the manoeuvre without undue risk. 4. Changing the field placings An orthodox field will involve the placing of fieldsmen on both sides of the wicket, with a majority coincident with the bowler’s planned line of attack. This means plenty of gaps in the field and, for inaccurate bowlers, plenty of scoring opportunities. Clearly, a preferred strategy would be to place all the fielders in a tightly confined space and have the bowler deliver a line and length that forces the batsman to play the ball there. This is a sound strategy for an accurate bowler – witness Laker’s devastating use of a packed leg-trap to a turning ball in 1956 – but for a batsman of Bradman’s ingenuity we now have a vacant off-side to be penetrated by inventive, unorthodox shots. 5. Introducing a ‘bodyline’ attack While a combination of numbers (3) and (4) above provides a partial solution in restricting scoring opportunities, only impatience will induce eventual dismissal. For Bradman it may do neither. However, together they provide the basis for a potentially successful solution: ‘leg theory’ will dictate the line of delivery, but we need also to control the height of the delivery to induce the batsman to give chances from playing the ball in the air. Herein lies the Jardine–Larwood proposition, initiated by Douglas Jardine, England’s captain in 1932 and executed brilliantly by Harold Larwood, the fastest and most accurate bowler of his time. Extreme fast bowling does not provide even Bradman with the time to move outside the line consistently without risk. The introduction of a high proportion of short-pitched balls rearing towards the throat or the rib-cage of the batsman makes scoring without risk extremely difficult. Batsmen are likely to fend off the ball defensively – to be caught in the leg-trap ring of close catchers. If they try to attack by hooking the ball over the in-field, they fall prey to a number of deep-set fielders on the leg-side and behind the wicket. Concentrating almost all the fielders in the arc between wicketkeeper and mid-wicket, five close to the bat and three close to the boundary, covers almost all the options. Scoring is restricted to risky options and bad balls. This form of attack was ‘successful’ in that it resulted in a 4–1 series victory for England and also provided an appropriate solution to ‘the Bradman problem’ in that he scored only one century in the series, a total of only 396 runs at an average of just 56.57. He still tried to play in a cavalier fashion, 34 Research Methods in Accounting moving outside the line to play tennis-like shots; the result was brilliant, hectic cameos that were over all too soon for his team’s requirements. However, the risks of bodily injury to the batsman from the bodyline solution were high, and its introduction was seen to be ungentlemanly and against the spirit of the game. Jardine and Larwood were never chosen to play against Australia again. The subsequent 1934 test series was very much a fence-building exercise, with the England bowling friendly and ‘bouncer’ free. Bradman was again unconstrained and scored 758 runs at an average of 94.75. The success of the leg theory solution generated further changes to the rules of the game, with a restriction on the number of fielders permitted on the leg-side behind the wicket and the number of short-pitched balls that were allowed to be bowled per over. The mo...
Purchase answer to see full attachment