British Accounting Review (1998) 30, 313–330 Article No. ba980074
UK Academic Accountants’ Perceptions of Research and Publication Practices TONY BRINN, MICHAEL JOHN JONES and MAURICE PENDLEBURY Cardiff Business School Published research output is widely used as a measure of performance in UK universities. This paper investigates the acceptability of the research practices which underpin publication in accounting journals. Eighty-eight UK accounting academics reported their perceptions of the acceptability of ten research practices and their views of five research publication statements. Three research practices were found to be particularly unacceptable, but were perceived to be frequently used. The paper concludes by listing nine guidelines which have implications for authors, editors and reviewers. 1998 Academic Press
INTRODUCTION Published research output is a widely used measure of the performance of both UK universities and individual academics (Lyall, 1978; Gray et al., 1987; Gee & Gray, 1989; Hutchinson, 1989; Gray & Helliar, 1994). Indeed, over time, the emphasis upon research has increased (Peasnell & Williams, 1986). At the institutional level, universities are graded and funded as a result of Research Assessment Exercises (RAEs) (for comment on this process, see Mace, 1993; Whittington, 1993; Puxty et al., 1994; Taylor 1994; Humphrey et al., 1995). Perhaps as a consequence, individual academics’ reward and promotion structures are becoming increasingly linked to research output. The pressures to publish have never been greater. Moreover, publication in the highest quality journals is particularly wellregarded. Indeed, journals are often ranked by quality (for example, Benjamin & Brenner, 1974; Howard & Nikolai, 1983; Nobes, 1985, 1986; Hull & Wright, 1990; Brown & Huefner, 1994; Doyle & Arthurs, 1995; Brinn We would like to thank those accounting academics who responded to our questionnaire. This paper has benefited from the comments of those at the 1997 British Accounting Association conference in Birmingham and of two anonymous referees. Any of the three authors will be pleased to receive correspondence at: Cardiff Business School, Cardiff University, Colum Drive, Cardiff CF1 3EU, UK. Received November 1996; revised October 1997; accepted April 1998. 0890–8389/98/040313+18 $30.00/0
1998 Academic Press
314
. .
et al., 1996). The highest ranked journals are generally considered the most prestigious. Typically, the most prestigious journals operate peer review. Authors submit manuscripts which are then distributed by the editors to reviewers. In theory, the reviewers and authors will be anonymous to each other. In practice, the likelihood is that because of prior exposure of the paper, the reviewers will be more readily able to identify the authors than the converse. The editors judge whether or not to accept papers on the advice of the reviewers, but also on their judgement of the paper’s quality and contribution. The peer review process has both an explicit and implicit agenda. Explicitly, the process enhances the quality of academic research, makes the research process more efficient and puts an imprimatur on an accepted paper. Less explicitly, a powerful minority or ideological elite are permitted to act as determinators of received wisdom (see Puxty & Tinker, 1995, pp.244–45, for a fuller discussion). All parties to this process (editors, reviews and authors) are bound by a set of rules and guidelines. However, these guidelines are often more implicit than explicit, are often not codified and are gained by experience rather than formally learnt. Indeed, a whole set of largely unwritten rules or conventions form a framework within which the research process operates. Despite the important and often subjective nature of these conventions, we know of no empirical accounting research conducted in this area in the UK.1 However, a limited amount of research has been conducted in the USA (Engle & Smith, 1990, 1992; Loeb, 1988, 1990, 1994). Research on acceptable publishing practices in the USA should not, of course, be seen as definitive in a UK context, given the differing environmental, organizational, professional and social contexts in the two countries. This lack of research is particularly important in the UK context for two reasons. Firstly, with increasing pressure upon academics to publish, there is a clear danger of the erosion of the conventions guiding research and publication practices.2 Authors may, for example, be tempted to ‘cut corners’ or take advantage of ambiguities in research practices. Reviewers may face similar pressures, especially as reviewing is typically ‘invisible and unrewarded’ (Puxty et al., 1994, p.163). Secondly, this potential erosion of the acceptable standards of academics is compounded by the lack of guidance as to what is and what is not acceptable practice. Although some practices, such as plagiarism, appear to be widely condemned, many more practices, such as the inclusion of names of minimal contributors to a paper, appear more equivocal. Indeed, in accounting, we know of no agreed codification of practices. Nor are there any consensual mechanisms by which penalties are inflicted if accepted practices are violated. Indeed, the situation in accounting appears analogous to that in organizational behaviour: ‘no formal mechanisms exist for communicating standards or for dealing with violations of those standards’ (Von Glinow & Novelli, 1982, p. 428) and ‘no code of ethics, no basis for consistent peer control, and no structures . . . currently
315
serve to maintain professional standards. Few standards of behavior are even widely agreed upon’ (Von Glinow & Novelli, 1982, p. 434). The purpose of this paper is to provide guidance on acceptable practices relating to research and publication in the UK by examining the attitudes and perceptions of a sample of leading UK accounting academics to a series of issues relating to research publication practices. This guidance should prove useful to individual researchers, referees and editors. In addition, we hope to stimulate a debate on the likely impact that the pressure to publish will have on research publication practices and to encourage more research into this area. Peer assessment, by its very nature, depends upon attempting to measure the attitudes of individuals. Thus, inevitably, a wide range of opinions will be recorded. Different individuals will have different views of the research process contingent upon their research interests, past experiences and present position. These differing life experiences will inevitably lead to varying stances towards research and publication practices. Individual scores, as captured by their recorded perceptions, will thus inevitably be subjective. However, the study of acceptable research practices is necessarily a subjective area. For example, Kerr et al. (1977) find that the acceptance or rejection of manuscripts by management and social science gate keepers is contingent upon individual predispositions and attitudes rather than more objective editorial norms, formats and policies. For example, senior and non-senior staff will tend to have different responses to research and publication, in general, and to the RAE, in particular. Non-senior staff may be concerned to build and develop their publications record in order to demonstrate a research reputation, so as to enhance their promotion prospects. Senior staff will not be subject to the same pressures. As a consequence, one might expect non-senior staff to be particularly concerned with practices which might damage their reputation, such as not being given due authorship credit. The remainder of this paper is divided into three sections, followed by a conclusion and recommendations. Section two, which follows, presents our review of the relevant literatures on the acceptability of research and publication practices. In section three, we detail our methods. The most substantive section of our paper, section four, outlines and discusses the main results and limitations of our study.
LITERATURE REVIEW Accounting ethics We know of no UK studies which have specifically investigated the acceptability of accounting research practices. In the USA, the acceptability of accounting research is typically subsumed under broader studies dealing
316
. .
with accounting ethics. For example, a number of US studies have been conducted into the ethical standards of accounting academics (Engle & Smith, 1990, 1992; Loeb, 1988, 1990, 1994). These US studies have often been extremely wide-ranging, and have typically not focused upon the ethics of research and publication practices.3 In relation to the ethics of accounting research publications, the study of Engle & Smith (1990) is most pertinent. Engle & Smith (1990), through the use of a mailed questionnaire, analysed the responses of US academics to questions about the ethicality of a range of practices, such as padding expense accounts, using university equipment for personal activities, accepting bribes from publishers for textbook adoption and a variety of student-related activities. Engle & Smith investigated seven major research activities (out of 29 overall activities): plagiarism; journal manuscript submissions; falsification of documentation;4 falsifying research data; graduate co-authorship; colleague co-authorship; and presenting research to more than one meeting. They found that over 98% of the 262 usable responses they received viewed the falsifying of research data, the falsifying of documentation for research grants and the plagiarizing of research as at least moderately to extremely unethical. In addition, at least 64% of their respondents believed that multiple manuscript submission (66%), the inappropriate granting of co-authorship status to a colleague (64%) or multiple presentation of the research at meetings (68%) was at least moderately to extremely unethical. Business and management research ethics We are aware of only two studies, both US, dealing with research practices in the business and management area: Von Glinow & Novelli (1982) in organizational behaviour and Sherrell et al. (1989) in marketing. Von Glinow & Novelli (1982) investigated 315 ‘gatekeepers’, i.e. editors, associate editors and review board members of leading organizational and behavioural journals as well as Academy of Management divisional chairpersons. The survey addresses five general areas: composition of review boards; review board members as reviewers; the review process; editor/program chairperson’s role; and submission and authorship credit. The results of the study listed eight inappropriate behaviours: making trivial manuscript changes in order to market it as a ‘different’ manuscript; milking data in order to increase the number of publications; assigning authorship based on prestige; using work without assigning appropriate credit; submitting substantially identical manuscripts simultaneously to two journals; employing students to do work, without giving appropriate credit; altering data to fit theory or increase statistical significance; and agreeing to review a manuscript and then distributing the manuscript to a graduate student. The study by Sherrell et al. (1989) into the research and publishing behaviour of 328 marketing academics provides some interesting background
317
information. These authors investigate 15 editor practices, seven reviewer practices, nine authorship credit practices, eight multiple submission practices, and four general authorship practices. Their respondents found the following practices the most unethical (scored out of 7, with 1 being unethical and 7 being ethical): a reviewer steals an author’s ideas after rejecting the manuscript (1·13); an author alters data to conform to a theory or increase significance (1·27); an editor or reviewer displays favouritism (1·36); a reviewer agrees to review, a graduate student then reviews, but the reviewer still lists his/her name (1·45); authors share manuscript contribution credits reciprocally without contribution (1·72); an author immediately destroys the original database (1·77); an author submits the same basic manuscript to multiple journals with trivial changes to differentiate (1·80); a prestigious name, a non-contributor, is added to the manuscript (1·82); and an editor selects reviewers with a strong bias (1·89). METHODS The data for this study were collected as part of a wide-ranging survey of UK academic accountants’ perceptions of journal quality5 and of research publication practices. The survey was undertaken by means of a questionnaire mailed over the summer of 1993 to a population of 260 ‘active’ academic researchers (241 from ‘old’ universities and 19 from ‘new’ universities). The ‘active’ researchers were drawn from the 1992 British Accounting Review Register (BARR) and consisted of all accounting academics who had published at least one academic/research article during the years 1990 and 1991 (the period covered by the 1992 BARR). Professors who had not published an academic/research article in this period were also included because their professorship was taken as being indicative of a strong, past research record. A total of 88 usable responses were received (a usable response rate of 33·8%). Of these, 38 were from professors (40% of the 94 professors in the sample frame), two were from readers and 19 from ‘old’ university senior lecturers. This means that 59 (or 66% of our sample) of the respondents were ‘senior’ staff. This seniority is likely to reflect a good awareness of the research and publication process. The respondents were asked to provide personal information about themselves. This indicated that 94% were from ‘old’ universities, 89% were male, 65% had a professional accounting qualification, and 37% had a PhD. Of those respondents who provided details of their age and length of service, 72% were over 40 (with 55% being between 40 and 49) and 72% had over 10 years service. When asked to rate themselves as a researcher (ranging from 1 ‘not at all active’, to 5 ‘very active’), 47% ranked themselves as ‘very active’, with a further 37% ranking themselves at point 4 on the scale. These self rankings were confirmed by the researchers’ publication
318
. .
record. There was a high positive correlation (at the 1% level) between selfrank as a researcher and the number of journals in which they published. None of the respondents ranked themselves as not at all active (point 1) which, given the criteria by which we selected the sample frame, is not surprising. We were thus sampling ‘successful’ academics in terms of publications. These academics may not have the same perceptions as the population of UK accounting academics as a whole. Nonetheless, our sample, closely approximated to Gray & Helliar’s (1994) analysis of the population of UK academics. In their analysis, 66% were professionally qualified accountants with 37% possessing a PhD. In our sample, 65% possessed a professional accounting qualification, with 37% possessing a PhD. A non-response bias test was not formally conducted as a postal strike created problems timing the receipt of the questionnaires. Our questionnaire comprised ten questions relating to possible research practices and four questions concerning research publication issues (see Tables 1 and 2 for a list of these questions). These questions were derived after a study of the pertinent prior accounting literature (especially, Von Glinow & Novelli, 1982; Sherrell et al., 1989; Engle & Smith, 1990). We focused mainly on those issues where lack of consensus about the acceptability of practices appeared likely. This, thus, excluded falsification and fraud which, we believed, would be considered universally unacceptable. Respondents were invited to make further comments when completing the questionnaire. We discuss selected comments in the text, where appropriate.
RESULTS Acceptability and perceived frequency of occurrence of research practices Respondents were asked to indicate their views of the acceptability or unacceptability of various research practices on a scale ranging from 1 (always unacceptable), through 2 (normally unacceptable), 3 (neutral), 4 (normally acceptable) to 5 (always acceptable). They were then asked to indicate how frequently they perceived these practices occurred on a scale ranging from 1 (never), through 2 (occasionally), 3 (often) to 4 (very often). The respondents were also partitioned into the categories of senior/nonsenior staff. For this purpose, senior staff were defined as professors, readers and senior lecturers and non-senior staff as lecturers. The reason for this distinction is that senior staff are likely to be more closely involved in the process of undertaking and publishing research because of the emphasis on research output as the basis for promotion in UK universities. Consequently, their views may be different from their less-experienced colleagues. The importance of seniority in ethical reasoning is confirmed by Ponemon (1990, 1992). Ponemon (1992, p.239) finds that US Certified Public Accountants (CPAs) are subject to ‘ethical socialization’, whereby those progressing to
Table 1 Acceptability and perceived frequency of research publication practices Acceptabilitya
Perceived frequencya
Senior
Research practices
n
Mean
n
Mean
1. Replicating other researchers’ methodology without giving due credit 2. Failing to give equal recognition to all colleagues who have contributed to a research paper by not including them as co-authors 3. Including someone who has made only a minimal contribution to a research paper as a co-author 4. Submitting the same paper simultaneously to more than one refereed journal 5. Publishing two papers with a considerable degree of overlap in two refereed journals 6. Publishing two papers with a considerable degree of overlap in a refereed and a professional journal 7. Publishing two papers with a considerable degree of overlap in a research report and a refereed journal 8. Presenting the same research paper at more than one conference 9. The editor of a refereed journal publishing in the journal he/she edits 10. The editor of a refereed journal reviewing a paper himself/herself
88
1·30
62
1·31
Analysis by seniority Overall responses
Non-senior b
Senior
n
Mean
n
Mean
n
26
1·27
84
2·08
58
Non-senior n
Meanb,c
2.05†††
26
2·15†††
†††
25
2·16††
Mean
b
88
1·55
62
1·61∗∗
26
1·38
83
2·10
58
2·07
88
2·44
62
2·48
26
2·35
86
2·52
60
2.43†††∗∗∗
26
2·73†††
88
1·60
62
1·56
26
1·69
83
2·07
59
2·05†††
24
2.13†††
88
2·53
62
2·66∗∗
26
2·23
86
2·53
60
2·63††∗∗
26
2·31†††
87
4·31
62
4·29
25
4·36
85
3·16
60
3·08†††∗
25
3.36†††
87
4·39
62
4·42
25
4·32
85
3·21
60
3·17†††
25
3·32†††
87
3·77
62
3·86
25
3·56
86
3·32
61
3·32†††
25
3.32†††
†††
88
3·27
62
3·32
26
3·15
84
2·79
58
2·69 ∗
26
3·00†††
87
3·25
61
3·31
26
3·12
82
2·51
58
2·53†††
24
2.46†††
Analysis by seniority Overall responses
a
319
Mean calculations are based on the underlying distribution of responses which for ‘acceptability’ ranged from 1=‘always unacceptable’, 2=‘normally unacceptable’, 3= ‘neutral’, 4=‘normally acceptable’ and 5=‘always acceptable’, and for frequency ranged from 1=‘never occurring’; 2=‘occasionally occurring’; 3=‘often occurring’; 4= ‘very often occurring’. b The starred research practices in the ‘senior’ mean column are those where the acceptability or perceived frequency scores of professors/readers/senior lecturers are significantly different (using the Mann–Whitney–Wilcoxon test) from the acceptability or perceived frequency scores of lecturers. The significance level is indicated as follows: ∗ significant at 10% level; ∗∗ significant at 5% level; ∗∗∗ significant at 1% level. c The research practices marked by † in the ‘senior’ and ‘non-senior’ mean columns are those where there is a significant correlation (using Spearman’s rank correlation coefficient) between the acceptability of a practice and its perceived frequency. The significance level is indicated as follows: †† significant at 5% level; ††† significant at 1% level.
320
. .
managers and partner positions within a firm tend to possess lower and more homogeneous levels of ethical reasoning. We also analysed our results by editorial/non-editorial involvement. Major editorial involvement was classified as serving as a journal editor (includes deputy editor/assistant editor/associate editor and/or serving on three or more editorial boards). The editorial partitioning did not typically provide important new information in addition to the seniority partitioning. Only two of the 20 editorial/noneditorial comparisons were statistically significant. Except for the two significant comparisons, we do not consequently discuss the editorial partitioning further. Table 1 shows the mean scores relating to the acceptability of research practices, both overall and partitioned by seniority. In addition, the significance levels for the correlation between acceptability and perceived frequency of occurrence, using Spearman’s rank correlation coefficients, are also reported.6 When evaluating these findings, it is important to bear in mind that the results of attitude measurement studies are necessarily subjective. No objective definition of acceptability is ever achievable given the lack of commonality of attitudes, the contradictory attitudes making up an individual’s responses and the potential for deliberate rationalization and deception (see Madge, 1980, pp. 108–203). The results presented here represent qualitative opinions in a quantitative way. This subjectivity must be appreciated when evaluating the findings. It can be seen from Table 1 that the practice of replicating other researchers’ methodology without giving due credit is generally regarded as unacceptable, with a mean score for the overall sample of 1·30. As many as 98% of the senior respondents and 96% of the non-senior respondents stated that the practice was normally or always unacceptable. The failure to give equal recognition as co-authors to all of the contributors to a paper is also widely believed to be unacceptable, with a mean for the overall sample of 1·55. However, the responses from the two categories of respondent show that this practice is perceived as significantly less acceptable (at the 5% level) by lecturers than by their senior colleagues. We tested for the existence of significant differences using the Mann–Whitney–Wilcoxon two-sample rank test and ANOVA. The Mann–Whitney and ANOVA tests revealed virtually identical results, and we therefore report only the Mann–Whitney results in Table 1.7 The existence of a significant difference for this particular practice is consistent with the idea that the number of publications is an important criterion in the promotion decision. We surmise that the failure to be included as an author is probably perceived more acutely by lecturers eager for publications in order to boost their promotion profiles. More senior staff, while still recognising the issue as important, might be expected to have a slightly more relaxed attitude towards the inclusion decision. Including minimal contributors as co-authors was thought by a majority of both senior and non-senior respondents to be always or normally unacceptable, but 18% of the senior and 11% of the non-senior categories felt
321
this to be normally acceptable. One respondent (professor, new university) suggested several people working on a major research project, but with different levels of contribution being made to different papers, as an example of acceptable practice. Submitting the same paper simultaneously to more than one refereed journal is generally thought to be unacceptable by both sets of respondents, with an overall mean of 1·60. One of the respondents (lecturer, old university) who felt this practice was acceptable explained that this was a reaction to a system which has provided such explicit rewards to individuals and institutions based on research output that this had led to a suspension of ‘ethical’ beliefs and the adoption of an ‘anything goes’ philosophy. Publishing papers with a considerable degree of overlap in different journals or reports was generally agreed to be acceptable if the sources of publication would have different readership profiles. Virtually all of the respondents felt that publishing overlapping material in a refereed and a professional journal or in a refereed journal and a research report was normally or always acceptable. Much less support was found for cases where the two sources of output were both refereed journals, and this practice was perceived as significantly less acceptable (at the 5% level) by lecturers than by senior staff. This can also be seen from the lower mean of 2·23 for nonsenior staff compared with 2·66 for senior staff. Prima facie, this is perhaps surprising, as the keenness of lecturers to increase the number of their publications might be expected to cause them to be relatively more relaxed about this practice. The fact that they are not may indicate that their primary concern is with other (presumably competing) lecturers achieving ‘unfair’ advantage from this practice. The attitude to multiple submissions of the same paper to different refereed journals provides the most striking contrast between the editorial/ non-editorial involvement categories. Those respondents with a major editorial involvement (mean score 1·22) feel this practice is significantly less acceptable at the 1% level than non-editors (mean score 1·70). This differential attitude may reflect the gatekeeping role of the editors. The pressures to publish, particularly for non-editors, clearly leads them to have a more relaxed view of multiple submissions. As the publication environment becomes increasingly competitive and as the penalties for violations of acceptable practice are unclear, differences between the attitudes of these two groups may further widen. The degree to which editors disapprove of this practice is illustrated by the fact that they consider it even more heinous than replicating other researchers’ methodology without credit (mean score 1·28). This is particularly interesting, because, prima facie, there is a violation of a free, competitive market in information. The market is distorted in the ‘buyers’ (journal editors) favour. Journal editors would argue this enhances quality standards, but significant drawbacks are created for suppliers (authors), particularly the time delays caused by sequential rejection. Presenting the same research paper at more than one conference was
322
. .
generally regarded as acceptable by both sets of respondents. The role of the editor in the publication process produced the widest dispersion of responses and also the greatest number of ‘further comments’. On balance, the acceptability of editors publishing in their own journal and reviewing papers themselves, are both supported. However, some concern was expressed over the need for a full double-blind review of papers submitted by editors and for editors not being the sole reviewers of papers submitted by others. Several respondents emphasized that any review of an editor’s submission should be truly double-blind refereed. Unfortunately, as many respondents also feel that anonymity in the refereeing process is difficult to achieve, the implications of such comments are unclear. Some respondents also expressed the opinion that editors not publishing in the journals they edit could be viewed negatively in certain circumstances. One, for instance, commented that: ‘I can see that there is some convention that editors do not publish in their own journal, but this poses problems where an editor holds that post because of special research interests of relevance to the journal. The reviewing process provides some protection here, I think, and it seems odd to devote much effort to promoting a journal and not deigning to appear in it.’ (lecturer, old university)
Other respondents, however, had different views, for example: ‘However, it does not seem quite right and, if something is worth saying an editorial is enough, otherwise, if it is good, it will be published elsewhere.’ (senior lecturer, old university)
While clearly the views expressed are mixed, perhaps the best summary position was expressed in one respondent’s (senior lecturer, old university) belief that most editors do not have massive egos and are able to ‘play fair’. This implicitly acknowledges that decisions of this type are judgmental and rely on a fundamental sense of fairness. In every case, a positive and significant correlation was found between the acceptability of each practice and its perceived frequency of occurrence (see the mean columns in the analysis by seniority of perceived frequency in Table 1). In 18 of the 20 cases, the significance was at the 0·01 level. Acceptable practices, therefore, are perceived to occur more frequently and unacceptable ones less frequently. Two, not mutually-exclusive, explanations are possible. Firstly, respondents may feel that the accounting research community generally behaves in a principled fashion. Secondly, lack of information on the frequency of occurrence may incline respondents to believe that acceptable practices must be frequent and vice versa. Unacceptable actions carry negative consequences and thus negative utility and, therefore, are ‘assumed’ to be infrequent. The converse is true for acceptable research practices.
323
However, despite the correlation between acceptability and perceived frequency, some interesting inconsistencies still emerge.8 For example, the research practices of replicating other researchers’ methodology without giving due credit, failing to give equal recognition as co-authors to all contributors to a research paper and submitting the same paper simultaneously to more than one refereed journal were perceived by over 90% of both sets of respondents to occur at least occasionally. The acceptability analysis, however, revealed that these practices were generally thought to be unacceptable. Similarly, 40% of the senior respondents and 58% of the non-senior respondents believed that the practice of including someone making only a minimal contribution to a paper as a co-author occurred often or very often. This is despite only 18% of the senior and 11% of the non-senior staff finding this normally acceptable. The higher perceived frequency by non-senior staff over their senior colleagues is also significantly different (at the 1% level). A further example is evident in the responses from senior staff to the practice of publishing a paper with a considerable degree of overlap in two refereed journals. Only 17% of the senior respondents felt this to be normally or always acceptable, but as many as 54% felt that the practice occurred often or very often. For non-senior staff, the differences were not so marked with 34% regarding the practice as normally or always acceptable, and 31% feeling that the practice occurred often or very often. In this case, the perceived frequency of occurrence by senior staff is significantly higher (at the 5%) level, than by their non-senior counterparts. The practice of publishing essentially similar papers in outlets with different readership profiles was perceived to occur frequently and this is consistent with the high level of acceptability of this practice. However, it is worth noting that non-senior staff perceive that the practice of publishing overlapping papers in a refereed and professional journal occurs significantly more frequently (at the 10% level) than do their senior counterparts. Consistency between the acceptability of presenting the same research paper at more than one conference and the perceived occurrence of this practice is not quite so high, with over 90% of both sets of respondents indicating that this occurs often or very often, but only 71% of senior staff and 56% of non-senior staff believing this to be normally or always acceptable. As discussed above, the acceptability of editors publishing in, and reviewing papers for, their own journal produced widely disparate responses. All of the respondents believed that editors do publish, at least occasionally, in the journals they edit, with as many as 71% of non-senior staff and 57% of senior staff reporting that this occurs often or very often. Similarly, over 90% of both sets of the respondents believed that editors themselves review papers submitted to their journal at least occasionally. Interestingly, the practice of editors of refereed journals publishing in the journal they themselves edit is perceived by those with major editorial involvement to be a significantly less frequent occurrence than by those without major editorial involvement. The acceptability of editors publishing in the journals they edit is, as was pointed out earlier, a practice that
. .
324
Table 2 Responses to research publication statements Analysis by seniority of respondents Overall responses Mean
Seniora
Non-seniora
n 1 2 3 4 5 Meanb n 1 2 3 4 5 Mean % % % % % % % % % %
Research publication statements
n
1. If the order in which the co-authors to a paper appears is not alphabetical this implies the contribution to the paper has not been equal 2. The policy of a refereed journal should always be to use a minimum of two reviewers 3. There are ‘gatekeepers’ to most areas of research who tend to make it difficult for newcomers to enter 4. If a reviewer knows the identity of an author he/she should not review the paper 5. If a reviewer knows the identity of an author he/she should review the paper but notify the editor that the identity of the author(s) is known
88 4·16
62 3 10 5 45 37 4·03∗∗ 26
88 3·98
62
86 3·44
61 3 17 30 40 10 3·37
25 4 16 20 40 20 3·56
84 2·77
59 13 41 22 17 7 2·63∗
25 4 36 20 24 16 3·12
82 3·90
58
24 8 4 8 59 21 3·79
8 20 40 32 3·96
10 16 43 31 3·95
26
8 38 54 4·46
4 26 35 35 4·00
a The columns headed: 1=relate to the percentage of respondents strongly disagreeing with each statement, 2=disagreeing, 3=neutral, 4=agreeing and 5=strongly agreeing. b The starred items in the ‘senior’ mean column are those where the mean level of agreement recorded by professors/readers/senior lecturers are significantly different (using the Mann–Whitney–Wilcoxon test) from the mean level of agreement recorded by lecturers. The significance level is indicated as follows: ∗ significant at 10% level; ∗∗ significant at 5% level.
attracted a number of ‘further comments’ and is clearly an issue of some concern. It is therefore interesting that non-editors feel that this practice occurs more frequently than those with major editorial involvement and perhaps indicates a suspicion by non-editors that editors might use their position to obtain some advantage in the publication process. Responses to research publication issues The extent to which the respondents agreed or disagreed with a number of research publication statements is shown in Table 2. The mean scores of
325
the overall sample and a more detailed analysis of the responses from senior and non-senior staff is provided. The response scale ranged from 1 (strongly disagree), through 2 (disagree), 3 (neutral), 4 (agree) to 5 (strongly agree). Respondents generally agreed or strongly agreed that non-alphabetical ordering of the authors to a paper implies unequal contribution (overall mean 4·16). Indeed, only 13% of the senior respondents and none of the non-senior respondents disagreed or strongly disagreed with this. However, the level of agreement from non-senior respondents was significantly higher (at the 5% level) than from their senior counterparts. This may reflect a greater sensitivity to this issue by those still aspiring for promotion. The practice of alternating the ordering of the names in the interests of equity (used by some authors who regularly work together and contribute equally to the papers they produce) may, therefore, send the wrong signals to the academic community. A footnote clarifying the order of names is one possible solution (Von Glinow & Novelli, 1982). The use of a minimum of two referees by refereed journals is strongly supported (overall mean 3·98). A majority of the non-senior respondents (60%) and 50% of the senior respondents agreed that gatekeepers to most areas of research make it difficult for newcomers to enter the market. The ‘gatekeepers’ statement attracted the greatest number of supplementary written comments. Considerable disagreement about both the existence and role of gatekeepers was evident. The extracts below, for example, give the flavour of those who disputed the existence of a gatekeeper effect: ‘Difficult to know what goes on in areas of research other than one’s own. Certainly I am unaware of a “gatekeeper” effect in my research area.’ (professor, new university)
and ‘. . . it is not the case but newcomers perceive it this way which is especially bad.’ (senior lecturer, old university)
At the other extreme, some respondents clearly perceived the existence of the gatekeeper phenomenon: ‘I am sure there are gatekeepers to some, and it is the job of the editor to know them and take account of their bias.’ (professor with editorial involvement, old university)
and ‘. . . there are gatekeepers in research areas and there are ways that a newcomer can make entry easier . . . for example, by talking to potential “gatekeepers” at conferences, writing to them, collaborating with them etc.’ (lecturer, old university)
The idea of a gatekeeper as the guardian of a particular research paradigm, or as a barrier to obtaining funding aroused particular concern. One respondent distinguished between the ‘legitimate’ role of quality control
326
. .
and the ‘illegitimate’ role of protecting or promoting particular research paradigms. Another pointed out that: ‘The more important issue is that there are research “paradigms” that develop that make new or unusual lines of enquiry more difficult to publish.’ (professor with editorial involvement, old university)
Obviously, the issue of doctrinal paradigm control is an emotive one, and drawing conclusions from a limited sample is potentially dangerous. However, the perception of negative gatekeeping may have a constraining effect on the development of new research paradigms and may discourage newcomers to a research area. The difficulty of striking the correct balance is well summarized by one respondent: ‘I suspect that some academics with an established research interest have a tendency to be tough reviewers of new material in their area, but I guess that new researchers are also going up a learning curve and may wrongly attribute the learning cost with lack of generosity on the part of reviewers. It is clear that the review process is very subjective and uneven.’ (lecturer with editorial involvement, old university)
The interesting question of the role of editor as gatekeeper was raised by one respondent (professor with editorial involvement, old university), who felt strongly about the ‘Editor overriding expert reviewers from his/her own interests or editorial bias’. This begs the question of what is the relationship between reviewers and the editors? In short, when is it legitimate for the editor to accept or reject the views of the reviewers? The lack of uniformity of views suggests that some, but not all, areas of research are perceived to be guarded by gatekeepers. The identification of areas perceived as strongly protected by gatekeepers would be interesting but the current research did not address this issue. Concern was also expressed about the role of special issues (and by extrapolation editorial commissioning). ‘Fast track publication (e.g. papers from a conference being included in a “special” issue is inappropriate and unfair. All papers published should be subject to the same refereeing process, particularly now that publications count so much!’ (senior lecturer, old university)
The statement that a reviewer who knows the identity of the author(s) to a paper should not review the paper produced the widest range of responses and a significantly higher (at the 10% level) degree of agreement from nonsenior respondents. This is intuitively logical as the lack of anonymity may potentially have a more serious impact on non-senior respondents, given the likelihood that many of the gatekeepers will be senior colleagues. As many as 80% of the non-senior respondents and 74% of the senior respondents agreed or strongly agreed with the related statement that where the identity of the author(s) is known, the review should be undertaken, but the editor notified. These two statements also attracted a large number
327
of ‘further comments’. Several respondents, for instance, drew attention to the need for care in allowing suspected knowledge of the authorship of a paper to affect the outcomes of a review. One for instance, felt that: ‘It is common to have read the paper in an early version. One should, however, be careful if the relationship with the author is a close one e.g., a friend or departmental colleague.’ (senior lecturer, old university)
Another pointed out that the ‘crucial point is whether the knowledge affects the assessment. I think referees will often have a suspicion who the author is, but not be sure.’ (professor with editorial involvement, old university)
Many of the comments also drew attention to the fact that conference presentations make anonymity difficult to achieve and that the solution of undertaking the review, but notifying the editor that the identify of the author(s) was known, was the only practical way to proceed.
CONCLUSION AND RECOMMENDATIONS Overall, our survey reveals that 88 UK accounting academics perceive substantial differences in the acceptability of ten research practices. Three research practices are found to be particularly unacceptable: replicating other researchers’ methodology without giving due credit; failing to give equal recognition to all colleagues contributing to a research paper and simultaneously submitting the same paper to more than one research journal. These three practices were thought to occur, at least occasionally, by more than 94% of respondents. Two research practices (including someone who has made only a minimal contribution as a co-author and publishing two papers with a considerable degree of overlap in two refereed journals) were found to be, on average, marginally unacceptable. Generally, neither seniority nor editorial involvement markedly affected respondents’ perceptions. However, those with editorial involvement (as compared to those with no editorial involvement) found submissions simultaneously to more than one refereed journal particularly heinous, while non-senior staff found failing to give equal recognition to all colleagues by including them as authors and publishing two papers with considerable overlap in refereed journals as being significantly more unacceptable than their senior colleagues. From this study of research practices, a set of guidelines can be distilled which refine best practice. The first five practices have particular importance for authors, practices (6) and (7) concern editorial policy, and practices (8) and (9) have particular relevance for reviewers.
328
. .
Recommended Best Practice Guidelines Authors (1) Other researchers’ methodology should always be given due credit. (2) All colleagues, who have contributed to papers, and only those colleagues, should be included in research papers as co-authors. (3) Presenting two papers which are the same at more than one conference is acceptable. (4) Publishing two papers which are the same, or have considerable overlap is acceptable where there are different audiences (e.g. professional and academic), but not acceptable when the audiences are the same (e.g. two refereed accounting journals). (5) The non-alphabetical listing of co-authors to a paper should only be used to signal unequal contributions. Editors (6) The practice of editors publishing in their own journals or reviewing papers themselves is acceptable, however, safeguards should be instituted to ensure that quality thresholds are maintained. (7) The policy of refereed journals should be to use a minimum of two reviewers. Reviewers (8) The role of reviewers is to act as guardians of research quality rather than as protectors of particular research paradigms. (9) If a reviewer knows the identity of an author, the reviewer should review the paper, but notify the editor of his/her recognition.
N 1. Grundy and Holt (1994) supply a review of cognate, mostly US, ethics literature. 2. This is true of non-accounting disciplines and in countries other than the UK. For example, the New Scientist’s editorial (9th November, 1996) highlights the problem of fraud in a US genetics laboratory. The editorial (p.3) states ‘what might help is for someone somewhere to curb the growing pressure on researchers to squeeze as many pages as they can get out of each project.’ The editorial continues ‘this emphasis on quantity rather than quality persists . . . leaving experienced scientists too busy keeping their own publications rolling off the press to spend much time checking the work of others.’ 3. Loeb (1990), for example, briefly discusses issues such as inattention to teaching and students; alcoholism and drug abuse; the need to teach accounting ethics; as well as more specific research issues such as research errors, fraud and plagiarism. Keys & Hendricks (1984) and Loeb (1990, 1994) advocate the need for a code of ethics for academic accountants, but do not formally conduct any empirical research. Loeb (1988) discusses some crucial issues when teaching students accounting ethics. 4. This study does not specifically cover the ethics of fraudulent research activity. Davis & Ketz (1991), after examining the potential for accounting research fraud, conclude that
5. 6. 7.
8.
329
accounting research has escaped many of the accusations and demonstrations of scientific fraud, offering the comforting but unproven implication of a relative lack of fraudulent behaviour in accounting research. Interested readers are referred to Brinn et al. (1996) for a discussion of the attitudes of 88 UK accounting academics to 44 leading research journals. We are grateful to an anonymous referee for pointing out the potential applicability of this approach. The Mann–Whitney–Wilcoxon and ANOVA tests were also carried out on a collapsed version of the data set. The collapsed data set was reduced from the original five-point scale to a three-point scale (representing unacceptable, neutral and acceptable). The three-point scale produced similar results to the original five-point scale, and is therefore not discussed further. These associations vary between 0·43 and 1·0. Rebasing the acceptability measure to a four point scale produced no substantial differences in the correlation levels.
R Benjamin, J.J. & Brenner, V.C. (1974). ‘Perceptions of Journal Quality’, The Accounting Review, April, pp. 360–362. Brinn, T., Jones, M.J. & Pendlebury, M. (1996). ‘UK Accountants’ Perceptions of Research Journal Quality’, Accounting and Business Research, Summer, pp. 265–278. Brown, L.D. & Huefner, R.S. (1994). ‘The Familiarity with the Perceived Quality of Accounting Journals: Views of Senior Accounting Faculty in Leading US MBA Programs’, Contemporary Accounting Research, Vol. 11, No. 1, pp. 223–250. Davis, S.W. & Ketz, J.E. (1991). ‘Fraud and Accounting Research’, Accounting Horizons, September, pp. 106–109. Doyle, J.R. & Arthurs, A.J. (1995). ‘Judging the Quality of Research in Business School: The UK as a Case Study’, Omega, Vol. 23, No. 3, pp. 257–270. Engle, T.J. & Smith, J.L. (1990). ‘The Ethical Standards of Accounting Academics’, Issues in Accounting Education, Vol. 5, No. 1, Spring, pp. 7–30. Engle, T.J. & Smith, J.L. (1992). ‘Accounting Faculty Involvement with Activities of Ethical Concern’, The Accounting Educators’ Journal, Spring, pp. 1–21. Gee, K.P. & Gray, R.H. (1989). ‘Consistency and Stability of UK Academic Publication Output Criteria in Accounting’, British Accounting Review, Vol. 2, No. 1, pp. 43–54. Gray, R.H. & Helliar, C. (1994). ‘UK Accounting Academics and Publication: An Exploration of Observable Variables Associated with Publication Output’, British Accounting Review, Vol. 26, No. 3, pp. 235–254. Gray, R.H., Haslam, J. & Prodham, B. (1987). ‘Academic Departments of Accounting in the UK: A Note on Publication Output’, British Accounting Review, Vol. 19, No. 10, pp. 53–73. Grundy, J. & Holt, G.J. (1994). ‘Ethical Judgements in Accounting: Personal and Professional’, paper presented at the 17th Annual Congress of the European Accounting Association, Venice. Howard, T.P. & Nikolai, L.A. (1983). ‘Attitude Measurement and Perceptions of Accounting Faculty Publication Outlets’, The Accounting Review, Vol. LVIII, No. 4, pp. 765–776. Hull, R.P. & Wright, F.B. (1990). ‘Faculty Perceptions of Journal Quality: An Update’, Accounting Horizons, March 1990, pp. 77–98. Humphrey, C., Moizer, P. & Owen, D. (1995). ‘Questioning the Value of the Research Selectivity Process in British University Accounting’, Accounting, Auditing and Accountability Journal, Vol. 5, No. 3, pp. 141–164. Hutchinson, C. (1989). ‘Consistency and Stability of UK Academic Publication Output
330
. .
Criteria in Accounting: A Comment’, British Accounting Review, Vol. 23, No. 3, pp. 279–284. Kerr, S., Tolliver, J. & Petree, D. (1977). ‘Manuscript Characteristics which Influence Acceptance for Management and Social Science Journals’, Academy of Management Journal, Vol. 20, No. 1, pp. 132–141. Keys, D.E. & Hendricks, J.A. (1984). ‘The Ethics of Accounting Research’, Journal of Accounting Education, Vol. 2, No. 2, pp. 77–88. Loeb, S.E. (1988). ‘Teaching Students Accounting Ethics: Some Crucial Issues’, Issues in Accounting Education, Vol. 3, No. 2, Fall, pp. 316–329. Loeb, S.E. (1990). ‘A Code of Ethics for Academic Accountants?’, Issues in Accounting Education, Vol. 5, No. 1, Spring, pp. 123–128. Loeb, S.E. (1994). ‘Accounting Academic Ethics: A Code is Needed’, Issues in Accounting Education, Spring, pp. 191–200. Lyall, D. (1978). ‘UK University Accounting Departments: Journal Output Performance 1972–1976’, British Accounting Review, Vol. 10, No. 1, pp. 22–25. Mace, J. (1993). ‘University Funding Changes and University Efficiency’, Higher Education Review, Vol. 25, No. 2, pp. 7–22. Madge, J. (1980). The Tools of Social Science, London: Longman. New Scientist (1996). ‘Cooking the books’, editorial, 9th November, p. 3. Nobes, C.W. (1985). ‘International Variation in Perceptions of Accounting Journals’, Vol. LX, No. 4, The Accounting Review, October, pp. 702–705. Nobes, C.W. (1986). ‘Academic Perceptions of Accounting Journals in the English Speaking World’, British Accounting Review, Vol. 18, No. 1, Spring, pp. 7–16. Peasnell, K.V. & Williams, D.J. (1986). ‘Ersatz Academics and Scholar Saints: The Supply of Financial Accounting Research’, Abacus, Vol. 22, No. 2, pp. 121–135. Ponemon, L.A. (1990). ‘Ethical Judgements in Accounting: A Cognitive Developmental Model and Experimental Evidence’, Critical Perspectives in Accounting, pp. 227–251. Ponemon, L.A. (1992). ‘Ethical Reasoning and Selection—Socialisation in Accounting’, Accounting, Organizations and Society, Vol. 17, No. 3/4, pp. 239–258. Puxty, A.G. & Tinker, T. (1995). ‘Policing Accounting: The Sociology of Knowledge as Praxis’, in T. Tinker & A.G. Puxty (eds) Policing Accounting Knowledge: The Market for Excuses Affair, (Princeton: Markus Wiener and London: Paul Chapman). Puxty, A.G., Sikka, P. & Willmott, H.C. (1994). ‘Systems of Surveillance and the Silencing of UK Academic Accounting Labour’, British Accounting Review, Vol. 26, pp. 137-–171. Sherrel, D.L., Hair, J.F. Jr., Griffin, M. (1989). ‘Marketing Academicians’ Perceptions of Ethical Research and Publishing Behavior’, Journal of the Academy of Marketing Science, Fall, pp. 315–324. Taylor, J. (1994). ‘Measuring Research Performance in Business and Management Studies in the United Kingdom: The 1992 Research Assessment Exercise’, British Journal of Management, Vol. 5, pp. 275–288. Von Glinow, M.A. & Novelli, L. Jr (1982). ‘Ethical Standards within Organizational Behavior’, Academy of Management Journal, Vol. 25, No. 2, pp. 417–436. Whittington, G. (1993). ‘The 1992 Research Assessment Exercise’, The British Accounting Review, Vol. 25, No. 4, December, pp. 383–395.