Research Policy xxx (xxxx) xxx–xxx

Contents lists available at ScienceDirect

Research Policy
journal homepage: www.elsevier.com/locate/respol

Towards a taxonomy of research misconduct: The case of business school

research
⁎
Jeremy Halla, Ben R. Martinb,
a
Surrey Business School, University of Surrey, Guildford, Surrey, GU2 7XH, UK
b
SPRU (Science Policy Research Unit), University of Sussex, Brighton, BN1 9SL, UK; also at Centre for Science and Policy (CSaP) and Centre for Business Research
(CBR), Judge Business School, University of Cambridge, UK

A R T I C LE I N FO A B S T R A C T

Keywords: This paper examines the growing pressures and incentives encouraging research misconduct, along with the
Research misconduct consequences, as illustrated by the case of business school research. Drawing on a review of the literature on
Taxonomy different theoretical approaches to analysing organizational misconduct, we develop a formal taxonomy dis-
Inappropriate conduct tinguishing appropriate conduct from blatantly inappropriate misconduct but with a specific focus on the ‘grey’
Questionable conduct
areas between these extremes in the form of questionable and inappropriate behaviour. We identify various
Competitive pressures
sources of research misbehaviour and different categories of those affected. The aim is to provide a clearer
Business schools
understanding of what research behaviour is deemed appropriate or not, which stakeholders it affects, and the
pressures and incentives likely to exacerbate such misconduct. We conclude with a discussion of how the tax-
onomy can help shape future good research practice (thereby setting a better example to students), and offer
some propositions for future research.

1. Introduction questionable: “how scientists perceive the line separating ethical from
unethical behaviour is likely to exhibit a much more ambiguous char-
Universities, like most institutions, are subject to growing perfor- acter than existing research acknowledges” (Johnson and Ecklund,
mance pressures. Competition for tenure, research funding, publication 2016, p. 990), although some form of misrepresentation, inaccuracy or
‘slots’ in elite journals, employment in leading universities, and re- bias is generally involved (Steneck, 2006).
putation within the academic community are major drivers of in- We address this ambiguity by proposing a taxonomy of various
dividual or institutional success (Martin, 2016). However, a pre- forms of research behaviour, ranging from appropriate practice to
occupation with publishing in elite journals and counting citations may blatant misconduct, specifically focusing on behaviour between these
create perverse incentives (Bouter, 2015), encouraging efforts to ‘push’ extremes that is ‘inappropriate’ or ‘questionable’. Our taxonomy dif-
or even transgress the boundaries of appropriate behaviour. Indeed, the ferentiates between these categories based on the stakeholders affected
number of articles retracted by journals has risen more than ten-fold in by the misconduct as well as the severity, ranging from premeditated
just ten years (Van Noorden, 2011). Such ‘system gaming’ may enhance dishonesty and intentional rule-bending to less intentional poor beha-
individuals’ chances in the competition for posts and funding, and the viour that may arise due to complexity, sloppiness, ignorance or honest
academic standing of the organization. Where some boundary has error. We provide examples drawn from business school research,
clearly been transgressed, this is regarded as ‘misconduct’. Yet as Butler where competitive pressures seem particularly acute. Our aim is to
et al. (2017) note, blatant misconduct such as fabrication, falsification provide clearer and more consistent guidelines for researchers, espe-
and plagiarism constitutes only part of a much wider problem. Such cially junior scholars, as well as for journal editors and others re-
blatant misconduct can be addressed, for example, through legal me- sponsible for monitoring and preventing academic misconduct.
chanisms, whereas the “less flagrant, more subtle cases of potential To develop our taxonomy1 of research misconduct, we draw on the
misconduct”, or what Fanelli (2009) and John et al. (2012) call ‘ques- organizational misconduct literature, which Greve et al. (2010) cate-
tionable research practices’, remain poorly understood. Moreover, there gorize as being derived primarily from rational choice, strain, cultural,
remains ambiguity about what types of research practices are network, and bounded rationality theories of misconduct. We provide

⁎
Corresponding author.
E-mail addresses: jeremy.hall@surrey.ac.uk (J. Hall), B.Martin@sussex.ac.uk (B.R. Martin).
Some distinguish between a ‘taxonomy’ and a ‘typology’, with the former being derived largely from data, and the latter based on categories derived from theory. The framework
1

proposed here involves elements of both.

https://doi.org/10.1016/j.respol.2018.03.006
Received 31 December 2016; Received in revised form 7 March 2018; Accepted 8 March 2018
0048-7333/ © 2018 Elsevier B.V. All rights reserved.

Please cite this article as: Hall, J., Research Policy (2018), https://doi.org/10.1016/j.respol.2018.03.006
J. Hall, B.R. Martin Research Policy xxx (xxxx) xxx–xxx

illustrative examples (Siggelkow, 2007) from business school research, context of a particular group, and hence its interpretation may vary
a field that, like others, has encountered growing problems with re- across groups and over time (see also Vaughan, 1999).
search misconduct (Honig and Bedi, 2012). This is not to imply that According to Ashforth et al. (2008), research on misconduct needs
business school research is more prone to misconduct, an issue on to go beyond static individual traits and behaviour to include the role of
which systematic data is lacking.2 However, as we shall see, competi- processes and systems, as well as how individual, organizational and
tive pressures on business schools and their faculty are at least as severe industrial levels interact to foster misconduct. Shadnam and Lawrence
as in many other areas, while the rewards for publishing in elite jour- (2011, p. 381) suggest that widespread misconduct, what they term
nals are often more lucrative. Furthermore, business school research ‘moral collapse’, results from the “breakdown of connections between
has become so theoretically sophisticated and methodologically com- moral communities, organizations and individuals which may be
plex that its direct impact on business practice is often rather limited avoidable if actors work to establish or restore those connections”. Note
(Bennis and O’Toole, 2005; Banks et al., 2016). Consequently, the that while there are numerous laws banning specific types of organi-
temptation to engage in questionable behaviour may be high, since zational misconduct, many types of behaviour are not illegal or are
academics may believe that the rewards outweigh the risks of detection, impractical to control through legal proceedings (Krawiec, 2005;
or that there is no direct impact on business, i.e. the ‘no harm, no foul’ Barnett, 2014), including the ambiguous ‘questionable research prac-
convention applies, especially for ambiguous areas of misconduct. tices’ (Fanelli, 2009; John et al., 2012; Johnson and Ecklund, 2016)
However, such a narrow perspective ignores the wider influence that discussed here. In what follows, the unit of analysis includes both in-
business schools have on promulgating ethical decision-making in the dividual and organizational-level misconduct – and the processes and
business community (Floyd et al., 2013). Following Eisenhardt et al. systems affecting these – with a particular focus on the relatively ne-
(2016), we chose business school research because it provides a useful glected non-legal ‘grey’ areas.
case for understanding ‘grey’ areas of misconduct by capturing certain In their literature review, Greve et al. (2010) note that organiza-
key aspects, specifically why it occurs, whom it affects and why it tional misconduct has been explored using numerous theoretical lenses,
matters to management education. which they categorize into five areas: rational choice, strain, cultural,
The paper is structured as follows. Section 2 considers the theore- network and accidental. Their categorization is primarily drawn from
tical foundations of organizational misconduct, setting out different organizational theory and related concepts common to business
theoretical frameworks for analysing and understanding the forms of schools, and hence makes assumptions that may not be shared by other
misconduct exhibited by academics and their organizations. Section 3 research traditions3 (for example, those in law, psychology or so-
explains why we focus on business school research, while Section 4 ciology).4 Nevertheless, it offers a reasonably comprehensive starting
introduces the entrepreneurial risk-return perspective on business point to develop a taxonomy of research misconduct.
school research. These more theoretical sections provide the founda- The rational choice perspective includes agency, contract, and re-
tions for developing a taxonomy of the various forms of research mis- putation theories that address the social control of individually rational
conduct and other questionable behaviour (Section 5), the different actors. Drawing on Arrow (1963), Greve et al. (2010) argue that these
sources of such misbehaviour (Section 6) and the main stakeholders theories have focused on inappropriate risk taking, accounts manip-
affected (Section 7). Finally, Section 8 summarizes the main conclu- ulation and blatant fraud, where “rational-action modeling assumes
sions and implications, including presenting propositions for further self-interested actors who need to be controlled in order not to choose
research. actions that would be beneficial for them but harmful for transaction
partners or third parties” (Greve et al., 2010, p. 60). The rational choice
perspective aligns closely with cost-benefit analysis as used in legal
2. Theoretical foundations of organizational misconduct theories of misconduct. Drawing upon Becker’s (1968) economic model
of crime and punishment, Hornuf and Haas (2014) note that an in-
Misconduct does not happen in a vacuum but generally emerges dividual’s propensity to commit a crime is a function of the probability
from organizational or institutional pressures and incentives − what of conviction and the punishment that may follow compared with the
has been termed ‘organizational misconduct’. According to Vaughan utility gained from committing that crime. According to Krawiec
(1999), with the rise of formal organizations have often come mistakes, (2005), legal-based systems that enforce organizations’ duties through
misconduct and disasters, in other words the ‘dark side of organiza- compliance structures are likely to fail because courts lack sufficient
tions’. She argues that organizational misconduct occurs when in- information about the effectiveness of such structures. She argues that
dividuals or groups violate internal or external rules, when attempts to compliance-based liability systems tend to encourage “cosmetic in-
engage in or encourage one type of behaviour unintentionally result in ternal compliance structures that reduce legal liability without reducing
another, and by accident. the incidence of organizational misconduct” (Krawiec, 2005, p. 572). In
Greve et al. (2010, p. 56) define organizational misconduct as short, legal theories of misconduct may fail to deter organizational
“behaviour in or by an organization that a social-control agent judges to misconduct, generating costly but largely ineffective compliance
transgress a line separating right from wrong; where such a line can structures.
separate legal, ethical, and socially responsible behaviour from their A more recent application of the rational choice perspective is
antitheses”. A social-control agent is an actor representing a collective tournament theory, a game-theoretic view of principal-agent relations5
that can impose sanctions on behalf of that collective. Organizations used to understand the causes of corporate misconduct, in particular
may be judged by multiple agents, such as international, national and those stemming from highly competitive promotion processes in orga-
local governmental entities, as well as professional associations like the nizations (Shi et al., 2016). Like agency and game theory, tournament
American Medical Association, the American Bar Association, uni- theory is based on information asymmetry between principals and
versity ethics boards and in our case the Committee of Publishing Ethics
(COPE). Greve et al. (2010) emphasize that their definition “avoids
treating misconduct as a straightforward implication of a set of laws, 3
Shortage of space preclude us from going further into the limitations here – for a
ethical principles, and/or social norms” (Greve et al., 2010, p. 56). recent summary of the limitations of theory and research on organizational misconduct,
Misconduct is thus essentially a social construct defined within the see Palmer et al. (2016).
4
For examples of legal, psychological and sociological approaches to aspects of gaming
and misconduct, see Adams and Pimple (2005), Umphress et al. (2010), and Espeland and
2
In his meta-review of the literature on research misconduct, the only field difference Sauder (2007) respectively.
that Fanelli (2009, p. 1) found was that “misconduct was reported more frequently by 5
For a summary of principal-agent models of organizational misconduct and their
medical/pharmacological researchers than others”. limitations, see Krawiec (2005).

2
J. Hall, B.R. Martin Research Policy xxx (xxxx) xxx–xxx

agents, and specifically the extent to which worker output can be networks over time. Network approaches do, however, address
monitored and how incentives can be used to align individual beha- Ashforth et al.’s (2008) call to go beyond individual behaviour to in-
viour with organizational goals. However, the rational choice per- clude the role of processes and systems embedded within a system, as
spective, and in particular the influence of economic applications like well as how individual, organizational and industrial levels interact to
agency theory, has been criticised for damaging the ethical behaviour result in misconduct, and how all this may be shaped by social-control
of practitioners (Ghoshal, 2005; Kidwell and Kidwell, 2008). agents. Thus, network approaches link micro individual and group le-
Greve et al.’s (2010) second area of organizational misconduct vels of analyses of misconduct (what Ashforth et al. (2008) call ‘bad
centres on strain theory, which posits that actors resort to misconduct apples’) with more macro organization, industry and national levels
when they fail to achieve their goals (or those imposed on them) (i.e. ‘bad barrels’).
through legitimate means. Originally formulated by Merton (1938) to Greve et al.’s (2010) final theoretical framework for analysing or-
understand why the impoverished were more likely to engage in illegal ganizational misconduct addresses the link between accidents and
activities, strain theory has been used to explore how misalignment misconduct, and is based on a recognition that organizations are com-
between goals and actual achievements may result in misconduct at the plex while managers are limited in what they know (March and Simon,
individual, organizational and societal level (Greve et al., 2010). A 1958; Simon, 1969). Because of such bounded rationality, accidental
related concept, general strain theory (Agnew, 1992), has been used by misconduct may be inevitable (Vaughan, 1999). As we discuss below,
Lewellyn et al. (2017) to explore the phenomenon of conference paper complexity and bounded rationality are undoubtedly causes and
‘double dipping’. They suggest that stressful circumstances create ne- sometimes even reasonable excuses to justify research misconduct,
gative emotions, which may be alleviated by engaging in adaptive be- given the shift towards increasingly complicated methodological ap-
havioural responses, including questionable conduct. proaches, greater emphasis on collaborations with specialized ex-
The third approach examined by Greve et al. (2010) involves cul- pertise, and demands for greater outputs. They are therefore important
tural theories of misconduct, again applied at various levels of analysis factors in the rational choice perspective as well as in the strain, cul-
including occupations, professions, organizations, industries and so- tural and network theories of organizational misconduct.
cieties. Organizational cultures incorporate certain norms, values and
beliefs about which attitudes and behaviours are deemed appropriate
3. Business school research
and good. According to Monteduro et al. (2016), social norms can ex-
plain some of the causal factors of misconduct, which in turn are in-
We focus here on business school research for several reasons. First,
fluenced by level of education, social development and national char-
management education is fast growing and increasingly international
acteristics.
(Honig et al., 2014; Morgeson and Nahrgang, 2008). According to
In contrast to the rational choice perspective, Greve et al. (2010)
Universities UK International (Stern, 2017), UK business schools are by
observe that research on cultural causes of corruption often focuses on
far the largest recipients of foreign students, more than double the next
ethics rather than the procedural rules, threats or coercion. They fur-
discipline (Engineering and Technology), offering lucrative growth
ther note that organizational cultures often implicitly encourage while
opportunities to many universities. They also have low operating costs
simultaneously condemning misconduct, for example, motivating
compared to disciplines that require expensive laboratories and
members to achieve particular ends without providing guidance on how
equipment. At the same time, there is a high demand for business
these should be achieved, or by exhibiting a certain tolerance of rule-
professors who have published in leading journals. Indeed, according to
breaking and undue risk-taking in the guise of innovativeness. This can
a recent survey, seven of the top ten highest paid professors in the US
give rise to what Umphress et al. (2010) call ‘unethical pro-organiza-
are affiliated with business schools.6 Business schools thus illustrate the
tional behaviour’, where individuals attempt to justify their misconduct
highly competitive promotion processes in organizations that Shi et al.
on the grounds that it helps their organisation. Such a culture may be
(2016) suggest may encourage misconduct.
exacerbated when there are pressures and rewards for extraordinary
Second, business schools and their faculty are subject to intense
performance, and it is particularly influenced by the organization’s
competitive pressures from numerous accreditation schemes (e.g.
leadership (Schein, 1983, 1985). According to Sims and Brinkmann
AASCB, EQUIS, AMBA), as well as several influential ranking schemes
(2003), deep cultural flaws driven by leadership and veiled by ‘window-
such as those published by the Economist, Financial Times, US News &
dressing ethics’ resulted in Enron’s collapse: “The culture at Enron
World Report, Wall Street Journal, Forbes and Bloomberg Business Week.
eroded little by little, by the trespassing of ethical boundaries, allowing
Efforts to improve their ranking (and the indicators on which such
more and more questionable behaviour to slip through the cracks”
rankings are based) frequently dominate a business school’s strategy in
(Sims and Brinkmann, 2003, p. 252).
their attempts to attract the best students, staff and funding.
A culture of misconduct, however, does not explain variation in
Third, while there is an growing range of publication outlets for
misbehaviour across organizational participants. This is the focus of
business school research, there are also strong career incentives and
Greve et al.’s (2010) fourth approach, network theories of organiza-
pressures to publish in a narrow range of leading journals, often de-
tional misconduct, which, as they note, “occupy an intermediate posi-
termined by impact factors or elite rankings such as the Financial Times
tion between individual-level theories and the organizational-level
‘FT50’ list or the UT Dallas list of 24 leading management journals.
theories” (p. 68). This perspective focuses on misconduct among in-
Many universities link tenure and promotion directly to publications in
dividuals linked by strong social ties but often rather isolated from
such elite journals, most of which have a rejection rate of over 90%.
other parts of the network, and on intentional collective efforts to de-
Morgeson and Nahrgang (2008) discuss how elite journal lists have
ceive, such as price-fixing. According to Breit and Forsberg (2016),
become extremely influential metrics for judging research performance,
networks between researchers and stakeholders are increasingly
which along with other indicators are then used to rank the institutional
common in contemporary academia, and may result in misconduct
quality and standing of business schools,7 frequently with detrimental
when, for example, actors are exposed to divergent expectations and
pressures from different sources and organizational cultures. They note
that “networks may influence people by providing information of 6
http://www.thebestschools.org/blog/2013/11/25/10-highest-paid-college-
practices (e.g., of how to do things or get away with things)” (Breit and professors-u-s/ (Accessed on 13 November 2017).
7
See the second case study described in Annex A, in which a journal article analysing
Forsberg, 2016, p. 8).
the relative publication performance of one business school reputedly played a significant
Grzesiuk (2016) suggests that research on network approaches to part in obtaining a major endowment. The methodology was later questioned, and when
misconduct provides inconsistent results and tends to be based on static it was revealed that the authors failed to declare they had worked at that school and that
analyses rather than a dynamic approach that explores changes to its director had helped draft the paper, the journal published an ‘expression of concern’

3
J. Hall, B.R. Martin Research Policy xxx (xxxx) xxx–xxx

effects (Mingers and Willmott, 2013). As Langford et al. (2006) stress, 4. The entrepreneurial risk-return perspective for business school
research proxies are often oversimplified, failing to capture important research
paths of knowledge flow; hence, a preoccupation with readily available
proxies can result in counterproductive activities, where the proxies Adopting an entrepreneurial12 risk-return perspective offers inter-
become the de facto goals of institutions or individuals (see also esting insights particularly relevant to business schools. Given that
Zimmerman, 2001; Gioia and Corley, 2002; Lawrence, 2008; Adler and analysing risks and returns is a basic prerequisite of business, it is little
Harzing, 2009). As Kerr (1975, p. 779) warned over 40 years ago, a surprise that some business professors might ‘practice what they
“fascination with an ‘objective’ criterion” may well result in goal dis- preach’ in terms of identifying and pursuing proxy performance mea-
placement, i.e. rewarding one specific behaviour while hoping for an- sures likely to result in personal and institutional success. Here, we are
other. not criticizing such entrepreneurial efforts by individual academics, nor
Yet such caution regarding proxies has seemingly been ignored in the systems that encourage such behaviour. Instead, our focus is on how
many business schools. For example, Honig et al. (2014, p. 120) state: such an orientation may occasionally result in undesirable outcomes,
“Collectively, this heightened level of competition, preoccupation with specifically in two areas. First, a myopic focus on proxies tends to
rankings, and rising research expectations, has resulted in significantly distance unethical behaviour from direct harm, thus incentivizing rule-
increased pressure worldwide on faculty to publish in ‘top-tier’ jour- bending because the risks of direct damage may seem remote. Second,
nals”. They further note that intense competition among business there is the indirect but very important impact on business education,
schools has placed a premium on elite journal publications, which in where, consistent with cultural theories of misconduct, unethical re-
turn is pressuring scholars to “see their work from an entrepreneurial search behaviour tends to erode a culture of integrity in the classroom,
perspective, driving them to consider the risk/return profile of the work which in turn may legitimize unethical business behaviour in practice
they may wish to undertake” (Honig et al., 2014, p.124). (Floyd et al., 2013). As Tang and Liu (2012) show, a supervisor’s per-
More comprehensive journal lists are also used by specific countries sonal integrity and character have significant effects on business stu-
to assess academic performance. These include the Australian Business dents’ subsequent propensity to engage in or refrain from unethical
Deans Council (ABDC) Journal Quality List, the German Academic behaviour.
Association for Business Research journal list (VHB-JOURQUAL), and
the UK-based Association of Business Schools (ABS) Academic Journal 4.1. No harm, no foul
Guide.8 In some cases, university funding is directly linked to pub-
lications in leading journals. For example, the allocation of research Fraudulent business research might seem to have rather limited
funds to universities by the Higher Education Funding Council for impact on the subject under investigation, i.e. business. In marked
England (HEFCE) is heavily determined by a periodic assessment ex- contrast to the medical field, it is rare for companies to blame their
ercise, the Research Excellence Framework (REF), which measures failure or poor performance on following the advice of a fraudulent
quality-related research outputs such as journal publications.9 For academic publication. Indeed, much business school research published
many UK business schools, meeting REF criteria is incorporated into in elite journals has become so narrow, theoretically sophisticated and
performance evaluations, where faculty promotions and salary in- methodologically complex that its direct impact on business practice is
creases are heavily influenced by publications in journals on the ABS limited (Bennis and O’Toole, 2005; Banks et al., 2016). Falsifying the
list.10 results for a new pharmaceutical drug could lead to deaths, whereas
Pressures to publish, a preoccupation with elite journals as proxies falsifying data for a business journal publication is unlikely to directly
for research excellence, and the potential for high rewards if successful affect specific businesses, because managers (including those seconded
are all likely to foster a more extreme entrepreneurial risk-return per- to MBA classes) are unlikely to read – let alone fully comprehend –
spective, perhaps even shading into amorality, an issue discussed by papers published in elite journals. The risks of fraudulent business
Honig et al. (2014) and the subject of the next section. While an en- school research may therefore be downplayed by some in terms of the
trepreneurial risk-return perspective in itself may not necessarily lead mantra ‘no harm, no foul’. This problem thus combines issues of com-
to misconduct, one may hypothesize that, from the rational choice plexity with the rational choice perspective, in which the potential
perspective, the stronger the competitive pressures and the greater the damage of misconduct may be perceived as minor by the perpetrator
payoffs, the greater will be the temptation to engage in questionable when compared to the potential gains.
behaviour or misconduct (Shi et al., 2016).11 Consistent with strain
theory, some academics desperate for tenure approval or promotion
4.2. Do as I say, not as I do
may be tempted to engage in misconduct if they risk failing to achieve
their goals through legitimate means. While awareness of new plagi-
The larger problem, however, involves the ethical culture that
arism detection technologies, high profile retractions (see e.g. Furman
business schools develop and promulgate. Floyd et al.’s (2013, p. 753)
et al., 2012), and exposure by social media mechanisms may limit more
literature review concluded that the “reputation of business has been
blatant forms of misconduct, more sophisticated metrics-gaming and
besmirched with a continuous parade of financial scandals” with
the deliberate exploitation of certain ‘grey’ areas of misconduct may
widespread economic repercussions. According to Cavanagh (2009, p.
emerge in their place.
20), the leaders responsible for these scandals “are graduates of our
‘best’ business programs”, which have “failed to convey ethics, social
responsibility, and good moral habits to their graduates”.13 In response,
Floyd et al. (2013) call for a greater emphasis on ethical priorities
within business schools. Cabral-Cardoso (2004, p. 87) similarly stresses
(footnote continued) the importance of ethics as part of an institution’s core values, but poses
(Barczak, 2015).
8
See http://www.harzing.com/download/jql_journal.pdf (Accessed on 13 November
2017) for a comprehensive list of journal ranking schemes. 12
The use here of the term ‘entrepreneurial’ is not to imply that an entrepreneurial
9
The REF results influenced the distribution of £1.6 billion of research funds for 2016- risk-return perspective invariably results in amoral behaviour. We are merely pointing
17 – see http://www.hefce.ac.uk/media/HEFCE,2014/Content/Pubs/2016/201607/ out that, when competitive and other pressures are extreme, those who adopt such an
HEFCE2016_07.pdf (Accessed on 13 November 2017). approach may be more inclined to cut corners.
10 13
For a critical analysis of the detrimental effects of an overemphasis on ABS-listed For discussion of the role of business schools and their teaching in corporate scan-
journals, see Mingers and Willmott (2013). dals and the 2007 financial crisis, see Swanson and Frederick (2003), Ghoshal (2005),
11
See the two case-studies in Annex A, which provide illustrative support for this. Pfeffer (2005), and Giacalone and Wargo (2009).

4
J. Hall, B.R. Martin Research Policy xxx (xxxx) xxx–xxx

the question: “How can ethics instruction be taken seriously when ex- public exposure17 or attempt to cover up their actions. A third test
pediency and self-interest appear to overrule ethical considerations concerns intent, and whether the perpetrators believed they would re-
among faculty?” ceive some benefit from the questionable behaviour. Table 1 below sets
Although the risks of getting caught while engaging in blatant out illustrative examples of blatant misconduct and of inappropriate,
misconduct are growing with new plagiarism detection technologies questionable and appropriate conduct.
(Lee, 2011) and with grass-roots initiatives like PubPeer and Retraction Matters are complicated by the fact that for certain offences (e.g.
Watch, there are, as we discuss below, numerous ‘grey areas’ involving data falsification and plagiarism) the degree of severity can vary. Data
questionable research practices that remain problematic. Yet there is fabrication is an exception – it is essentially binary in that either the
limited discussion in the literature as to what research behaviour is data have been produced from the empirical study reported or they
specifically deemed questionable or inappropriate. In what follows, we have been fabricated. One prominent example involves Diederik Stapel,
attempt to fill this gap by developing a taxonomy of different forms and the social psychologist18 who fabricated data in several dozen studies
levels of research misconduct and other questionable behaviour. This (Tilburg University, 2012); by 2016, over 50 of his papers had been
includes examining the sources of these different forms of misconduct retracted (in some cases for data falsification rather than fabrication).
as seen from the rational choice, strain, cultural and network theories of Another is the accountancy professor, James Hunton, who fabricated
organizational misconduct, as well as the more specific entrepreneurial data in several articles (Bentley University, 2015) and has now had over
risk-return perspective reflected in much business school research. We 30 publications retracted.
discuss the paradoxical impact of over-simplistic research proxies and Data falsification represents a clear case of ‘blatant misconduct’. In
bounded rationality, where expectations of increasingly complicated the case of Dirk Smeesters, professor of consumer behaviour at
methodological approaches, greater emphasis on collaborations with Rotterdam School of Management, four papers were found to have
specialized expertise, and demands for more outputs in a narrow range “severe problems” with data manipulation (Erasmus University, 2014)
of elite journals are all playing a role in shaping research behaviour. To and his employment was terminated. Similarly, several papers by Ulrich
complete the taxonomy, we draw on the ‘no harm, no foul’ and ‘do as I Lichtenthaler were discovered to have problems relating to the mis-
say, not as I do’ discussion above to consider the degree of misconduct representation of the degree of statistical significance and/or deliberate
and to examine who is affected. ‘omitted variable bias’. His Habilitation was revoked19 and he had to
resign from his Chair of Management and Organisation at Mannheim
5. Towards a taxonomy of research misconduct University in 2015.20 However, data falsification can take less severe
forms including selective reporting and omitted data, and as such may
Blatant research misconduct is often contrasted with appropriate be regarded as ‘inappropriate’. Although there are rules regarding such
conduct − i.e. good scientific practice. There is not, however, a practices, they are sometimes ‘fuzzy’ in that there is no clear boundary
simple dichotomy but rather a continuum of behaviour that also in- (for instance, how many data outliers can be dropped before this be-
cludes ‘questionable’ and ‘inappropriate’ conduct. Blatant research comes ‘inappropriate’ or blatant?).
misconduct is clearly defined, with universally accepted rules covering Another form of questionable data manipulation is HARKing or
plagiarism and data falsification and fabrication in particular. The US Hypothesizing After the Results are Known (Kerr, 1998) – i.e. retro-
Federal Policy on Research Misconduct defines plagiarism as “The fitting hypotheses to one’s data in order to achieve high(er) statistical
appropriation of another person’s ideas, processes, results, or words significance (Martin, 2016). Such a practice may be encouraged by
without giving appropriate credit”. Data fabrication is “Making referees trying the help authors improve the statistical significance of
up data or results and recording or reporting them”, while data their findings (Anon., 2015), but it is generally considered ‘un-
falsification is defined as “Manipulating research materials, equipment, scientific’.21 On the other hand, many would regard Winsorization of
or processes, or changing or omitting data or results such that data – the assigning of lesser weight to an apparently spurious outlier –
the research is not accurately represented in the research record”.14 as acceptable (see e.g. Dixon and Tukey, 1968; Ghosh and Vogt, 2012).
Anderson et al. (2013) note that the US government limits legal In short, data manipulation ranges from blatant misconduct through
jurisdiction to blatant misconduct, leaving other questionable practices inappropriate and questionable conduct to broadly acceptable forms of
largely unregulated. ‘trimming’ or ‘tidying up’ of one’s data. The precise boundary regarding
‘Inappropriate’ research conduct is where certain rules exist, al- what is deemed unacceptable is at best indistinct and at worst a matter
though they may vary across countries, disciplines, institutions and of rather subjective interpretation.
journals (see Resnik et al., 2015 for an analysis of the variations in Plagiarism likewise covers a wide range in terms of seriousness.
misconduct definitions and policies). In contrast, ‘questionable’ re- Plagiarism of an entire paper (e.g. by serial plagiarist Hans Werner
search conduct15 is where there is an absence of clear and explicit rules. Gottinger – see Martin, 2007, 2012) or large sections (e.g. by Tony
However, one test here is that of ‘the reasonable reader’16 – i.e. what Antoniou, whose Deanship of Durham Business School was terminated
would s/he see as acceptable or unacceptable? Another test is whether when substantial parts of his PhD thesis and a journal article were
the individual engaged in the questionable behaviour would resist found to have been drawn from other sources – see Tahir, 2008) clearly
constitutes serious misconduct. Such cases are, however, relatively rare.
14
More common is lifting of entire sentences (e.g. in the literature review
See http://grants.nih.gov/grants/research_integrity/research_misconduct.htm (ac-
section of a paper) with no use of quotation marks and no indication of
cessed on 13 November 2017). See also Anderson et al. (2013) for a an extensive review
of the literature on research misconduct in the US, including definitions, prominent cases,
evidence on prevalence, factors encouraging misconduct, and how best to deal with the
problems.
17
“Would such behavior pass the New York Times front page test (i.e., how would the
15
One could further separate ‘dubious conduct’ from ‘questionable conduct’, with the author feel if the New York Times did a front page exposé of such academic practices)?”
former being defined as research conduct that most ‘reasonable readers’ would regard as (Anon., 2015, p. 216)
18
inadequate or unsuitable, while for ‘questionable conduct’ opinions are more evenly di- He had been Director of the Tilburg Institute for Behavioral Economics Research
vided between those who regard the conduct as acceptable and those who do not. (TIBER) and later Dean of the Tilburg School of Social and Behavioral Sciences.
19
However, it is probably better to keep the taxonomy relatively simple at this stage. See http://www.rhein-zeitung.de/region_artikel,-whu-erkennt-einstigem-starprofessor-
16
The notion of the ‘reasonable reader’ is well established in the legal sphere, in lehrbefaehigung-ab-_arid,1037700.html (Accessed on 15 December 2016; an archived ver-
particular with regard to law on libel. According to ‘the innocent-construction rule’, US sion can be found at https://archive.is/vDJLo).
20
courts must interpret the words “as they appeared to have been used and according to the This case is examined in more detail in Annex A, in particular the role that com-
idea they intended to convey to the reasonable reader” (see https://definitions.uslegal. petitive pressures played.
21
com/i/innocent-construction-rule/- accessed on 13 November 2017). See also McCraw One exception is when authors are completely open in acknowledging that their
(1991). hypotheses had been modified in the light of emerging results (Anon., 2015, p. 216).

5
 J. Hall, B.R. Martin

Table 1
A Taxonomy of Research Misconduct, Inappropriate and Questionable Conduct.
Source: An extensive review of the literature along with the authors’ own experiences as journal editors.

Nature of Appropriate conduct Questionable conduct Inappropriate conduct Blatant misconduct

behaviour
Widely accepted as good scientific practice Absence of clear rules but test of ‘the reasonable reader’; Rules exist although there may be some variation by Clearly defined and universally accepted rules
perpetrator embarrassed/reluctant to be revealed field, country, institution and/or journal

*
Data mani-pulation • Winsorization • HARKing • Selective reporting • Data fabrication (e.g. Stapel, Hunton)
• Omitted data • Data falsification (e.g. Lichtenthaler, Smeesters)
Use of work by • Drawing from and building on work of • Short phrases lifted from others and not put in • Entire sentences reproduced without source or • Plagiarism of entire article, whole section(s) etc. (e.g.
others others quotation marks quotation marks Gottinger, Antoniou)
• Failure to cite or acknowledge others • Wilfully omitting an entire body of work (e.g. in a
proposal)

6
Use of own work • Making every effort to diffuse one’s • Hyping own work/excessive self-citation • Self-plagiarism (e.g. Frey, Lichtenthaler) • Using the same theory or data to arrive at different
work conclusions (just for the sake of publishing another
paper)
• Avoiding excessive self-citation • Partial overlap with other papers by that author • Redundant publication
• Maximizing one’s research output • Salami publishing
Authorship • Including as authors all who have made • Obligatory authorship (e.g. expectation that a • Ghost authorship (e.g. Song) • Failure to declare an interest (e.g. Yang & Tao)
a substantial contribution PhD supervisor should be an author) • Gift authorship
• Gift colluding
Note: The above categories are not exhaustive; there are many other forms of misconduct (e.g. fake referees, citation cartels, journal impact factor (JIF) manipulation by editors) and the examples listed here are merely illustrative of the spectrum.
Note also that some of the above examples may not fall neatly into a single ‘box’ but extend over two or more degrees of severity (e.g. use of a ghost author to improve the written English may not be considered as ‘inappropriate’ or even
‘questionable’).
* Winsorization is the assigning of lesser weight to an apparently spurious outlier or modifying its value so it is closer to other sample values (Dixon and Tukey, 1968; Ghosh and Vogt, 2012).
Research Policy xxx (xxxx) xxx–xxx
J. Hall, B.R. Martin Research Policy xxx (xxxx) xxx–xxx

the source. While one or two cases in a paper might be passed off, most version of the paper was addressing a different readership. However,
editors will deem repeated instances to constitute ‘inappropriate’ re- with digital search engines enabling a paper on a given topic to be
search conduct. Another example is where material has been taken from quickly found wherever it is published, such behaviour is generally now
a particular source and paraphrased to some extent but the original deemed inappropriate24 unless explicit editorial approval has been
source is not cited at that point in the text. Both examples probably obtained. If found out, a duplicate or redundant paper may be retracted
come under the heading of ‘questionable’ research conduct – seen as by the journal involved.
wrong by many22 but not necessarily by everyone. Note, however, that Also closely related is the phenomenon of salami publishing, de-
editors tend to regard such errors as symptomatic of sloppy research fined as:
and, if detected and widespread, will often lead to desk rejection.
the deliberate attempt by an author or team of authors to in-
Besides the direct lifting of text, plagiarism also includes taking
appropriately inflate the total of publications yielded by a particular
ideas or other research material from others and failing to acknowledge
research study (or database, survey, experiment, project or what-
the source. Often harder to prove, this is perhaps best seen as a form of
ever) through a process of subdividing the published output into a
‘inappropriate’ conduct unless it is widespread and clearly deliberate. A
number of thin ‘slices’ or ‘least publishable units’, thereby either
less serious form may consist of hyping of one’s own work or excessive
generating a greater number of separate publications than is merited
self-citation, where the perpetrator may view this as causing no harm
by the overall contribution to knowledge offered by that study, or
and thus no foul, or even as a form of protection against self-plagiarism,
creating a situation where the research community would instead be
even though the outcome is an inflated citation score (Seeber et al.,
better served by the results being combined in a single or a smaller
2018). Here, there are often no clear rules so perhaps this is best de-
number of publications. (Martin, 2013, p. 1008)
scribed as ‘questionable’ conduct.
While plagiarism is a form of intellectual theft, self-plagiarism is a Salami publishing is becoming more common due to growing
form of intellectual deception (Martin, 2013, p. 1008), and can be de- competitive pressure on researchers and funding agencies expecting
fined as: value for money in the form of publications. Currently there are no
clear rules about how many papers one can legitimately obtain from a
“the practice by an author (or co-authors) of reproducing text, ideas,
given study (something that clearly depends on the scale of the project).
data, findings or other material from one or more earlier (or con-
Less scrupulous authors or those under intense competitive pressures
temporaneous) papers by the same author(s) without explicitly
may try to exploit that ambiguity. For example, Lichtenthaler produced
citing or otherwise acknowledging those other papers, thereby
over 20 articles from his PhD study (see Annex A, Supplementary ma-
misleading the reader (and in particular referees and editors) as to
terial).
the level of originality of that paper” (Martin, 2013, p. 1008).
Gift authorship, a form of network misconduct, is where researchers
Self-plagiarism includes a range of forms of differing severity. In are included as co-authors despite not having “participated sufficiently
some cases, essentially the same paper is republished in different in the work to take public responsibility for the content” (Vancouver
journals. For example, Bruno Frey (formerly professor at University of Group, 1985, p. 722). This can include adding senior or prestigious
Zurich) and colleagues published four versions of a study of who sur- authors to improve the appeal of the paper or out of a sense of ob-
vived the Titanic disaster, with each paper containing no cross-refer- ligation or pressure from senior colleagues. Junior authors may be in-
ences to the others. Such behaviour is generally seen as ‘inappropriate’. cluded in papers to help their tenure file or enhance their employment
For example, the editor of one of the affected journals publicly stated prospects. Some authors may ‘gift collude’ – where one author puts the
that he found such conduct “ethically dubious” (Autor, 2011, p. 239). name of another on one paper and in return has his/her name included
An investigation at Frey’s university concluded that “repeated failure to on another. Most journals have rules against such practices so this
publish the same work without consent of the editors and without cross- constitutes ‘inappropriate’ behaviour. However, it is almost impossible
references (‘Eigenplagate’)” was “improper” (University of Zurich, to detect, and often only becomes an issue when the integrity of a paper
2011), and his university contract was not renewed. is challenged and one of the authors claims they were unaware of what
For most journals, authors must attest that submitted papers have went wrong (Smith, 1994).
not been published nor are they under consideration elsewhere. There The converse of gift authorship is ghost authorship, where an actual
are thus clear rules, although it is unknown how effectively they are author is not named (e.g. for commercial or conflict-of-interest rea-
monitored. While authors may answer ‘Yes’ to this question, there are sons), a practice considered ‘inappropriate’ and proscribed by most
many cases where there is some overlap between the paper submitted journals.25 An example involves Michael Song, who, according to a
and others already published by the author(s) or currently under con- subsequent investigation (UMKC, 2015, p. 14), admitted that he may
sideration by other journals. Here, the issue centres on the degree of have written parts of the Yang and Tao (2012) paper described in
overlap. In the mind of the author, the papers may appear distinct, but Annex A, Supplementary material, which claimed that Song was the
the ‘reasonable reader’ referred to above may perceive there to be world’s top innovation management scholar!
significant overlap that affects the level of a paper’s originality, and The last category discussed here is failure to declare an interest in a
hence the decision on whether to publish. publication. A history of authors failing to declare a material interest
Related to self-plagiarism is redundant or duplicate publication.23 with the result that dubious research went unnoticed (at least initially)
This can be defined as “a paper where the existence of one or more prior has resulted in fairly universal rules among journals requiring the de-
(or contemporaneous) papers by the same author(s) means that the new claration of any relevant professional, institutional, commercial or
paper offers insufficient of an original contribution to knowledge to other material interest. Again, there are grey areas – i.e. when is an
merit publication in its own right” (Martin, 2013, p. 1008). In the pre- interest significant enough to declare and when not? This ambiguity has
digital age (where researchers could scan only a limited set of journals), been exploited by some who think that not declaring an interest may
duplicate publication may have been regarded as acceptable if each increase the likelihood of publication. For example, Yang and Tao’s

22 24
See, for example, the excellent online tutorials on plagiarism on the Indiana For example, it might sometimes be deemed acceptable for a second version to be
University website at https://www.indiana.edu/∼plag/examples.html (Accessed on 13 published in a practitioner journal with a very different readership, or a foreign language
November 2017). journal, but explicit permission from the editors involved and cross-referencing would
23
Redundant or duplicate publication may or may not involve self-plagiarism, de- still be required.
25
pending on whether the author cites or otherwise discloses the existence of the other For a legal approach to the problem of ghost authorship, see Stern and Lemmens
paper. (2011).

7
J. Hall, B.R. Martin Research Policy xxx (xxxx) xxx–xxx

Table 2
A Taxonomy of the Sources of Research Misconduct and Other Questionable Behaviour.

Type Examples of Behaviour Theoretical Sources of Misconduct (Greve Severity & Sample Corrective Measures
et al., 2010, plus others as shown)

Premeditated - Fully aware of rules but intent on breaking because risk- Rational choice (e.g. Arrow, 1963)
dishonesty reward not aligned
- Belief that getting caught is unlikely Cost-benefit analysis (Becker, 1968; Hornuf Very high
and Haas, 2014)
- Desperate to get published for fear of losing career Strain theory (Agnew, 1992; Lewellyn et al., Loss of research funding, employment
2017; Merton, 1938) termination, criminal charges
- Others have got away with it, so belief that this is the only Cultural theories (Schein, 1983, 1985;
way ahead Monteduro et al., 2016; Sims and Brinkmann,
2003)

Bending or gaming - Aware of rules but attempt to shift boundary between Rational Choice (Arrow, 1963) High
the rules appropriate and inappropriate conduct, exploiting unclear or
inconsistent rules for personal gain
- Belief that “anything goes” and “all that is not forbidden is Entrepreneurial risk-return perspective Public exposure, retraction of papers,
allowed”, often with specious post hoc justification (e.g. “I was (Honig et al., 2014) formal warning
told not to self-cite”) but with evidence of premeditation and/
or covering of tracks
Cultural theories (Schein, 1983, 1985;
Monteduro et al., 2016; Sims and Brinkmann,
2003)

Complexity and - Unclear or different rules, editorial policies, conventions, Bounded rationality (March and Simon, 1958;
ambiguity etc. Simon, 1969; Vaughan, 1999)
- General awareness of rules but open to interpretation Ambiguity (Fanelli, 2009, John et al., 2012; Medium, but potentially more serious if
(ambiguity) Johnson and Ecklund, 2016). signs of premeditation or cover-up
- Many co-authors, all of whom assume someone else makes Bounded rationality (March and Simon, 1958; Improved awareness, COPE guidelines,
final check Simon, 1969; Vaughan, 1999) clearer expectations of responsibilities
of co-authors
- Individual co-authors submit slightly different versions to Network theories (Ashforth et al., 2008; Breit
different journals and Forsberg, 2016)

Ignorance and - Researchers from different ‘cultures’ where norms/ Cultural theories (Schein, 1983, 1985;
sloppiness conventions different Monteduro et al., 2016; Sims and Brinkmann,
2003; Umphress et al., 2010).
- Lack of experience, research capabilities (e.g. PhD students, Bounded rationality (Simon, 1969; March and
junior researchers) Simon, 1958; Johnson and Ecklund, 2016)
- May have intended to sort out problem but ‘never got round
to it’

Honest mistake - Often claimed, but less credible for established researchers, Accidental misconduct (Vaughan, 1999). Low
and only valid if not systematic or part of a pattern Better training and supervision

(2012) paper on the best innovation management researchers and de- organizational misconduct. Table 2 presents a taxonomy of the various
partments failed to declare the pertinent fact that the authors had been types and examples of research misconduct, the theoretical foundations
visiting researchers at what the paper claimed was the top department on which they are based, the severity of the misbehaviour involved and
in the field. Furthermore, it conspicuously failed to mention that the relevant corrective measures. While we would contend that competitive
individual named as top researcher in the field (Song) had actually pressures and incentives are an ultimate cause of growing mis-
drafted parts of the paper (UMKC, 2015, p. 14). If in any doubt, the conduct,26 this table sets out different types of ‘proximate’ causes
author should simply declare that interest to the editor and let the latter commonly identified during misconduct investigations.
decide if it should be made public. Premeditated dishonesty, the most serious type of misconduct, is
So far, we have summarized various forms of questionable or in- where the perpetrator is aware of the rules and hides the misconduct in
appropriate conduct and blatant research misconduct. Although there the hope of getting away with it. From the perspective of rational
are relatively clear rules for what is regarded as appropriate and for choice theory (e.g. Arrow, 1963; Greve et al., 2010) or cost-benefit
what constitutes blatant misconduct, there is considerable ambiguity in analysis (Becker, 1968; Hornuf and Haas, 2014), the culprit has im-
between these two extremes. We have argued that questionable conduct plicitly if not explicitly weighed up the balance between the benefits of
can be distinguished from inappropriate conduct by the degree to which the misconduct (e.g. another publication) against the risk of being
there are rules precluding that specific form of conduct. However, an caught and the sanctions that may follow. According to strain theory
effective taxonomy ideally requires a sharper set of distinctions. To (Greve et al., 2010; Merton, 1938), the behaviour might be attributed to
achieve this, we next analyse the sources of research misconduct to help stressful circumstances (Lewellyn et al., 2017) or shear desperation,
us understand the pressures and thus the underlying intent and severity such as fears over tenure, employability or meeting grant obligations. It
of the behaviour. We then discuss the degree to which the behaviour may also be justified through cultural theory (Greve et al., 2010;
affects various stakeholders, providing a further means to distinguish
between questionable and inappropriate conduct.
26
We do not explore here the impact of technology on research misconduct, as our
primary focus is why, rather than how, research misconduct may occur. In the case of
6. Sources of research misconduct plagiarism, the growing availability of academic material online has facilitated and even
encouraged ‘cutting and pasting’, and a tendency among some to then ‘forget’ to para-
Section 2 outlined various theoretically derived sources of phrase or cite the lifted material.

8
J. Hall, B.R. Martin Research Policy xxx (xxxx) xxx–xxx

Schein, 1985), for example, where the culprit believes that social norms aware of the rules but perhaps rather ‘hazy’ on exact details. For ex-
(Monteduro et al., 2016) or deep cultural flaws (Sims and Brinkmann, ample, the perpetrator may have been aware of a problem (e.g. a failure
2003) encourage systemic corruption and the only way to succeed is to paraphrase material taken from other sources) and intended to re-
through misconduct. Often the misconduct is accompanied by evidence solve the problem later but failed to do so. Failure to properly check the
of intent, such as covering one’s tracks. An example is the Gottinger rules or simple incompetence may be the cause, especially for early-
case of serial plagiarism mentioned earlier, where the beginning and career researchers and PhD students (i.e. bounded rationality).
end of a plagiarized paper would be altered in an attempt to hide the Researchers from less scientifically developed countries may also lack
fact that the main body of the paper was plagiarized. The range of adequate training or claim cultural differences regarding research
sanctions involved in such cases include revocation of research funding, norms and conventions. The defence of ignorance is frequently invoked
being banned from the affected journals, loss of one’s job and perhaps but is not always credible, the problem instead often being more due to
even criminal proceedings. sloppiness or to naivety regarding the editorial process. Moreover, as in
The second type of misconduct consists of bending the rules, i.e. the other areas of life, “Ignorance of the law is no excuse”. Note that if the
individual is aware of the rules but is unilaterally trying to shift the publication reveals signs of premeditation or attempts by the author(s)
boundary between appropriate and inappropriate conduct in their fa- to cover their tracks, claims of ignorance or sloppiness are invalid,
vour, often exploiting unclear or inconsistent rules. Like premeditated placing the case in a more serious category. The corrective measures for
dishonesty, there may be elements of rational choice, cost-benefit and such cases are broadly similar to those for the previous category, but
strain theories, although the underlying logic is an extreme version of with particular emphasis on better training and increased awareness.
the entrepreneurial risk-return perspective (Honig et al., 2014), The least serious type of behaviour involves honest errors or gen-
whereby culprits often justify their actions by seeing unclear or in- uine mistakes, which, while not true forms of misconduct or in-
consistent rules as opportunities to exploit for personal gain. When appropriate behaviour, are nevertheless problematic. The key issue is
challenged, the individual frequently offers specious reasons for their how the issue is addressed once identified. Such cases can only be a
behaviour (for instance, if challenged about self-plagiarism, they may one-off – if repeated or part of a pattern, then this is not a credible
claim “My supervisor told me not to cite my own work”). Such culprits defence, and there must also be no evidence of premeditation or at-
tend to have a very restricted and self-serving sense of morality, best tempted cover-up, otherwise it belongs in a more serious category.
summed up in the phrase “All that is not forbidden is allowed”. As Honest mistakes are often claimed but are a less credible defence for
discussed above, such amoral attitudes towards ethical behaviour in established researchers. They may, however, be common among scho-
business schools have been partly responsible for a culture that may lars under pressure to juggle numerous projects, teaching and admin-
have facilitated many of the financial scandals in recent years (Cabral- istrative tasks, i.e. a complexity issue (Vaughan, 1999).
Cardoso, 2004; Cavanagh, 2009; Floyd et al., 2013). Corrective mea- Consistent with our distinction between inappropriate and ques-
sures include rejection or retraction of papers and the issuing of formal tionable research conduct, the severity of the case (and the ensuing
warnings, as well as punitive measures such as public exposure and, sanction) is greater when: (i) there is evidence of attempts by the per-
depending on the severity and persistence of the behaviour, termination petrator to hide the mistake; (ii) a repeat offender is involved; (iii) the
of employment. offender is a more experienced researcher in an institution or country
The next two types of behaviour are less serious, though still un- where the rules are clear and the action would thus be seen as un-
acceptable. The first includes complexity, a situation where many in- acceptable to the reasonable reader; and (iv) the offender would receive
teracting variables result in bounded rationality (Simon, 1969) making some undue benefit from the ‘mistake’. The main approach to dealing
mistakes likely (Vaughan, 1999), and ambiguity, where rules may be with honest errors or genuine mistakes consists of better training and
open to interpretation (Fanelli, 2009, John et al., 2012; Johnson and supervision of younger researchers and others who have erred.
Ecklund, 2016). Examples include unclear or inconsistent rules, con- Having outlined the various forms of research conduct and the
ventions or editorial policies. Another example might be a paper with sources from which they are derived, we now turn to the final com-
numerous co-authors where nobody made a final check, or where one ponent of our taxonomy, the degree to which the behaviour affects
co-author submitted one version of a paper to one journal and another different stakeholders.
to a second journal without checking with each other.27 However, if
there are any signs of premeditation, this puts the case in a more serious
7. Who is affected by research misconduct, and by how much?
category. Although complexity can be a legitimate cause of mistakes
(discussed below), it is often used in an attempt to retrospectively
Misconduct, inappropriate and questionable conduct can have an
justify questionable behaviour.
impact on a range of stakeholders. Table 3 below sets out a typology of
Note that complexity is likely to increase over time, given that elite
the main stakeholders affected and indicates, for illustrative purposes,
journals expect ever more sophisticated methodologies and theoretical
the severity of the impact on each. Most directly, misconduct may have
contributions, while funding agencies may call for greater inter-
consequences for other researchers. Those who build on research sub-
disciplinary collaborations among stakeholders with specialized ex-
sequently found to be tainted lose credibility. Victims of plagiarism
pertise (Hall et al., 2014) along with demands for more outputs. Breit
often feel a strong sense of violation, and may subsequently suffer
and Forsberg (2016) suggest that such increased pressures for greater
where, for example, citations are effectively ‘stolen’ from them, redu-
network collaboration can contribute towards research misconduct
cing their standing within the academic community and hindering
because inter alia collaborators may provide information of dubious
promotion opportunities. Other forms of misconduct with serious con-
practices, such as how to get away with things in order to meet per-
sequences include failure to declare an interest and ghost authorship,
formance targets. Among the corrective measures are clearer rules and
both of which may result in biased or unreliable results being given
guidelines, improved awareness, greater use of COPE guidelines,28 and
undue credibility, potentially misleading those who later draw upon
clearer expectations with regard to co-authors’ responsibilities.
that work.29 Likewise, undue weight may be given to a paper’s results in
Ignorance and sloppiness may occur when the perpetrator is broadly
the case of gift authorship involving a high-status researcher.
For self-plagiarism, the direct consequences are felt mainly by the
perpetrator, who may lose peer status and perhaps even their job if
27
Another example of complexity is ‘Climategate’, where the researchers, in at-
tempting to deal with a wide array of stakeholders with very different agendas, ended up
29
resorting to somewhat dubious research practices (Garud et al., 2014). In medical research, the consequences of ghost authorship can be far more severe,
28
See https://publicationethics.org/ (Accessed on 13 November 2017). for example, influencing regulators to approve a drug.

9
 J. Hall, B.R. Martin

Table 3
Stakeholders affected by research misconduct.

Type of behaviour Other researchers Employers Students Editors/journals Societal stakeholders

Blatant Misconduct Data fabrication Authors who build on Perpetrator’s inflated publication record May lead some to feel rewards Serious damage to journal’s Erroneous implications for
Data falsification tainted work lose → others lose out Damage to reputation of of cheating greater than risks & reputation Major effort needed to practice
credibility institution costs of getting caught investigate
Plagiarism Authors whose work stolen Possible effect on meta-reviews
lose status etc.
Failure to declare an Biased/unreliable results – Damage to reputation of institution Legitimizes unethical behaviour Erroneous implications for
interest misleads others practice

Inappropriate Selective reporting, Unreliable results – misleads Perpetrator’s inflated publication record → May lead some to feel that can ‘bend’ Major effort needed to Erroneous implications for practice
conduct omitted data others others lose out Salary, promotion costs based on the rules or belief that ‘all that is not investigate/police Repeated
Sentences lifted Those plagiarized lose potential dubious productivity Damage to reputation of forbidden is allowed’ infringe-ments → damage to Possible effect on meta-reviews
without attribution citations + status institution journal’s reputation
Failure to cite or
acknowledge others

10
Self-plagiarism Takes publication slots away
Redundant publication from legitimate research
Gift authorship If high status co-author, undue If high status co-author, undue
credibility given to results weight given to results
Ghost authorship Misleads others Reputational damage Erroneous implications

Questionable HARKing Undue weight given to May lead some to feel that can Major effort needed to investigate/ May → erroneous implications for
conduct results/author (→ others lose ‘bend’ the rules or belief that ‘all police Repeated infringe-ments may → practice
out) that is not forbidden is allowed’ damage to reputation of journal
Short phrases lifted & Those plagiarized lose potential Perpetrator’s inflated publication record → Possible effect on meta-reviews –
not put in quotes citations + status others lose out may → erroneous implications for
Hyping own work/ Annoyance with author – Damage to reputation of institution practice
excessive self-citation weakens reputation
Partial overlap with
other papers by author
Salami publishing

Code: Effect on others shown in bold – Severe Impact. Effect on others shown in italics – Significant Impact; Effect on others shown in normal typeface – Minor Impact.
Research Policy xxx (xxxx) xxx–xxx
J. Hall, B.R. Martin Research Policy xxx (xxxx) xxx–xxx

discovered (for example, Frey’s contract was not renewed by Zurich main conclusions emerging from the literature review and from our
University). Similarly, the effects of redundant publication and salami efforts to develop a useful taxonomy of research misconduct.
publishing are directly felt mainly by the perpetrator in the form of
diminished reputation, although other researchers may spend un-
necessary time ploughing through additional papers that make little or 8. Discussion and conclusions
no additional contribution. However, as discussed below, there are
important indirect effects, such as perpetrators being hired or promoted Growing publishing pressures in academia for improved individual
over others who do not engage in such behaviour. Furthermore, senior and institutional success appear to have created perverse incentives
scholars often set the cultural norms within their communities, which (Bouter, 2015), encouraging gaming of the system and pushing the
may encourage others to cut corners. In the case of HARKing, the boundaries of appropriate research behaviour (Martin, 2016). While
consequence is that other researchers may give undue weight to the examples of blatant misconduct are sometimes easier to identify, they
results or to the author, with others thus losing out in terms of the are only part of a far wider problem of inappropriate research beha-
attention and status accorded. It may also raise expectations for future viour (Butler et al., 2017), much of which remains poorly understood
research, artificially increasing the bar for publication. (Fanelli, 2009; John et al., 2012) and is often open to ambiguous in-
A second stakeholder category consists of employers. Perpetrators of terpretation (Johnson and Ecklund, 2016). To address this ambiguity,
misconduct may have been rewarded with promotion or increased we have developed a taxonomy that differentiates various forms of
salaries for an inflated publication score at the expense of others in the research behaviour, ranging from appropriate practice to blatant mis-
organization. In addition, there will almost certainly be some damage conduct, and specifically focusing on behaviour between these ex-
to the organization’s reputation depending on the severity of the mis- tremes. Our aim is to provide clearer, more consistent guidelines for
conduct. However, in some cases, a business school may benefit (at researchers, especially those beginning their careers, in an attempt to
least in the short run) through enhanced funding or renewed accred- reduce ambiguity and hence reverse the trend towards dubious research
itation. Indeed, some individuals may even attempt to justify their behaviour apparent in many fields (Fanelli, 2009; Van Noorden, 2011).
misconduct in terms of helping their organisation, i.e. what Umphress Following Anderson et al.’s (2013) call for research on actual re-
et al. (2010) term ‘unethical pro-organizational behaviour’. search behaviour rather than hypothetical situations, we cite real ex-
The third category of those affected by misconduct consists of stu- amples from business school research, where competitive pressures and
dents. If they become aware that their lecturers are engaging in and incentives to engage in misconduct seem particularly acute (Honig and
being rewarded for misconduct, some may sense that the rewards for Bedi, 2012). At the same time, increasingly sophisticated theoretical
cheating are greater than the risks of getting caught, or that they, too, and methodical approaches may have distanced much business school
are entitled to ‘bend’ the rules. As various studies have suggested, this research from practice (Bennis and O’Toole, 2005; Banks et al., 2016),
tendency may be particularly pronounced among economics and busi- encouraging a ‘no harm, no foul’ justification for dubious behaviour.
ness school students (e.g. Frank et al., 1996; Klein et al., 2006; McCabe We attempted to refute this excuse by identifying the stakeholders af-
et al., 2006; Giacolone and Thompson, 2006; Giacalone and Wargo, fected by such behaviour, as well as the pernicious influence it may
2009; Brown, 2011; Wang et al., 2011). have on promulgating unethical decision-making in the wider business
A fourth category of stakeholders comprises journal editors. Serious community (Floyd et al., 2013). If faculty are perceived to be engaged
forms of misconduct require considerable resources to investigate the in gaming the system or other dubious conduct, such an amoral culture
allegations. This includes putting together a case, seeking a response is likely to be reproduced among graduates, affecting their future be-
from the defendant(s), considering that response, determining the haviour and the organizations in which they work.
outcome, and, where the misconduct is proven, deciding an appropriate Several authors have distinguished appropriate research behaviour
sanction. This generally involves following a formal procedure as set from blatant misconduct, while recognizing there is a range of beha-
out by COPE. Such cases typically take several person-months of effort, viours in between. Building on this, we developed a taxonomy that
time that could be better spent on other editorial tasks. Moreover, when includes the nature and severity of different forms of research mis-
misconduct has been detected, their journal’s reputation can suffer conduct, its sources and the stakeholders it affects. We have shown how
since their editing and review procedures may be blamed for not de- one can combine several bodies of theoretical work, including rational
tecting the problem sooner. choice, strain, cultural, network, and bounded rationality theories of
The last category consists of societal stakeholders such as firms, misconduct. By doing so, we have presented one of the first attempts to
government departments and NGOs that utilize the research. Here, construct a more formal and theoretically based taxonomy30 that can
forms of misconduct such as data fabrication and falsification, failure to help deal with the growing problem of research misconduct. In parti-
declare an interest, ghost authorship and HARKing can all result in cular, it may be useful for training PhD students and young researchers,
erroneous practical implications for management or policy. In the case providing clarity regarding the boundary between what is acceptable
of gift authorship involving a high-status individual, the possible con- and what is not, and encourage scholars to think about the propriety
sequence is undue weight being accorded to the results. However, for and ethics of their behaviour rather than just relying on ‘the rules’. It
plagiarism, self-plagiarism, redundant publication and salami pub- may also be useful for researchers when reflecting on their research and
lishing, the consequences for external stakeholders are generally less publication strategies, especially when confronted with pressures to
pronounced, except where meta-reviews are involved, in which case the increase their output and maximise their ‘score’ on some particular
findings of the undetected plagiarist or self-plagiarist may be given metric. In addition, the typology may be useful for editors and others
undue weight. responsible for ensuring research integrity and determining where au-
This section completes our taxonomy of research misconduct and thors have strayed across the boundary between what is acceptable and
other forms of dubious behaviour. By distinguishing between the dif- what is not, whether intentionally or otherwise.
ferent types of stakeholders affected by misconduct, we can obtain a From the taxonomy and the theoretical perspectives on which it is
clearer view as to who is most seriously affected and to what extent. based, we present in Table 4 a number of testable propositions that
The results provide further grounds for classifying particular forms of might be the subject of future empirical work on research misconduct.
behaviour in the different categories of blatant misconduct (where the
impact on others is most severe), inappropriate conduct (where the 30
The only other taxonomy we have come across is that developed by Helton-Fauth
impact on others is generally not as severe but is nevertheless sig- et al. (2003) but this is rather different in nature, providing a taxonomy of ‘ethical events’
nificant) and questionable conduct (where the impact on others is re- in research, with the four main categories relating to ‘data’, ‘study conduct’, ‘professional
latively minor or less direct). In the final section, we synthesize the practices’ and ‘business practices’.

11
J. Hall, B.R. Martin Research Policy xxx (xxxx) xxx–xxx

Table 4
Some testable propositions.

Theoretical approach Proposition

Rational choice theory • The greater the rewards and pressures for high performance in academic organizations, the more prevalent research misconduct tends
to be.
• When penalties for misconduct are seen as minor or trivial, the more prevalent research misconduct tends to be.
• The lower the probability that misconduct cases are followed up and investigated, the more prevalent research misconduct tends to be.
• When rules on misconduct are lacking or unclear, the more prevalent research misconduct tends to be.
• When the direct impact of misconduct is seen as trivial (i.e. ‘no harm, no foul’), the more prevalent research misconduct tends to be.
Strain theory • The greater the pressures on individual academics in an organization to perform, the more they come to see research misconduct as
justified and the more prevalent it tends to be.
• The more intense the competitive pressures on academic organizations, the more prevalent research misconduct tends to be.
Cultural theory • The more senior academics (e.g. leading researchers, journal editors, business school deans) are seen to be engaged in gaming or other
dubious practices, the more prevalent research misconduct tends to be.
• The less effort invested by academic organizations in the ethical socialization of their staff, the more prevalent research misconduct
tends to be.
• The laxer the enforcement of ethical policies by academic organizations, the more prevalent research misconduct tends to be.
• The laxer the enforcement of ethical policies by journal editors, the more prevalent research misconduct tends to be.
• The more a national or disciplinary culture adheres to a belief that ‘all that is not forbidden is allowed’, the more prevalent research
misconduct tends to be.
Network theory • The more organizations are involved in a research collaboration with different performance expectations, the more prevalent research
misconduct tends to be.
• The
be.
more internally well connected but globally isolated a part of a research network, the more prevalent research misconduct tends to

Complexity and bounded rationality • Research that involves complicated methodologies is more likely to involve misconduct than in the case of simpler methodologies
where it is more difficult to mask.
• misconduct
The more a research collaboration requires different specialist expertise, the more difficult it is for members to know whether
is present, and thus the more likely it is to be prevalent.

Note: ‘Misconduct’ is used in the above table to include the wider spectrum of dubious and inappropriate behaviour as well as more blatant misconduct.

(Note that these are merely illustrative rather than intending to offer a provided it is not detected.
comprehensive list.) Studies investigating the propositions derived from network the-
The rational choice propositions are predicated on situations where ories of misconduct could compare how increased network collabora-
the pressures and rewards are high, while the misconduct or ques- tion in academia affects misconduct, especially when there are di-
tionable research behaviour may be rationally (albeit not morally) vergent expectations and pressures from different sources and
justified because it is perceived as marginal or not easily detected. To organizational cultures (Breit and Forsberg, 2016). For example, studies
test these propositions, one could compare organizations characterized comparing networks that are internally well connected but globally
by high pressure and rewards with those that are more relaxed. In cases isolated could be conducted, specifically focusing on whether it is pri-
where the misconduct is serious, corrective measures could include marily individuals who influence the network (i.e. the ‘bad apples’ as-
threats to research funding, termination of employment or even crim- sumption) or vice versa (‘bad barrels’), as suggested by Ashforth et al.
inal charges. However, more common is where it is implicitly under- (2008).
stood that, if dubious activities seemingly result in limited harm, then Studies relating to the propositions on complexity and bounded
no great efforts are necessary to follow up and investigate them. rationality could be conducted from various perspectives. One might
However, such a ‘no harm, no foul’ attitude misses the wider implica- assess whether increasingly complicated methodologies offer greater
tions of academic misconduct, where a ‘do as I say, not as I do’ attitude opportunities for misconduct. As Bennis and O’Toole (2005) and Banks
among faculty fails to convey appropriate ethical standards to students et al. (2016) note, business school research has become quite theore-
(Cabral-Cardoso, 2004; Cavanagh, 2009; Floyd et al., 2013). The ra- tically sophisticated and methodologically complex. Hence, reviewers
tional choice perspective is hence rather limited when it comes to un- and editors may be unable to easily detect misconduct. Such research
derstanding and avoiding less blatant forms of research misconduct. may also be too narrow or abstract to directly affect practitioners, with
Strain theory differs from rational choice perspectives by focusing the result that misconduct may be seen as justified because it causes
on career concerns and the perceived risk of losing out to one’s peers as little real harm, and thus ‘no foul’. Another approach might be to ex-
justification for dubious behaviour. To test our propositions here would plore whether relying on specialist expertise opens up opportunities for
therefore require understanding psychological factors such as how misconduct, which in turn may depend on how far partners are trusted
stressful circumstances create negative emotions (Lewellyn et al., (Anderson et al., 2013).
2017), which in turn are used to justify such behaviour. However, there One overarching proposition related to complexity (not listed in the
remains the question whether career threats are indeed the true cause table) is whether misconduct can be attributed to a single factor, or
of misconduct or merely a post hoc justification. whether it stems from several causes. We would suggest that mis-
The propositions associated with cultural theories involve social conduct is more likely to derive from a number of complex factors ra-
norms that are influenced, for example, by level of education, social ther than a single cause, and then may be ‘justified’ in terms of other
development and national characteristics (Monteduro et al., 2016). reasons. This points to a potential limitation of our theoretical frame-
These propositions might be tested by examining the degree to which work, derived as it is from Greve et al.’s (2010) categorization of the-
organizational cultures are competitive, or whether an organization’s oretical approaches to organizational misconduct. Nevertheless, by
norms, values and beliefs might lead some to conclude that certain improving our understanding of why misconduct may occur, whom it
forms of misconduct were acceptable. As Greve et al. (2010) note, it is affects and the direct and indirect impacts it may have, we hope to
the ethical perspective rather than procedural rules, threats or coercion reduce its occurrence.
(the focus of rational choice and strain theories) that should be the This brings us to the limitations of our study and options for future
focus of such studies, particularly if the organizations involved foster a research. First, the taxonomy presented here is not intended to provide
culture that outwardly condemns misconduct but implicitly supports it, a comprehensive list of all forms of misconduct and questionable

12
J. Hall, B.R. Martin Research Policy xxx (xxxx) xxx–xxx

behaviour. Indeed, new forms of misconduct will doubtless emerge as Acknowledgements

old ones become more easily detected and policed. Further research
could explore, for example, how new technologies are utilized to en- The authors are grateful to three Research Policy referees and two
gage in, detect or pre-empt misconduct. However, our classification of referees of the SPRU Working Paper Series for detailed comments on an
the main sources of misconduct and dubious behaviour, ranging from earlier version, along with much helpful advice from the Guest Editors
the most serious and unambiguous through to honest error, provides a of this Special Issue.
framework to help understand why different types of misconduct
emerge. As we have shown, misconduct in its various forms has adverse Annex A Supplementary material - two case studies
consequences for a range of stakeholders. By analysing and classifying
the impacts, we have sought to explain why certain behaviours result in Supplementary material associated with this article in the form of
unfair advantage, providing a more complete picture of who the mis- two case studies can be found in the online version, at https://doi.org/
conduct affects and how. However, further research and discussion 10.1016/j.respol.2018.03.006.
within the research community is needed to clarify the sometimes
ambiguous boundaries between appropriate and inappropriate research References
conduct, and to improve awareness so there is less scope for the un-
scrupulous to exploit current ambiguities. Adams, D., Pimple, K.D., 2005. ‘Research misconduct and crime: lessons from criminal
Secondly, our study was confined to business school research. Other science on preventing misconduct and promoting integrity’. Account. Res. 12,
225–240.
fields may experience different problems with misconduct, so our Adler, N.J., Harzing, A.W., 2009. ‘When knowledge wins: transcending the sense and
findings may not be generalizable to all academic areas and circum- nonsense of academic rankings’. Acad. Manag. Learn. Educ. 8, 72–95.
stances. However, business schools provide a useful illustration because Agnew, R., 1992. Foundation for a general strain theory of crime and delinquency.
Criminology 30, 47–87.
they exhibit a number of traits, including intense competitive pressures Anderson, M.S., Shaw, M.A., Steneck, N.H., Konkle, E., Kamata, T., 2013. Research in-
and significant incentives and rewards for gaming or pushing the tegrity and misconduct in the academic profession. In: In: Paulsen, M.B. (Ed.), Higher
boundaries in order to gain advantage. Such traits are likely to be Education: Handbook of Theory and Research 28. Springer, Dordrecht, pp. 217–261.
Anon, 2015. The case of the hypothesis that never was; uncovering the deceptive use of
prevalent (if not as extreme) in other fields, particularly those that are
post hoc hypotheses. J. Manag. Inq. 24, 214–216.
relatively fast-moving, and may be particularly pertinent with regard to Arrow, K.J., 1963. Uncertainty and the welfare economics of medical care. Am. Econ.
‘grey’ areas of misconduct. Rev. 53, 941–973.
Ashforth, B.E., Gioia, D.A., Robinson, S.L., Trevino, L.K., 2008. Re-viewing organizational
Thirdly, our taxonomy is derived mainly from an organizational
corruption. Acad. Manage. Rev. 33 (3), 670–684.
studies perspective – a field with an extensive history of exploring Autor, D.H., 2011. Letter to Professor Bruno Frey. J. Econ. Perspect. 25, 239–240.
misconduct – and specifically from Greve et al.’s (2010) categorization Banks, G.C., Pollack, J.M., Bochantin, J.E., Kirkman, B.L., Whelpley, C.E., O’Boyle, E.H.,
of the sources of misconduct. We recognize that other perspectives such 2016. Management’s science-practice gap: a grand challenge for all stakeholders.
Acad. Manag. J. 59, 2205–2231.
as those from psychology,31 sociology or law, may also provide useful Barczak, G., 2015. Expression of concern. J. Prod. Innov. Manag. 32, 655.
insights. In particular, a legal perspective could shed light on the jur- Barnett, M.L., 2014. Why stakeholders ignore firm misconduct: a cognitive view. J.
isdictional dimension to the taxonomy, i.e. who should be responsible Manag. 40 (3), 676–702.
Becker, G.S., 1968. Crime and punishment: an economic approach. J. Polit. Econ. 76,
for monitoring and policing, and what legal measures could be em- 169–217.
ployed.32 Indeed, since misconduct is often likely to involve a number Bennis, W.G., O’Toole, J., 2005. How business schools lost their way. Harv. Bus. Rev. 83
of factors, exploring them through only one theoretical lens may not (5), 96–104.
Bentley University, 2015. Report of Judith A. Malone, Bentley University Ethics Officer,
adequately explain the phenomenon. For example, collaboration is Concerning Dr. James E. Hunton. . downloaded from https://www.bentley.edu/files/
often hailed as a fruitful mechanism for generating new knowledge and Hunton%20report%20July21.pdf. (Accessed 13 November 2017).
research capabilities (as seen from the perspective of network theory) Bouter, L.M., 2015. Commentary: perverse incentives or rotten apples? Acc. Res. 22,
148–161.
among research disciplines, regions/countries and industry/policy Breit, E., Forsberg, E.-M., 2016. Scientific misconduct and integrity: an organizational
makers (cultural theories). However, increased incentives to collabo- perspective. Report for the EU Horizon 2020 Project on ‘Promoting Integrity as an
rate (rational choice theory), coupled with more intense pressures to Integral Dimension of Excellence in Research’. . downloaded from http://printeger.
eu/wp-content/uploads/2016/10/D2.6.pdf. (Accessed 13 November 2017).
publish (strain theory), are likely to increase the number of co-authors
Brown, B.S., 2011. The academic ethics of graduate business students: 1993 to 1998. J.
(complexity), some of whom may not be legitimate contributors or who Appl. Bus. Res. 16 (4), 105–112.
may engage in collusion or gift authorship. Butler, N., Delaney, H., Spoelstra, S., 2017. The grey zone: questionable research practices
In conclusion, the taxonomy developed here represents an attempt in the business school. Acad. Manag. Learn. Educ. 16, 94–109.
Cabral-Cardoso, C., 2004. Ethical misconduct in the business school: a case of plagiarism
to achieve greater clarity with regard to research misconduct, particu- that turned bitter. J. Bus. Ethics 49, 75–89.
larly the more ambiguous forms. It is not intended to be a static fra- Cavanagh, G., 2009. What’s good for business. America 20 (4), 20–22.
mework, but rather to offer a starting point for understanding why Dixon, W.J., Tukey, J.W., 1968. Approximate behavior of the distribution of Winsorized t
(trimming/Winsorization). Technometrics 10, 83–98.
research misconduct emerges, who it affects (both directly and in- Eisenhardt, K.M., Graebner, M.E., Sonenshein, S., 2016. Grand challenges and inductive
directly) and to what degree. Indeed, we recognize that professional methods: rigor without rigor mortis. Acad. Manag. J. 59 (4), 1113–1123.
norms and rules need to respond to emerging new types of misconduct Erasmus University, 2014. Report of the Smeesters Follow-Up Investigation Committee. .
downloaded from https://www.eur.nl/fileadmin/ASSETS/press/2014/maart/
by those determined to cheat or game the system. By providing greater Report_Smeesters_follow-up_investigation_committee.final.pdf. (Accessed 13
clarity, we seek to counter some of the excuses commonly used to November 2017).
justify misconduct, and to temper the more extreme forms of en- Espeland, W.N., Sauder, M., 2007. Rankings and reactivity: how public measures recreate
social worlds. Am. J. Sociol. 113, 1–40.
trepreneurial risk-return/rational choice outlook adopted by certain Fanelli, D., 2009. How many scientists fabricate and falsify research? A systematic review
researchers and their organizations. We hope that the taxonomy pre- and meta-analysis of survey data. PLoS One 4, e5738.
sented here, and other articles in this special issue, can serve to dis- Floyd, L., Xu, F., Atkins, R., Caldwell, C., 2013. Ethical outcomes and business ethics:
toward improving business ethics education. J. Bus. Ethics 117, 753–776.
courage dubious behaviour and thus help to shape future good research
Frank, R.H., Gilovich, T.D., Regan, D.T., 1996. Do economists make bad citizens? J. Econ.
practice. Perspect. 10, 187–192.
Furman, J.L., Jensen, K., Murray, F., 2012. Governing knowledge in the scientific com-
munity: exploring the role of retractions in biomedicine. Res. Policy 41, 276–290.
Garud, R., Gehman, J., Karunakaran, A., 2014. Boundaries, breaches, and bridges: the
case of Climategate. Res. Policy 43, 60–73.
31 Ghosh, D., Vogt, A., 2012. Outliers: an evaluation of methodologies. Section on Survey
See, for example, Mumford et al. (2009) for a study of the influence of personality on
Research Methods – JSM 2012. pp. 3455–3460. downloaded from http://ww2.
research misconduct.
32 amstat.org/sections/srms/Proceedings/y2012/files/304068_72402.pdf. (Accessed
We are grateful to an anonymous referee for this suggestion.

13
J. Hall, B.R. Martin Research Policy xxx (xxxx) xxx–xxx

13 November 2017). BusinessWeek rankings. Acad. Manag. Learn. Educ. 7, 26–41.

Ghoshal, S., 2005. Bad management theories are destroying good management practices. Mumford, M.D., Antes, A.L., Beeler, C., Caughron, J.J., 2009. On the corruption of sci-
Acad. Manag. Learn. Educ. 4, 75–91. entists: the influence of field, environment, and personality. In: Burke, R.J., Cooper,
Giacalone, R.A., Wargo, D.T., 2009. The roots of the global financial crisis are in our C.L. (Eds.), Research Companion to Corruption in Organizations. Edward Elgar,
business schools. J. Bus. Ethics 6, 1–24. Cheltenham, pp. 145–170.
Giacolone, R.A., Thompson, K.R., 2006. Business ethics and social responsibility educa- Palmer, D., Smith-Crowe, K., Greenwood, R., 2016. The imbalances and limitations of
tion: shifting the worldview. Acad. Manag. Learn. Educ. 5, 266–277. theory and research on organizational wrongdoing. In: Palmer, D., Smith-Crowe, K.,
Gioia, D.A., Corley, K.G., 2002. Academy of Management Learning and Education 1. pp. Greenwood, R. (Eds.), Organizational Wrongdoing: Key Perspectives and New
107–120. Directions. Cambridge University Press, Cambridge.
Greve, H., Palmer, D., Pozner, J., 2010. Organizations gone wild: the causes, processes, Pfeffer, J., 2005. Why do bad management theories persist? A comment on Ghoshal.
and consequences of organizational misconduct. Acad. Manag. Ann. 4, 53–107. Acad. Manag. Learn. Educ. 4, 96–100.
Grzesiuk, K., 2016. Unethical behaviour of organisations from a social network per- Resnik, D.B., Rasmussen, L.M., Kissling, G.E., 2015. An international study of research
spective – a literature review. Ann. Ethics Econ. Life 19 (4), 23–37. misconduct policies. Account. Res. 22, 249–266.
Hall, J., Matos, S., Bachor, V., Downey, R., 2014. Commercializing university research in Schein, E.H., 1983. The role of the founder in creating organizational culture. Organ. Dyn.
diverse settings: moving beyond standardized intellectual property management. Res. 12, 13–28.
Technol. Manag. 57 (5), 26–34. Schein, E.H., 1985. Organizational Culture and Leadership. Jossey-Bass, San Francisco.
Helton-Fauth, W., Gaddis, B., Scott, G., Mumford, M., Davenport, L., Connelly, S., Brown, Seeber, M., Cattaneo, M., Meoli, M., Malighetti, P., 2018, Self-citations as strategic re-
R., 2003. A new approach to assessing ethical conduct in scientific work. Acc. Res. 10, sponse to the use of metrics for career decisions, Research Policy (forthcoming –
205–228. available online at https://doi.org/10.1016/j.respol.2017.12.004).
Honig, B., Bedi, A., 2012. The fox in the hen house: a critical examination of plagiarism Shadnam, M., Lawrence, T.B., 2011. Understanding widespread misconduct in organi-
among members of the Academy of Management. Acad. Manag. Learn. Educ. 11, zations: an institutional theory of moral collapse. Bus. Ethics Quart. 21 (3), 379–407.
101–123. Shi, W., Connelly, B., Sanders, G., 2016. Buying bad behavior: tournament incentives and
Honig, B., Lampel, J., Siegel, D., Drnevich, P., 2014. Ethics in the production and dis- securities class action lawsuits. Strateg. Manag. J. 37, 1354–1378.
semination of management research: institutional failure or individual fallibility? J. Siggelkow, N., 2007. Persuasion with case studies. Acad. Manag. J. 50 (1), 20–24.
Manag. Stud. 51, 118–142. Simon, H.A., 1969. The Sciences of the Artificial. MIT Press, Cambridge, Mass.
Hornuf, L., Haas, G., 2014. Regulating fraud in financial markets: can behavioural designs Sims, R., Brinkmann, J., 2003. Enron ethics (or: culture matters more than codes). J. Bus.
prevent future criminal offences? J. Risk Manag. Financ. Inst. 7, 192–201. Ethics 45, 243–256.
John, L.K., Loewenstein, G., Prelec, D., 2012. Measuring the prevalence of questionable Smith, J., 1994. Gift authorship: a poisoned chalice? Br. Med. J. 309, 1456–1457.
research practices with incentives for truth telling. Psychol. Sci. 23, 524–532. Steneck, N., 2006. Fostering integrity in research: definition, current knowledge, and
Johnson, D.R., Ecklund, E.H., 2016. Ethical ambiguity in science. Sci. Eng. Ethics 22 (4), future directions. Sci. Eng. Ethics 12, 53–74.
989–1005. Stern, S., Lemmens, T., 2011. Legal remedies for medical ghostwriting: imposing fraud
Kerr, S., 1975. On the folly of rewarding A while hoping for B. Acad. Manag. J. 18, liability on guest authors of ghostwritten articles. PLoS Med. 8, e1001070.
769–783. Stern, V., 2017. International Facts and Figures, Universities UK International (down-
Kerr, N.L., 1998. HARKing: hypothesizing after the results are known. Pers. Soc. Psychol. loaded from http://www.universitiesuk.ac.uk/policy-and-analysis/reports/
Rev. 2, 196–217. Documents/International/International_Facts_and_Figures_2017.pdf. (Accessed 19
Kidwell, L.A., Kidwell, R.E., 2008. Do the numbers add up to different views? Perceptions October 2017).
of ethical faculty behavior among faculty in quantitative versus qualitative dis- Swanson, D.L., Frederick, W.C., 2003. Are business schools silent partners in corporate
ciplines. J. Bus. Ethics 78, 141–151. crime? J. Corp. Citizenship 9 (Spring), 24–27.
Klein, H.A., Levenburg, N.M., McKendall, M., Mothersell, W., 2006. Cheating during the Tahir, T., 2008, Dean dismissed for plagiarism, The Times Higher (6 March 2008) down-
college years: how do business school students compare? J. Bus. Ethics 72, 197–206. loaded from https://www.timeshighereducation.com/news/dean-dismissed-for-
Krawiec, K.D., 2005. Organizational misconduct: beyond the principal-agent model. plagiarism/400950.article. (Accessed 13 November 2017).
Florida State Univ. Law Rev. 32, 571–615. Tang, T.L.-P., Liu, H., 2012. Love of money and unethical behaviour intention: does an
Langford, C., Hall, J., Josty, P., Matos, S., Jacobson, A., 2006. Indicators and outcomes of authentic supervisor’s personal integrity and character (ASPIRE) make a difference?
Canadian university research: proxies becoming goals? Res. Policy 35, 1586–1598. J. Bus. Ethics 107, 295–312.
Lawrence, P.A., 2008. Lost in publication: how measurement harms science. Ethics Sci. Tilburg University, 2012, Flawed Science: the Fraudulent Research Practices of Social
Environ. Polit. 8, 9–11. Psychologist Diederik Stapel, Final Report of the Levelt Committee, Noort Committee
Lee, Y., 2011. Understanding anti-plagiarism software adoption: an extended protection and Drenth Committee downloaded from https://www.tilburguniversity.edu/
motivation theory perspective. Decis. Support Syst. 50, 361–369. upload/3ff904d7-547b-40ae-85fe-bea38e05a34a_Final%20report%20Flawed
Lewellyn, K.B., Judge, W.Q., Smith, A., 2017. Exploring the questionable academic %20Science.pdf. (Accessed 13 November 2017).
practice of conference paper double dipping. Acad. Manag. Learn. Educ. 16, 217–236. UMKC, 2015, Investigation into Allegations Raised in Kansas City Star Article UMKC’s
March, J.G., Simon, H., 1958. Organizations. John Wiley, New York. Misleading March to the Top, Report by PricewaterhouseCoopers to University of
Martin, B.R., 2007. Keeping plagiarism at bay a salutary tale. Res. Policy 36, 905–911. Missouri Kansas City (UMKC) (downloaded from https://www.umkc.edu/
Martin, B.R., 2012. Does peer review work as a self-policing mechanism in preventing facultysenate/currentminutes/FSMinutes/2015-2016%20minutes/120115/
misconduct: a case study of a serial plagiarist. In: Mayer, T., Steneck, N. (Eds.), PWCreport.pdf. (Accessed 13 November 2017).
Promoting Research Integrity in a Global Environment. World Scientific Publishing/ Umphress, E.E., Bingham, J.B., Mitchell, M.S., 2010. ‘Unethical behaviour in the name of
Imperial College Press, London and Singapore, pp. 97–114. the company: the moderating effect of organizational identification and positive re-
Martin, B.R., 2013. Whither research integrity? Plagiarism, self-plagiarism and coercive ciprocity beliefs on unethical pro-organizational behavior’. J. Appl. Psychol. 95,
citation in an age of research assessment. Res. Policy 42, 1005–1014. 769–780.
Martin, B.R., 2016. Ethics and integrity in publishing. In: Clark, T., Wright, M., Ketchen, University of Zurich, 2011, Allegations of Publications Misconduct against Bruno S. Frey
D.J. (Eds.), How to Get Published in the Best Management Journals. Edward Elgar, et al., Report of the Ad-hoc Commission (downloaded from https://docs.google.com/
Cheltenham, pp. 29–48. file/d/1GeaNAFcKZAEPJp69F_Dc3Res6VwP5vtKys_e3si2ZeMfWF3iAopIZ3wvoDSY/
McCabe, D.L., Butterfield, K.D., Trevino, L.K., 2006. Academic dishonesty in graduate edit?pli=1. (Accessed 13 November 2017).
business programs: prevalence, causes and proposed action. Acad. Manag. Learn. Van Noorden, R., 2011. The trouble with retractions. Nature 478, 26–28.
Educ. 5, 294–305. Vancouver Group (International Committee of Medical Journal Editors) (1985),
McCraw, D., 1991. The right to republish libel: neutral reportage and the reasonable Guidelines on authorship, British Medical Journal, 291, 722.
reader. Akron Law Rev. 25, 335–365. Vaughan, D., 1999. The dark side of organizations: mistake, misconduct, and disaster.
Merton, R.K., 1938. Social structure and anomie. Am. Sociol. Rev. 3, 672–682. Annu. Rev. Sociol. 25, 271–305.
Mingers, J., Willmott, H., 2013. Taylorizing business school research: on the one best way Wang, L., Malhotra, D., Murnighan, J.K., 2011. Economics education and greed. Acad.
performative effects of journal ranking lists. Hum. Relat. 66, 1051–1073. Manag. Learn. Educ. 10, 643–660.
Monteduro, F., Hinna, A., Moi, S., 2016. Governance and corruption in the public sector: Yang, P., Tao, L., 2012. Ranking of the world’s top innovation management scholars and
an extended literature review. In: Hinna, A., Gnan, L., Monteduro, F. (Eds.), universities. J. Prod. Innov. Manag. 29, 319–331.
Governance and Performance in Public and Non-Profit Organizations. Emerald, Zimmerman, J.L., 2001, Can American business schools survive?, Bradley Policy Research
Bingley, pp. 31–52. Center, Financial Research and Policy Working Paper No. FR 01-16 downloaded from
Morgeson, F., Nahrgang, J., 2008. Same as it ever was: recognizing stability in th http://papers.ssrn.com/abstract=283112. (Accessed 13 November 2017).

14