Research misconduct complaints and institutional logics: The case of Hans Eysenck and the British Psychological Society.

A formal complaint was lodged with the British Psychological Society in 1995 that alleged serious scientific misconduct by Hans J Eysenck. The complaint referred to research into the links between personality traits and the causes, prevention and treatment of cancer and heart disease. Using a framework of institutional logics, we criticise the Society's decision not to hear this complaint at a full disciplinary hearing. We urge the BPS to investigate this complaint afresh. We also support calls for the establishment of an independent National Research Integrity Ombudsperson to deal more effectively with allegations of research misconduct.

Hans Eysenck (1916–1997) has been described as ‘one of the most famous and productive psychologists of the 20th century . . . [and as] Britain’s most prolific writer and researcher in psychology’ (Andersen et al., 2020). By the time of his death, he was the most frequently cited living psychologist and the third most cited of all time, coming after Sigmund Freud and Jean Piaget (Haggbloom et al., 2002). Despite this reputation, in May 2019 an inquiry by King’s College London (2019) concluded that the reported results in 26 of his articles were ‘unsafe’ (p. 2).

The inquiry followed publication of an open letter to the Principal of King’s College London by the editor of Journal of Health Psychology (Marks, 2019). The 26 papers flagged as ‘unsafe’ emerged from research carried out in collaboration with Ronald Grossarth-Maticek, a physician and social scientist based in Heidelberg. This research examined the association between personality factors and the causation, prevention and treatment of fatal diseases. The King’s College inquiry left much to be desired because it ignored many publications that emerged from the same disputed research programme (Everall, 2019; Hawkes, 2019; Marks, 2019; Marks and Buchanan, 2020). One of Eysenck’s biographers, Rod Buchanan, believes that 87 publications in the area under challenge should be retracted from the scientific literature (Marks and Buchanan, 2020).

This has been a long running saga. In the early 1990’s, numerous psychologists, statisticians, epidemiologists and doctors raised serious concerns about the research of Eysenck and Grossarth-Maticek (Fox, 1991; Pelosi and Appleby, 1992, 1993), but with little impact (Smith, 2019). In 1995, Anthony Pelosi, a Scottish doctor, submitted a written complaint detailing his concerns to the British Psychological Society (henceforth, BPS) (Pelosi, 2019).

In this paper, we use an institutional logics approach to review how the BPS dealt with Pelosi’s complaint. We draw attention to pertinent rationales (or ‘logics’) that are likely to have motivated action or inaction by the BPS in dealing with the accusations it received against one of its prominent members. Analytical approaches of this type, based on institutional logics, have been used widely in sociology, management studies and related fields for about 40 years to explain how broad belief systems influence the behaviour of institutions.[1] Our discussion of competing logics may help in developing an appreciation of why an institution such as the BPS behaved as it did.[2] We also highlight the need to improve complaint-processing protocols of professional bodies and scholarly academies. Additionally, we briefly consider the response of journal editors to the recommendations of the King’s College London inquiry.

The following discussion has wider significance than the individual case under consideration. There is mounting concern among psychologists that their discipline is experiencing a crisis (Chambers, 2017; Hughes, 2018; Ritchie, 2020). As with other academic disciplines, the research literature in psychology has been subject to a growing number of retractions, including 57 sole authored or co-authored journal articles of a Dutch social psychologist, Diederik Stapel (Craig et al., 2020). There have been embarrassing failures to replicate findings of major studies in psychology (Open Science Collaboration, 2015). As well, the discipline has witnessed rapid growth in a variety of Questionable Research Practices (Agnoli et al., 2017; John et al., 2012) including the improper use of statistical significance tests (Leggett et al., 2013; Masicampo and Lalande, 2012).

These problems come at a time when universities and scholarly journals have exhibited a persistent and widespread reluctance to investigate research misconduct properly (Tourish, 2019). This brings into sharp relief the remit of professional associations to maintain the integrity and public image of their discipline. The present paper documents and criticises the response of one such association, the BPS, when it was confronted with detailed allegations of serious research misconduct by one of its most prominent members.

As noted, we frame our analysis using an institutional logics approach. We then detail the complaint against Hans Eysenck and follow this by discussing the BPS’s response. Thereafter, we consider the practical implications for how research misconduct should be investigated. We conclude by offering several recommendations that are intended to help the academic community respond more effectively to allegations of research misconduct.

An institutional logics perspective

The term ‘institutional logics’ explains contradictory practices and beliefs that are inherent in the major institutions of modern society: family, community, religion, state, market, profession and corporation (Friedland and Alford, 1991; Thornton et al., 2012). Institutional logics can be viewed as

. . .socially constructed, historical patterns of material practices, assumptions, values, beliefs and rules by which individuals produce and reproduce their material subsistence, organize time and space, and provide meaning to their social reality (Thornton and Ocasio, 1999: 804).

Institutional logics are embedded in practices that are sustained and reproduced by cultural assumptions and political struggles. Such embedding shapes cognition, guides decision-making and helps organisational actors focus on a limited set of issues, problems and solutions that are consistent with a prevailing logic (Jackall, 1988; Thornton, 2002).

We suggest that professional associations face a conflict between a logic that prioritises preserving the integrity of a disciplinary field, and another logic that emphasises the importance of preserving the reputation of that field. While these logics are not mutually exclusive, there is often a tension between them. Some professional associations seem to favour a logic of preserving a field’s integrity to avoid losing legitimacy. Yet, steps to preserve a field’s integrity can often involve unwelcome publicity and imperil legitimacy. This risk is particularly acute if an investigation draws attention to dubious practices on the part of pioneering figures who have a prominent public profile. We make a case that those responsible for adjudicating these tensions are likely to feel bonded to the pioneering figures in the field. This bonding arises through common work interests, shared professional networks, and a feeling that preserving the reputation of a field requires protecting the reputation of the individuals who have played a central role in its development.

Drawing from Berggren and Karabag (2019), we frame our analysis by applying the following three institutional logics to the institutional field of medical science.

Classical medical logic

This ‘focuses on patient care and is embodied in the Hippocratic Oath (“First do no harm”) which requires physicians to uphold a basic ethical standard and to act accordingly’ (Berggren and Karabag, 2019: 429). Whilst a medical logic is most relevant to physicians, it also encompasses the activities of researchers who address issues of physical and mental health. In addition, this logic applies to the professional associations who have a role in legitimising the activities of clinicians and biomedical researchers.

Academic logic

This rests ‘on the intrinsic value of new knowledge and on the principles used to obtain this knowledge’ (p. 429). Academic logic embodies the view that ‘truth-claims, whatever their source, are to be subjected to pre-established impersonal criteria’ (Merton, 1973: 270) and that ‘truth-claims’ should prevail over the reputation of any individual, however distinguished.

Market-oriented logic

This prioritises ‘external performance and indicators’ such as publication scores, funding success and ‘brand and image’ (p. 429). It encourages a cynical cost benefit calculus. In light of integrity/reputation issues, journals, professional associations and universities may tend to act only when the likely reputational cost/benefit of doing nothing outweighs the likely reputational cost/benefit of conducting some sort of inquiry or review. As more parts of the academy become corporatised, commodified, ranked, audited and listed in numerous league tables, the influence of this logic has grown throughout universities (Huzzard et al., 2017).

In the following sections, we highlight the tension between preserving the integrity of the field of psychology and preserving the reputation of that field. We draw particular attention to the subsidiary tension between an academic logic and a market-oriented logic.

The roles that various professional bodies, scholarly academies, universities and publishers should play in dealing with research misconduct are unclear. Each of these bodies seems to assume that the other should act, intends to act, or will eventually act. This leads to a situation of ‘diffused responsibility’ that encourages professional societies to ‘pass the buck’ rather than engage willingly and substantively with the problems concerned. The failure of professional bodies to act discourages the will of others to intervene even more. Such failure may be compounded by adherence to a market-oriented institutional logic. The latter logic will often conflict with pressures to confront allegations of research misconduct that arise from the exercise of classical medical logic and/or an academic logic.

We now explore the institutional logics in the BPS’s processing of the complaint against Eysenck, including whether the Society’s response was motivated by concerns for integrity or for reputation.

The complaint

On 9 July 1995 Dr Pelosi wrote to the BPS making a formal complaint about the conduct of one of its members, Professor Eysenck. The complaint drew the BPS’s attention to peer-reviewed publications about Eysenck’s work on personality and fatal diseases and to the assertion by Bernard Fox, the then leading authority in biopsychosocial cancer epidemiology, that Eysenck’s and Grossarth-Maticek’s claims were ‘simply unbelievable’ (Fox, 1988). Enclosed with the complaint was a copy of two critical articles in the British Medical Journal (Pelosi and Appleby, 1992, 1993) and Eysenck’s reply to the first of these (Eysenck, 1992).

Eysenck and Grossarth-Maticek had reported hitherto unseen strengths of association between particular personality types and the development of cancer and ischaemic heart disease. For example, they concluded that people who were hypothesised to have a ‘cancer-prone’ personality type were 121 times more likely to die from a cancer compared with those hypothesised to have a healthy personality. This relative risk of 121 is ‘. . .perhaps the highest ever identified in non-infectious disease epidemiology’ (Pelosi and Appleby, 1992). Eysenck and Grossarth-Maticek also claimed that their method of psychological treatment resulted in massive reductions in all-cause mortality over the ensuing decade (Eysenck and Grossarth-Maticek, 1991; Grossarth-Maticek and Eysenck, 1991).

Pelosi and Appleby (1992, 1993) raised numerous concerns about how such results could have been obtained. They questioned how Grossarth-Maticek could have had the time personally to carry out the psychotherapy within large randomised trials while also conducting cohort studies involving approximately 30,000 participants. They challenged the reporting of no classification error in the complicated assessments of personality that were conducted by ‘more than 100’ student interviewers (Lee, 1991). In a sentence that had been carefully re-worded by the British Medical Journal’s solicitors, they stated that:

. . . .one is left to speculate whether the authors have made the mistake, during reanalyses of their data, of reassigning individuals to personality types after causes of death were known (Pelosi and Appleby, 1992: 1297).

Other scientists had raised similar concerns, sometimes in much starker terms. Van Der Ploeg (1991, 1992), for example, reported that there had been unequivocal manipulation of data sheets in parts of this research programme.

Pelosi’s complaint brought an even more serious concern to the BPS’s attention. Grossarth-Maticek et al. (1991) described in detail how a subgroup of 41 people with malignant hypertension had been included in their randomised trial of individual psychotherapy. These research subjects were described as ‘stressed but healthy’. However, their reported clinical features clearly indicated that they were at imminent risk of stroke, heart failure and kidney failure and that they needed urgent hospital treatment. This subgroup had an extremely high mortality on follow-up. It is difficult to imagine a more unethical clinical experiment (Pelosi, 2019; Pelosi and Appleby, 1993).

Pelosi’s complaint was considered at meetings of the BPS’s Investigatory Committee on 24 July 1995 and 30 August 1995. A letter notifying dismissal of the complaint was sent to Pelosi on 14 September 1995. Before we examine the processing of this complaint, it is important to be aware of a prior controversy in which the BPS had been embroiled, and which may have influenced its response to Pelosi’s complaint.

The British Psychological Society’s investigation in context

In mid-1995, when the complaint against Eysenck was received, the Society was living in the shadow of the reputational odium it suffered as a result of the controversial claims that prominent psychologist and BPS member, Sir Cyril Burt (1883–1971), had engaged in fraudulent research. If the Society had openly investigated the integrity of Eysenck’s research this could have reopened the Burt controversy and the schisms it had caused within the psychology community.

For several decades, Burt had been ‘a commanding figure in British psychology; productive, influential and publicly honoured’ (Connolly, 1980: i). He had published many ground-breaking studies. For example, his studies of separately reared twins led him to the controversial conclusion that hereditary factors were the main cause of differences in intelligence tests between children with working class backgrounds and those with upper class backgrounds. However, several years after his death, evidence emerged that some of the data he relied on in these studies were fake.[3]

In 1980, the BPS published the highly critical results of a symposium it convened on Burt (Beloff, 1980). Writing on behalf of the Society’s Council, its then President, Kevin Connolly, stated that ‘. . .we now sadly accept that Sir Cyril was guilty of violating a fundamental cannon of science. . .the presumption of honesty’ and that ‘there now seems no reasonable doubt that Sir Cyril Burt perpetrated fraud in that he fabricated data. . .’ (Connolly, 1980: i). An academic institutional logic clearly dominated these proceedings. In 1980, the Society undertook to produce a list of Burt’s unreliable publications but it never did so (Samelson, 1997). In 1992, after a campaign to have the issue re-opened, the BPS decided that it would no longer take a position on alleged misconduct by deceased members (Tucker, 1997). The Society withdrew its judgement on Burt. This pleased some but it infuriated others (Morris, 2019).

Given this divisive and infamous history, we contend that key figures within the Society may have been reluctant publicly to adjudicate on allegations of fraudulent and unethical research behaviour by another of its most prominent members.[4] If the Society had investigated the allegations against Eysenck, it would most likely have been subject to strident demands from opponents and supporters of Burt. For example, some would have been eager to know why the Society had stepped back from taking a position on Burt but was now willing to consider Eysenck’s case. An investigation of Eysenck could therefore have become a proxy argument for the controversial research agenda that Burt (and Eysenck) had championed. There was a clear tension, once more, between preserving the integrity of the field (a task which would be rendered difficult if its two most eminent members were disgraced) and preserving its reputation. We suggest that the actions of the BPS were partially motivated by desire to avoid this tension, and to pass the responsibility to others.

The BPS’s investigation of Pelosi’s complaint

Our understanding of the processing of the complaint has been developed from archival records held by the Society and by the complainant. The Society provided access to minutes of relevant parts of the two meetings of its Investigatory Committee at which the complaint was considered. The Society’s archivist informed us that ‘The BPS does not hold any other papers and correspondence in relation to the work of the Investigatory Committee’ and that we had been supplied with ‘all the information that [the BPS] have found in the archives relating to [the consideration of the complaint]’.[5] We also drew upon secondary library resources, including biographies and a large literature in scholarly journals that has assessed Eysenck’s contribution to psychology.

Some of the statements we make emerge from our interpretation of the evidence available to us. Such an interpretative approach ‘can produce many useful insights [and]. . . lead to contestable conclusions because of. . . .a “plurality of plausible explanations”’ (Ron, 2008: 291). We do not assert that the explanations and interpretations entered below are necessarily better than competing ones.

Statute 14(4) of The Royal Charter, The Statutes and The Rules, British Psychological Society (approved September 1994) governed the way in which complaints were dealt with. A three stage process was envisaged. In the first stage:

The Investigatory Committee shall consider all allegations bought to its attention. Normally the member who is the subject of the allegation shall be invited to make such written observations on the allegation as he or she deems appropriate. The Member shall be warned that such observations would be taken into account at any hearing to consider whether or not he or she is guilty of professional misconduct.

Consistent with this first stage procedure, the complaint was considered by the Investigatory Committee on 24 July 1995.

The second stage of the complaint consideration process is explained in Statute 14(4) as follows:

Having considered any observations that the subject of the allegation has made and taken such additional steps as it deems necessary to decide whether further investigation is justified, the Investigatory Committee shall either appoint an Investigatory Panel to undertake further investigation of the allegation or recommend to the Chairman of the Disciplinary Board [always a non-psychologist[6]], or a non-psychologist member of the Board nominated by the Chairman, that further investigation is not justified.

The actions envisaged in this second stage were completed by the Investigatory Committee at its meeting on 30 August 1995.

Depending on the outcome of the second stage processing, the third stage generally involved one of two outcomes: either setting up an Investigatory Panel which could lead to a Disciplinary Committee hearing or having a non-psychologist member of the Disciplinary Board agree that further investigation was not required.

The use of two adjudicative bodies (an Investigatory Committee referring complaints to a Disciplinary Board) is common in disciplinary processes in professions. This raises questions regarding the role of an Investigatory Committee in such circumstances. Should it be viewed principally as a mechanism to protect a profession and its members from unmeritorious complaints? Or should it be viewed principally as a screening mechanism to determine ‘whether there is sufficient evidence to warrant referral to’ a Disciplinary Board (Singleton, 2009/2010)? This question was considered in an Ontario Divisional Court hearing of a professional misconduct case against a teacher.[7] The court confirmed the view that the Investigatory Committee should perform ‘a limited screening role’: that is, determine whether there is sufficient evidence to warrant referral to a disciplinary board. In our view, such a role was not performed by the BPS’s Investigatory Committee, despite there being ‘sufficient evidence’ to warrant a referral. The actions of the Investigatory Committee in dismissing the complaint were consistent with a desire to protect the reputation of the Society and its member by not deeming the complaint meritorious. A market logic seemed to prevail.

In her discussion of medical ethics and professional conduct, Beloff (2003) refers specifically to how the BPS investigated complaints. She draws attention to the society’s obligation to protect the public ‘from incompetence, misconduct, fraud and to uphold the good name of the profession’ (p. 13). She saw [as we do] an unsatisfactory ‘duality’ because ‘at the moment the society is both the members’ association and yet the advocate for the public’ (p. 13). In our view, such a ‘duality’ is inherently inclined to subordinate academic and medical institutional logics to a market-based logic. The way the complaint was processed is consistent with such a view because the committee abrogated its ‘obligation to the public’. Beloff argued that the role of the Investigatory Committee is to determine ‘whether the complaint is serious and informative enough to be taken further’ (p. 13). This is consistent with the decision of the Ontario Divisional Court, discussed earlier. In our view, the complaint was both serious and informative and should have been considered further.

The decision of the Society’s Investigatory Committee not to recommend that the complaint be heard at a disciplinary hearing can be explained by the nature of those disciplinary hearings at the time; and by the Society’s market-oriented sensitivity, especially in the wake of the Burt controversy.

Disciplinary hearings were announced publicly. They were held in public session, often lasting several days, and usually the parties were represented by legal counsel (Beloff, 2003: 14). The hearings were conducted this way ‘to maintain public confidence’ (p. 14). The Clerk of the Investigatory Committee was required to ensure proceedings were ‘entirely logically “independent and impartial”’ (p. 15). Of note is Beloff’s (2003) biting final sentence in which she claims ‘there are some psychologists, of good faith and probity, who still think that these formal registers are against civil liberties and that it is all about creating a closed shop’ (p. 15).

Meeting of the investigatory committee on 24 July 1995

Those present at this meeting were:

Professor Steve Newstead – (Chair) President of the BPS, past Chair of Special Group of Teachers in Psychology [now Division of Academics, Researchers, and Teachers in Psychology]

Anne Richardson – Course director for the Diploma of Clinical Psychology at University College London – later Head of Mental Health, Department of Health

Gerry Mulhern – Honorary General Secretary of the BPS, Lecturer (later Professor) in Psychology Queen University Belfast, later President of the BPS

Margaret McAllister – Educational Psychologist, later President of the BPS

Clerk: Graham Geldart (BPS Assistant Executive Secretary, qualified solicitor).[8]

Aspects of the profiles of two of the committee members invite confidence that the complaint would have been assessed fully and fairly. The Chair, Professor Newstead lists a research interest in academic dishonesty.[9] Margaret McAllister (2011) describes how her involvement in the BPS led to her gaining ‘an excellent grounding in effective committee work, for example, considering all sides of an argument, achieving consensus and getting things done’ (see also Henderson, 1996).

The minutes of the meeting record the initial processing of the complaint as follows:

Noted 1: Letter of complaint from Dr Anthony Pelosi dated 9th July 1995 and attendant documentation.

Noted 2: The Clerk should seek Hans Eysenck’s comments as he would in the normal course requesting him to address the ethical issues raised in the letter of complaint relating to the research which was the subject of these complaint matters.

Meeting of 30 August 1995

Eysenck’s response [which was required to be in writing] dated 16 August 1995 was considered at this meeting. The BPS has been unable to find any record of Eysenck’s response. The same members who attended the previous meeting of the committee were again in attendance. The extract from the minutes of the meeting that dealt with the processing of the complaint is reproduced below.

Noted 1: Letter from the Clerk to Hans Eysenck dated 2nd August 1995 seeking comments and written submissions on the complaint; and a reply from Hans Eysenck dated 16th August 1995.

Noted 2: The only real issue of possible professional misconduct in relation to these allegations were those of a contamination of the research sample and in this regard the subject of the complaint had pointed out that this was not undertaken by him, but by a third party.

Agreed: There were no elements of professional misconduct to answer in this matter and a recommendation would be put to the nominated representative of the Disciplinary Board that it would not be appropriate to appoint an Investigatory Panel to make further enquiry.

The minutes suggest that assessment of the complaint and the issues it raised was superficial. The previous meeting of the committee asked Eysenck to ‘address the ethical issues raised’. However, there is no record in the minutes of any response to the ‘ethical issues’, although there may well have been. The minutes record that the only possible issue was ‘contamination of the research sample’. Several aspects of the matters mentioned as ‘Noted 2’ in the minutes warrant comment.

‘Real issues’: This implies that some matters raised in the complaint were relevant and within the jurisdictional ambit of the committee (that is, ‘real’) and others were possibly outside the committee’s jurisdictional ambit and were ‘not relevant’ (that is, ‘not real’). We return to this point later.

‘Professional misconduct’: Is there a difference implied between professional misconduct and other misconduct? It is plausible that the committee confined itself to investigating ‘professional [conceived as the behaviour of a practising psychologist] misconduct’. Perhaps professional misconduct was perceived as a ‘real issue’ and other than professional misconduct was perceived as a ‘non-real issue’. This would have been an easy course of action for the committee because ‘professional misconduct’ was not defined directly and unambiguously by the BPS at the time. Even in the Code of Ethics and Conduct published by the Society’s Ethics Committee in 2009, and updated in 2018, ‘professional misconduct’ is confined largely to behaviour that involved harming clients or behaving in a way that brought the Society or the reputation of the profession into disrepute (e.g. by a member being found guilty of a criminal offence). So it is plausible that the Investigatory Committee applied a practitioner-oriented view of ‘professional misconduct’ to side-step some of the research-related issues in the complaint.

‘Contamination of the research sample’: The complaint drew attention to the unethical inclusion in a clinical trial of gravely ill people with malignant hypertension that had been identified by the researchers. The language used in the minutes shifts the focus of the complaint deftly. In referring to ‘contamination of the research sample’, Eysenck may have been accepting that some participants were inappropriately included in this trial but he then distanced himself, strategically, from Grossarth-Maticek by denying any responsibility for their recruitment to the research programme. This would be consistent with the tactics that, according to one of his biographers (Buchanan, 2010: 391), Eysenck often used to respond to the ‘fearsome assault’ on his work.

Eysenck is the lead author or sole author of more than 60 articles and book chapters on the research he conducted with Grossarth-Maticek. In reply to criticisms of it, Eysenck stated:

There were many aspects of the work about which I could only say that I would not have done it that way, and regretted that I had not been there when decisions about methodology and statistical analysis had been made (Eysenck, 1991: 298).

This is a further instance of diffused responsibility, in this case by Eysenck, as he attempted to shift responsibility for problems onto Grossarth-Maticek. The BPS’s apparent willingness to accept this is puzzling. Eysenck’s collaboration with Grossarth-Maticek was very close (Andersen et al., 2020; Buchanan, 2010) and it would be strange had he not been more aware of the authenticity of their data and the rigour of its analysis. Even if we grant ignorance on these points, many observers would consider it a form of misconduct for a lead (or any) author not to be a party to such decisions, at least to some extent. By distancing himself from the methodological and the ethical problems complained about, Eysenck ignores his personal agency in condoning such actions (implicitly or explicitly) as a co-author and sometimes sole author of articles describing this research. It is difficult to understand why the BPS would consider that they did not have any role in doing something about the publication and dissemination of unethical research by one of its members.

The recommendation of the Investigatory Committee that ‘there were no elements of professional misconduct to answer’ and that it would not be appropriate to appoint an Investigatory Panel, was conveyed by the Clerk to the Chair (or the Chair’s nominated representative). This person was required to be a non-psychologist member of the Disciplinary Board. Thus, one of the first six members of the Disciplinary Board listed below would have been approached:

a. David Richards (Chair) (Deputy Chairman of Monopolies and Mergers Commission)

b. Colin Prestige (Law Society recommen-dation)

c. Louis Kramer (General Dental Council)

d. Margaret Packham (General Optical Council)

e. Susan Ritter (Institute of Psychiatry)

f. David Lindsay (Institute of Chartered Accountants)

g. David Legge (psychologist)

h. Lea Pearson (psychologist)

i. Tony Chapman (psychologist)

j. Halla Beloff (psychologist)

k. Tony Gale (psychologist).[10]

The professional backgrounds of the non-psychologist members were in law, dentistry, optometry, nursing and accountancy. The BPS informed us that they do not hold information that might enable them to identify the non-psychologist approached.[11] In any event, the precise identity of the person involved does not matter greatly. The BPS may have wished to give non-psychologists major roles in this process to prevent accusations that it was merely ‘looking after its own’. But this approach has unintended consequences. In particular, all six non-psychologists were unlikely to be equipped, by way of experience and knowledge, to comprehensively assess the epidemiological, statistical, clinical and ethical issues raised in the complaint.

On 14 September 1995 the Clerk, Graeme Geldart, wrote to Pelosi on behalf of the Society, as follows:

After full consideration of all the material before it, The Investigatory Committee decided that it would not be appropriate to appoint an Investigatory Panel to conduct further enquiries into the matter. Its decision has been confirmed by the independent non-psychologist representative of the Disciplinary Board.

The Investigatory Committee sought comment from all relevant parties on the matters of complaint raised by you and, having considered the matter carefully, and with the benefit of all the documentation before it, concluded that Professor Eysenck’s conduct was not such as to amount to misconduct, and an Investigatory Panel was not therefore appointed.

The Committee has asked me to assure you that it is confident that its purposes have been properly and satisfactorily served in bringing this matter to the attention of the subject of the allegations and trusts that you accept its position. The matter is now closed as regards the Society.

The wording of this letter raises further queries.

‘Sought comment from all relevant parties’

Is the plural ‘parties’ simply loose wording by the Clerk? Who else other than Eysenck was approached for a written response? The minutes suggest Grossarth-Maticek was not approached. Was other unrecorded information sought and considered? If so, from whom was this information sought? It seems unlikely that the committee contacted the editors of the journals in which the research at issue was published.

‘Considered the matter carefully’

The matters raised in the complaint required understanding of methodological and clinical issues. However, the terseness of the minutes suggests that these matters received only cursory attention.

Commentary

The processing of the complaint highlights several inadequacies in the Society’s procedures at the time. The complainant was not given an opportunity to respond to Eysenck’s response. The non-psychologist member of the Disciplinary Board who was consulted was unlikely to have been qualified to assess the numerous research improprieties that had been alleged. The complaint should have been processed more thoroughly and transparently by persons more familiar with the research methods being challenged.

As we explain below, the processing of the complaint was inconsistent with the Society’s own explanation of ‘How the Investigatory Committee Works’, published as ‘Ethics Column No 5’ in The Psychologist in March 2006 (BPS, 2006). We accept that the composition and protocols of the Investigatory Committee might have changed between 1995 and 2006, and that ‘Ethics Column No 5’ could reflect improvements in procedures. If so, then such improvements were well overdue. Several of the protocols mentioned do not appear to have been applied in considering Pelosi’s complaint. Point (i) of the Society’s procedures, in ‘Ethics Column No. 5’ (outlined below), was not allowed. Based on the information available to us, it seems likely that the investigatory protocols outlined in points (ii) to (vi) below were also not followed.

(i) give the complainant ‘an opportunity to comment on the response’ from the psychologist who is the subject of the complaint;

(ii) ask a member of the committee who was knowledgeable in the ‘type of work complained about’ to lead the discussion of the complaint;

(iii) make decisions on ‘the balance of probabilities’;

(iv) be prepared ‘to explain the ways in which [members] made their decision’;

(v) use a ‘sliding scale that takes into account elements such as the seriousness of the allegation, the decision-making processes used, the potential harm, the strength of feeling among committee members and specialist knowledge’; and

(vi) ‘tend towards caution’ in circumstances of doubt by erring on the side of recommending further investigation of the complaint.

The broader corpus of Eysenck’s work

How has the BPS responded to inevitable questions about the reliability and validity of Eysenck’s other work? Andersen et al. (2020) have argued that the papers on personality and fatal diseases ‘. . .stand apart from [Eysenck’s] many other seminal contributions to psychological knowledge’. We do not share this confidence. It strains logic to believe that Eysenck would reserve defective research habits for only one specific domain of inquiry (Smith, 2019). He had strong views about many controversial topics and often drew incendiary conclusions that differed from those of others working in these areas (Eysenck, 1997). Legitimate questions should be asked about the evidence base on which he relied. As one of Eysenck’s biographers noted: ‘Many of his publications, his books especially, tended to skim over methodological and procedural details. His documentation and referencing habits left much to be desired’ (Buchanan, 2010: 167). Eysenck also relied heavily on data obtained from many postgraduate students and research assistants. The testimony of some of them, cited by Buchanan (2010), suggests that he was often more interested in obtaining spectacular results that confirmed his theories than in ascertaining the reliability and validity of the data he relied upon.

Eysenck (1994: 126) provided insight into his thinking on these issues when writing about the Burt affair:

Scientists have extremely high motivation to succeed in discovering the truth; their finest and most original discoveries are rejected by the vulgar mediocrities filling the ranks of orthodoxy. They are convinced that they have found the right answer. . .The figures do not quite fit, so why not fudge them a little bit to confound the infidels and unbelievers? Usually the genius is right, of course (if he were not, we should not regard him as a genius), and we may in retrospect excuse his childish games, but clearly this cannot be regarded as license for non-geniuses who foist their absurd beliefs on us.

This comes perilously close to justifying research misconduct on the part of ‘geniuses’. Eysenck certainly seems to have regarded himself as some sort of genius. He was encouraged in this belief by younger scientists whom he influenced (Nyborg, 1997; Saklofske, 1998).

Colman et al. (2019) have called on the BPS to conduct a ‘formal investigation and audit’ into Eysenck’s wider publications. While it may be impractical to audit the entirety of his voluminous output, there are sufficient grounds for supporting a targeted audit of articles where there is good cause to doubt the veracity, integrity or validity of the evidence presented and the conclusions drawn.

The BPS’s response to such suggestions has been to ‘pass the buck’ and to claim that although it reaffirmed the importance of research integrity, ‘the conduct of research lies with the academic institution which oversees the work carried out by its academics and we welcomed the investigation into this research carried out by King’s College, London’ (BPS, 2019). This seems to be a further indication that, so far as the BPS is concerned, classical medical logic and academic institutional logic remains subordinated to a market-oriented institutional logic. An unfortunate implication of the Society’s stance is that a large number of questionable publications will remain in the scientific literature with no obvious remedy in sight. This seems incompatible with a serious intent to act in the best interests of public health and broader society and to address psychology’s credibility crisis.

How journal editors have responded – further evidence of diffused responsibility

The King’s College London report was completed in May 2019. In September of that year the editors of 11 journals were contacted by King’s College’s Director of Research Governance, Ethics and Integrity. So far, three journals (International Journal of Sport Psychology, Psychological Reports and Perceptual and Motor Skills) have collectively retracted 14 papers, including three not highlighted by the inquiry. Two of these journals (Perceptual and Motor Skills, 2020; Psychological Reports, 2020) have also issued ‘expressions of concern’ in regard to 61 papers that were authored or co-authored by Eysenck on topics other than personality and fatal diseases. Two other journals, International Journal of Social Psychiatry (2020) and Journal of the Royal Society of Medicine (2020) that have never published work by Eysenck on personality and fatal diseases have recently made ‘expressions of concern’ about seven articles that were written by him on other subjects.

Seven journals have yet to respond: Behaviour Research and Therapy (founded in 1963 by Eysenck), Journal of Social Political and Economic Studies, Integrative Physiological and Behavioral Science, Neuropsychobiology, Journal of Behavioral Therapy and Experi-mental Psychiatry, International Journal of Stress Management, and Journal of Clinical Psychology. This may indicate that journal editors and publishers are content to prevaricate when faced with allegations of research misconduct. If journals fear reputational damage from publishing retractions, it suggests that a market-oriented institutional logic has taken precedence over classical medical logic and academic logic. That is, the journals concerned are prone to prioritise public relations concerns over concerns for scholarly integrity and for public health.

We have found it helpful to use an institutional logics approach when examining the response of Personality and Individual Differences to the King’s College London inquiry. This journal was founded by Eysenck in 1978 (Eysenck, 1988). He was the first editor and his wife, Sybil Eysenck, was the second editor. It is the official journal of the International Society for the Study of Individual Differences that was co-founded by Eysenck. Personality and Individual Differences did not accept the recommendation from King’s College that three articles co-authored by Eysenck and Grossarth-Maticek should be retracted from the scientific literature. Instead, the journal made ‘expressions of concern’ and stated:

It would simply be unwise to rely upon the veracity of the reported results until either evidence appears of deliberate intention to deceive (which would result in immediate retraction), or the results are replicated by an independent group of researchers (Saklofske et al., 2020).

This is an extraordinary decision. Do the signatories (the editor and the senior associate editors) believe that the reported findings could actually be credible? Have they read Eysenck’s and Grossarth-Maticek’s publications that carefully describe unethical experimentation on people who were dangerously ill due to malignant hypertension? Would the journal have reached such a verdict in the case of someone who was not its founding editor? The current editor and at least three co-signatories had been Eysenck’s postgraduate students and/or junior collaborators. Has this association affected their judgement?

Personality and Individual Differences seems to suggest that only a ‘deliberate intention to deceive’ should result in retraction. It is impossible to determine precisely the mental state of scientists who produce unbelievable or unethical research. If the approach of these editors was applied more widely, then hardly any published papers – except those whose authors openly admit to fraud – would ever be retracted. Although an academic institutional logic is partly in evidence, it seems to have been subordinated to other concerns.

Diffusion of responsibility is also evident. In making the decision not to retract these articles, the editors ‘were minded of the previous articles, claims/counter-claims and the formal investigation into this matter by the BPS during the early 1990s. The latter investigation declined to proceed with the complaint against Professor Eysenck’. This makes it all the more important that the BPS should carry out a new and more thorough investigation of Pelosi’s complaint than it did in 1995.

Conclusion: The way ahead

There are significant conflicts in the institutional logics that guide professional associations, academic institutions, and scholarly journals. In this case, a major national professional association, appears to have prioritised a market-oriented institutional logic over a classical medical logic or an academic logic. A persistent failure of ‘transparency of process’ is also displayed. This has involved prioritising self-serving behaviour over ethical propriety. Rather than behave in this way, we urge all professional associations to investigate complaints of research misconduct against members in a way that prioritises integrity over reputation.

A clear diffusion of responsibility effect is evident. Many of the actors implicated in this case have been reluctant to discharge their responsibilities – presumably in the hope that someone else would do so instead. Most of the journal editors who have received clear recommendations following a properly-constituted inquiry by King’s College London have yet to take any action. There is also little indication to date that at least 60 other publications by Eysenck, based on the same flawed dataset (Marks and Buchanan, 2020), have been the subject of any serious re-consideration by university authorities, the BPS, or publishers of scholarly journals.

We support the call for all of Eysenck’s publications on the links between personality and fatal diseases to be thoroughly investigated (Marks and Buchanan, 2020). Furthermore, we support calls for other parts of Eysenck’s research output to be audited by appropriate authorities (Colman et al., 2019; Smith, 2019). As a part of this re-investigation and auditing, we call on the BPS to reconsider the substance of Pelosi’s complaint – and to do so transparently, and in accord with best ethical practice.

After the Burt affair, the BPS took the view that it would no longer investigate the work of deceased members. Journal editors and former employers may well feel likewise. This is not an acceptable position. The integrity of the scientific record is at stake. The reliability of research must be assessed independently of whether its authors are alive to mount a defence. When manifest failings have been identified, it is negligent to allow publications to remain in the scientific literature, where they can continue to influence researchers, be included in meta-analyses (Chida et al., 2008; Shields et al., 2020), and undermine public health. This is not primarily a question of establishing guilt or innocence but of determining the integrity of the evidence base on which scientists and the broader academic community can build further work and advise on matters of policy.

Beyond this case, we concur with Marks and Buchanan’s (2020) recommendation that the United Kingdom should create an independent National Research Integrity Ombudsperson. The United States already has an Office of Research Integrity. This acts as a central resource for investigations into research malpractice and it has played an important role in promoting integrity within academia. While the United Kingdom has a Research Integrity Office, this is ‘an independent charity, offering support to the public, researchers and organisations to further good practice in academic, scientific and medical research’.[12] A properly resourced Ombudsperson’s Office could ease the diffusion of responsibility problem and help to restore academic and medical institutional logics to their rightful positions at the forefront of attempts to improve scientific integrity.

The Eysenck case is a stain on the record of psychology and on science itself. It is time for appropriate measures to be taken to ensure that scandals such as this are addressed more effectively in the future. We urge the BPS to play its part, by investigating afresh the formal complaint it received in 1995 concerning the research of Hans Eysenck.