Fair game? Producing and publishing gambling research

Abstract

This brief article reviews the findings of Fair Game (2014) and discusses their implications for journals publishing gambling research. Drawing parallels with critiques in tobacco and alcohol, it adds to the growing number of voices arguing for reform of the gambling field.

Keywords:

• ethics
• research
• funding
• gambling industry
• knowledge

Introduction

Earlier this year, a team of anthropologists from Goldsmiths, University of London, launched Fair Game, a report which uses interviews with stakeholders to investigate how gambling research is produced and used (Cassidy, Loussouarn, & Pisac, 2014). We spoke with members of the industry, treatment providers, regulators and policymakers, as well as with researchers. The majority of the 109 stakeholders (67) were based in the UK, but we also interviewed people in south-east Europe, North America and Oceania. The purpose of Fair Game is to provide a small and suggestive body of evidence which will encourage others to continue to reflect on their own activities and to expose gambling research to critical attention. If gambling studies is to attract and retain talented early career scholars, make contributions to policy, and be taken seriously, it has to become a more reflexive, critical and open field. In this short article I provide a brief overview of the findings of Fair Game before moving onto a critical discussion of the implications of the report for gambling journals.

Methodology

Before we began interviewing for Fair Game we held a focus group of four research users (a treatment provider, a policymaker and two members of the industry) and four research producers (an anthropologist, a historian and two psychologists). Members of this group devised 40 questions to ask during semi-structured interviews. They also provided the names of possible contributors. We began by inviting influencers: researchers who sat on boards, advised committees and produced authoritative, widely cited work; policymakers at the level of national governments; treatment providers working for internationally respected institutions; industry executives at board level. We then looked at early career researchers, those working in less mature jurisdictions including Hong Kong, Macau and south-east Europe, and those who had left the field of gambling research.

Of the 143 people we invited to participate, 109 agreed: 35 members of the gambling industries, 49 researchers, 11 treatment providers, 6 regulators and 8 policymakers. The breakdown by gender, years of experience and jurisdiction can be viewed in Fair Game. Interviews ranged from an hour to many hours over several meetings. All interviews were transcribed and entered into Nvivo, software designed to identify common themes in qualitative data. We chose to structure the report around five themes: problems with gambling, evidence, the field of gambling studies, money and access. We selected quotes from the interviews to illustrate the views that we had collected, giving greater weight to those which were most representative of our participants, but also preserving diversity. So, for example, even though 40 researchers told us that taking funding directly from the industry could be expected to influence the outcome of research, we also included the views of those who believed that doing so had no impact on research. More subtle differences were also preserved where possible. For example, one researcher felt that industry funding would affect the outcome of most research but not his own. The interviews were replete with strategic segues and pragmatic elisions. As might be expected, the theory and practice of gambling research are not identical. On the basis of our findings we made a number of observations and recommendations, many of which have been made before by a variety of people (Adams, 2013; Cosgrave, 2010; Young, 2012). The particular contribution of Fair Game is that it is based on the views of stakeholders and includes their perspectives on the everyday practices of producing gambling research. Our recommendations included the following:

• gambling research should broaden its perspective beyond the narrow confines of problem gambling conceptualized as an individual weakness;

• policymakers should consult a wider range of evidence and pay more attention to the derivation of all evidence;

• gambling studies should have a code of ethics;

• journals should include a wider range of approaches and consult a wider range of referees;

• research should be funded by hypothecated taxes, the proceeds of which should be distributed by research councils; and

• access to data and environments should be part of licensing.

The field of gambling studies

An area of consensus between stakeholder groups was that the standard of research in gambling studies was poor, relative to other fields including tobacco and alcohol research. According to our participants, gambling studies is an insular and uncritical homogenous field which suffers from unproductive repetition and rivalries. These weaknesses are reproduced by funding which rewards conformity and marginalizes critical voices. The impact of this relationship, between what some of our participants referred to as ‘safe’ research and funding, is felt most acutely by early career researchers:

Plenty of times I have been pushed to take up more of a conventional perspective on problem gambling, or measuring or using existing work to rehash ideas that are already out there. There is support in that there is money, even. There is career progression. This is the amazing thing for a new scholar in your field. And discouraging too. It is very hard to do something new. You are discouraged, because to work with people you have to choose someone who has a record of getting money. But if you do that the likelihood is that they are a person who sticks just to problem gambling. They may be completely genuine and their research may be excellent, in those terms, but those terms are not the ones on which I want to work. I want to go beyond that and there is absolutely no chance to do that in gambling studies. (Male researcher who has been working on gambling in Europe, but not the UK, for six years.)

Several related trends strengthen and tighten this cycle. Some are not exclusive to gambling studies including the reduced availability of public funding for research and the marketization of academia, which has increased pressure on academics to engage in lucrative collaborations with the private sector (Thomas, 2012). Others are more distinctive, including investment by the industry in research, treatment and education as an aspect of brand management/lobbying and the oligarchical structure of the field. Financial support and therefore influence is concentrated in the hands of a few individuals who are trusted by the industry, regulators and governments to produce research that is acceptable to all parties. Industry support for large projects by high-profile academics, including the editor-in-chief of this journal, have attracted negative media attention and raised unresolved questions about relationships between researchers and operators (Nicholls, 2014).

A growing body of research outside gambling studies has shown that, for reasons that are poorly understood, financial relationships with industry increase the likelihood of pro-industry findings, biased interpretations of data and under-representation of alternative interpretations (Babor & Miller, 2014). To take one striking example, Barnes and Bero (1998) found that review papers funded by the tobacco industry were 88 times more likely than non-industry-funded studies to conclude that passive smoking is not harmful to health. Systematic biases have also been recorded in industry-funded research into sugar-sweetened beverages (Bes-Rastrollo, Schulze, Ruiz-Canela, Martinez-Gonzalez, & Stuckler, 2013), climate change (Shrader-Frechette, 2011) and pharmaceuticals (Bekelman, Li, & Gross, 2003).

The mechanisms through which these biases are transmitted vary and may be subtle and complex. However, in gambling, some are straightforward. A number of our research participants described a sense of indebtedness to industry funders, whose contributions to research are conceptualised as ‘gifts’:

The research we do has to be of some use to the industry, after all they are paying for it. You can't just choose something completely esoteric – you have to demonstrate that what you produce is going to be of some benefit to them otherwise you won't get funding. They are paying for it after all. I think that's reasonable! (Female researcher based in the UK who has been working on gambling for four years.)

It is not easy to change any field, to ensure that it continues to adapt and develop. However, the torpor in gambling studies is not accidental. Nor does it provoke the anxiety that it should.

Implications for journals

Emphasis on safe research

Specialized gambling journals play a central role in sustaining the field in its current form. The small number of journals is dominated by individuals acting as contributors, editors, board members and peer reviewers. According to our participants this perpetuates a focus on ‘safe’ research, mostly (but not always) conducted by a wide range of psychologists and medical researchers, including psychiatrists, nurses, psychotherapists and biomedical researchers, and focusing on the individual consumption of gambling by people who are identified using screens and questionnaires as ‘problem gamblers’. Relatively little attention is paid to the conditions of production which enable gambling (the political economy of gambling), the history of relationships between operators and the state, the wider public health implications of gambling expansion, or the archaeology and technology of gambling products and places. In this case, the effect of peer review – often portrayed as an assurance of quality – is to stifle innovation (Adams, 2008; Horrobin, 1990, Young, 2013).

The problem is not (just) the under-representation of qualitative research. Both quantitative and qualitative work can be ‘safe’, or critical (in the sense of being analytical, searching and significant), rigorous or unsound. Moreover, the majority of disciplines and sub-fields have moved beyond this artificial opposition to embrace mixed methods which are collaborative, iterative and responsive to the questions at hand and the nature of the data available. With a few exceptions, gambling studies is woefully inadequate in these terms. Funding streams and calls for proposals help to maintain the artificial separation between these approaches. Researchers are able to construct silos from which they defend their approaches and disparage alternatives. Tables of contents show that that publishing is a conservative game: rearrangement and repetition produces an impression of activity at the same time as it undermines any serious attempts to encourage diversity or change.

Follow the money

Journals in the fields of tobacco, alcohol and pharmaceuticals research insist that vested interests are identified (Casswell, 2013). Indeed, the BMJ (formerly the British Medical Journal) ‘will not consider for publication any study that is partly or wholly funded by the tobacco industry’ (Godlee et al., 2013). In gambling studies it is virtually impossible to detect corporate interests, even in research published in peer reviewed journals. Babor and Miller (2014) conducted a survey of 30 randomly selected papers from the National Centre for Responsible Gaming's (NCRG) library. Only one declared the NCRG in its conflict of interest statement. Less than half (48%) referred to funding from the NCRG in the acknowledgements section or a footnote (Babor & Miller, 2014, p. 341). Unlike the fields of tobacco, alcohol and pharma, there is no tradition of declaring interests in gambling studies. Journals including Addiction have moved towards increasingly stringent requirements, prompting animated discussions (Babor & Miller, 2014; Conibear, 2014; Ellison, 2014). Despite pointed and strenuous interventions, sometimes in the very journals that are being criticized, gambling studies has resisted reform. Journal policies are inconsistent and applied in an ad hoc manner which does little to assist the reader in contextualizing the research. At conferences, researchers are forced to share rumours (accurate or not) about funding sources. As other fields have learned, a statement of interests, whether negative or positive, should be a basic requirement of every publication and presentation. Declarations of interest can take many forms, but the most effective are simple, including this example from Addiction:

The authors have no sources of funding, direct or indirect, nor any other connection with the tobacco, alcohol, pharmaceutical or gaming industries nor any body substantially funded by one of these organizations. (McCambridge, Hawkins, & Holden, 2014)

Declarations of interest are imperfect tools. Epistemologically, they do not enable us to metaphorically ‘strip’ research down to objective findings having eliminated the impact of vested interest. It is difficult to draw the limits of what information might help to interpret a particular piece of research. They are difficult to enforce, and to limit. What constitutes a conflict of interest in this field? Where do the edges of declarable interests lie? However, these challenges should not be used as an argument for inactivity (Gmel, 2010). There are many definitions of conflicts of interest which could be helpful to gambling studies. The Committee on Publication Ethics (COPE), for example, states in its Guidelines on Good Publication Practice (2003):

Conflicts of interest arise when authors, reviewers, or editors have interests that are not fully apparent and that may influence their judgments on what is published. They have been described as those which, when revealed later, would make a reasonable reader feel misled or deceived. (Committee on Publication Ethics [COPE], 2003)

Gestures to situate knowledge do not resolve bias or facilitate a single ‘correct’ reading of findings, but they do inform the reader or listener and contribute to a more open and critical field. They are one of a raft of related measures that is required to improve the quality and credibility of published research in this field.

Reproducibility

Concerns about the reproducibility of research have spread recently from medical science to the life sciences (Anonymous, 2013) including popular science (McNutt, 2014). Ledgerwood has recently described a ‘crisis of confidence’ and a ‘sea-change’ in psychological science, following

a number of largely unrelated events that happened to coincide – Jonah Lehrer's (2010) widely read New Yorker article on the effects of publication bias in science, Bem's (2011) controversial paper on precognition, a rising concern about direct replication, the Stapel fraud case (Tilburg University, 2011), and the publication of several troubling critiques of current practices in research and publishing (e.g., Simmons, Nelson, & Simonsohn, 2011; Vul, Harris, Winkielman, & Pashler, 2009). (Ledgerwood, 2014)

Gambling has a particularly wide credibility gap, with much work funded directly or indirectly by the industry, small pools of reviewers, the absence of a tradition of disclosure of sources of funding or data and a naive approach to conflicts of interest. A reproducibility policy which requires authors to disclose the data and computer code used for analysis would place gambling studies journals at the cutting edge of debates in academic publishing.1

International Gambling Studies: a brief case study

International Gambling Studies (IGS) is a member of COPE, but the disclosure policy provided online under ‘Instructions for Authors’ makes no reference to this membership and in fact states:

The Editor(s) accept that such support is often essential to enable research to occur; they will seek further clarification where appropriate. (IGS, 2014)

It is not clear exactly what ‘such support’ refers to – presumably from industry? It is not only accepted, but also described as ‘often’ (not ‘sometimes’, or ‘regularly’) ‘essential’. One might think that this would necessitate a particularly clear and robust policy on disclosure. On the contrary, IGS policy is discretionary and places no particular responsibilities on the author. It continues:

For more information about the declaration of potential conflicts see the guidelines complied (sic) by the International Society of Addiction Journal Editors http://www.parint.org/isajewebsite/conflict2.htm (IGS, 2014)2

The International Society of Addiction Journals Editors (ISAJE) webpage referred to provides a comprehensive ‘transparency form’ for completion by contributors and editors to journals, which has recently been adopted by Addiction. It requests information for the previous three years on (1) funding sources for the work; (2) constraints on publishing; (3) competing interests (financial and non-financial). It includes examples of competing interests:

RS has been reimbursed by the International Centre for Responsible Drinking for attending several conferences; TD has been paid by Monte Carlo Resorts Casino for running educational programmes and has her research registrar paid for by the company; JS has shares in the Kingfisher Brewing company. (International Society of Addiction Journal Editors, 2014)

Importantly, it also suggests that those who do not see themselves as having a conflict of interest must declare as such.

The electronic submission of articles to IGS requires authors to answer the following questions:

Do you have any conflict of interest including any financial/professional associations or interests which should be disclosed? If yes, please state.

Have you received any funding for this manuscript? If yes, please state the source of funding.

Authors are also required to tick a box which confirms that ‘I have uploaded a title page with a clear statment [sic] acknowledging any funding, support, and conflicts of interest where relevant in the title page along with the brief biographies of each author.’ Articles in the current edition of IGS continue to make no reference to declarations of interest, negative or positive (see, for example, Gainsbury, Hing, Delfabbro, & King, 2014 or Philander & MacKay, 2014). The purpose of invoking but not enforcing guidelines from ISAJE or purchasing membership of COPE is therefore opaque. This problem is not confined to IGS.

Conclusions

In tobacco research over the past 40 years, revelations about financial relationships between industry and researchers have discredited evidence, arguments and institutions as well as individuals (Petticrew & Lee, 2011). Those of us who are fed up with the current situation are beginning to ask similar questions about gambling studies. Who is paying for what? The spotlight will fall on research funding, consultancy, sponsorship and charitable donations to educational programmes and institutions. In Fair Game we referred to the now defunct Centre for the Study of Gambling at the University of Salford in the United Kingdom which received £100,000 a year from casino corporations including MGM Mirage, Isle of Capri and Kerzner International. Then head Peter Collins declared these relationships on the website and when asked about them by a journalist suggested that operators chose to support his centre because they shared the views that he held, independently (Barnett, 2007). How many currently active institutions who present themselves as independent sources of evidence for policymakers declare financial relationships with operators?

Individuals, conference organizers, editors and directors of research institutes who choose not to declare financial relationships with industry are playing a dangerous game. In the past, some might have thought it acceptable that financial support from charities set up to raise and distribute funding from the industry need not be declared. This is no longer a sustainable position. The body of evidence about the impact on research of industry funding is growing. It is becoming increasingly clear that all researchers must be explicit about their own position and allow readers and listeners to make judgements about how their work should be understood. Gambling studies has a number of other profound structural issues to address, including improving the mechanisms for funding research, but until journal authors and editors devise a systematic approach to disclosures of interest it will always be a second-rate field with unresolved methodological, ethical and epistemological issues which reflect poorly on every one of its practitioners.

Declaration of interests

Funding sources: The research leading to these results has received funding from the European Research Council under the European Union's Seventh Framework Programme (FP/2007–2013)/ERC Grant Agreement n. 263443. Between 2006 and 2009 I received £90,697.22 from ‘Research into Problem Gambling’, a collaborative research initiative between the Economic and Social Research Council (ESRC; lead organization) and the Responsibility in Gambling Trust (RiGT), now the Responsible Gambling Trust (RGT). All aspects of the grant were administered by the ESRC.

Constraints on publishing: Near the end of the project, I was asked by the RiGT to submit press releases for prior approval, a request that I declined. The ERC has not placed any restrictions on publishing other than that they be acknowledged as a source of support.

Competing interests: Between 2007 and 2009 I received ad hoc support from the National Lottery Commission for the Gambling Research Network, a group of early career and PhD researchers coming together in London two or three times a year. Money covered refreshments and no explicit restrictions or inducements were placed on the group by the NLC. I have paid to attend industry-sponsored events and attended free, industry-supported events. I have not received any other direct or indirect payments from the industry or any other groups substantially funded by gambling to conduct research or to speak at conferences or events. I have no other conflicts of interest to declare.

Additional information

Notes on contributors

Rebecca Cassidy

Rebecca Cassidy is an anthropologist and author of several books and articles on horseracing and gambling, including Sport of Kings: Kinship, Gender and Thoroughbred Breeding (Cambridge University Press, 2002) and Qualitative Research in Gambling: Exploring the Production and Consumption of Risk, edited with Claire Loussouarn and Andrea Pisac (Routledge, 2013). Her recent work focuses on the political economy of gambling research. In 2014, with Loussouarn and Pisac, she compiled Fair Game: Producing Gambling Research, a report based on interviews with 109 stakeholders in the field.

Notes

^1. The Transparency Project (http://www.thetransparencyproject.org/) is an example of an attempt by a group of researchers (lead by Howard Shaffer and Debi LaPlante of the Division on Addictions at the Cambridge Health Alliance, a teaching affiliate of Harvard Medical School) to share data contributed by an industry sponsor (bwin.party digital entertainment) with a wider audience. The fact that this attempt has not been universally well received is further evidence of the credibility gap in gambling research.

^2. The online ‘General Guidelines’ for authors submitting to IGS also indicate that ‘Authors must also incorporate a Disclosure Statement which will acknowledge any financial interest or benefit they have arising from the direct applications of their research. Disclosure statements should be included on the title page and also stated and described during the submission process. Full disclosure of any conflicts of interest is required at the time of submission.’ ‘Disclosure statement’ is hotlinked to ‘Disclosure of Conflicts of Interest’ a subsection of Taylor and Francis' author services (see http://journalauthors.tandf.co.uk/copyright/assignmentAndYourRights.asp#link3).