Academia in a Spin
Research-as-advocacy and the subversion of the Gambling Act Review
14th March 2022
Advocacy research is skewing the evidence informing the government’s review of gambling laws – restricting personal freedoms and further undermining an already highly regulated industry, writes Dan Waugh.
When Nigel Huddleston launched the government’s once-in-a-generation review of Britain’s gambling laws in 2019, he emphasised that it was to be an exercise in modernisation, grounded in hard evidence. After 27 months, three gambling ministers and four Secretaries of State, publication of a White Paper is now understood to be imminent; but there are also signs that the process has worked loose of Huddleston’s sensible moorings. In particular, the abasement of academia as a tool of advocacy rather than enquiry now presents a strong risk that this long awaited and long overdue review will result in regulatory dysfunction.
An analysis of Hansard reveals that the most commonly cited pieces of research in Parliamentary Questions on gambling since the launch of the Gambling Act Review are last year’s Public Health England report on gambling harms and the 2021 Oxford University study of Lloyds Bank data on gambling expenditures (Muggleton et al.). It is not surprising that these two reports have grabbed political attention. The former alleged social and economic costs of gambling in England of at least £1.2billion a year; while the latter claimed to have identified relationships between spending on gambling and a range of ‘harms’, including increased mortality. However, both studies contained serious methodological flaws; both made immoderate and unsubstantiated claims; both appeared to have been conducted with the aim of influencing the outcome of the government review. As we have seen only too well in recent years, ‘science’ can all too easily be bent to the will of vested or ideological interests.
To be fair to the Public Health England (PHE) researchers (now under the Office of Health Improvement and Disparities (OHID)), their task of estimating the social and economic costs of gambling was always going to be a challenge. Given the dearth of actual data that might be used to make scientific assessments, the team scratched around in the research literature for the odd straw to grasp at. A study of teenagers and twenty-somethings in Manitoba (2007 to 2012) was used to estimate rates of depression among gamblers of all ages in England; while hospital records from Sweden (2006 to 2016) were used for a wild guess at suicides. The PHE team either did not understand – or deliberately misrepresented – the research findings that they reviewed. Around half of the alleged £1.2billion cost related to increased mortality from suicide but the calculation in the report was based on the premise that Gambling Disorder is the same as ‘problem gambling’, which is categorically false.
This is not a question of splitting hairs. The former is a psychiatric disorder that has been recognised in clinical diagnosis guidelines since 1980 and is included in the World Health Organisation’s International Classification of Diseases. The latter (as used by PHE) is a sub-clinical classification obtained from surveys that indicates the likely presence of ‘adverse consequences from gambling’ and possible ‘loss of control’ but cannot be used to diagnose psychiatric disorder. People with Gambling Disorder will almost always be classified as ‘problem gamblers’; but a substantial proportion of ‘problem gamblers’ are unlikely to meet the diagnostic criteria for Gambling Disorder.
PHE’s error of misclassification was bad enough – but worse followed. As we have noted already, the Swedish study used as the basis for extrapolation (Karlsson & Hakansson, 2018) analysed hospital records of people receiving clinical treatment for Gambling Disorder. It is highly probable therefore that the records represent more extreme and complex cases, often involving a range of other mental health problems, including Major Depressive Disorder, Anxiety Disorders, Trauma and Stressor-Related Disorders, Alcohol Use Disorder and Substance Use Disorders (it is worth noting that, on average, fewer than 0.4% of ‘problem gamblers’ in Sweden receive hospital treatment for Gambling Disorder). The American Psychiatric Association states that ‘individuals presenting for treatment of gambling disorder typically have moderate to severe forms of the disorder’.
PHE compounded these basic errors by assuming that all excess harms experienced by people with Gambling Disorder (that is, over and above what might be expected in the general population) should be attributed to gambling – a claim that ignores well-established research findings on comorbidity and sequencing of disorders. To demonstrate the absurdity of this position, it implies that anyone with both a Gambling Disorder and an Alcohol Use Disorder suffers no harmful effects from alcohol (we imagine that when OHID looks at alcohol harms, this finding may be excised).
PHE neglected to make any estimate of positive effects from participation in gambling, despite a broad research literature on the psycho-social benefits of recreational betting and gaming. British Health Surveys have repeatedly shown that ‘non-problem gamblers’ (as measured by the Problem Gambling Severity Index) that is, the overwhelming majority of gambling consumers, enjoy better mental health on average than non-gamblers. This fact somehow eluded the researchers – perhaps because its incorporation into their model would have turned a £335million cost of depression from gambling into a £247million benefit (to be clear, we are not suggesting that this would be a robust finding; simply that it is the result of a more diligent application of PHE’s methodology).
When asked why benefits had been excluded from scope, the researchers replied that they thought a focus on costs would grab the attention of the government more firmly; and that they didn’t really think that consumers obtained any benefits from gambling anyway. This version of events is however, contradicted by the minutes of PHE meetings with its expert advisory group. The advisory group, which included at least two prominent anti-gambling activists, discouraged PHE from considering benefits on the grounds that they were already ‘well promoted’ and simply ‘not relevant’.
Gambling as the ‘new tobacco’
Right from the start, there were reasons to be suspicious of the PHE report. The public health lobby typically cites gambling as the ‘new tobacco’ – something that is inherently harmful. It is now common for public health researchers to state – without providing much in the way of justification – that gambling is an ‘Unhealthy Commodity’, causing damage regardless of dosage. In a few years, the public policy lens has shifted from concern about Gambling Disorder (suffered by around 0.2% of the population across mild, moderate and severe forms) and ‘problem gambling’ (an additional 0.3%); to concern about any participation in gambling at all (a little over half the adult population). In a journal article published in 2020, around a year after the PHE project had kicked off, its lead researcher abandoned any pretence that the exercise would be objective, writing: ‘more research and evidence are needed to support advocacy and action’. Such assertions are consistent with the research-as-advocacy phenomenon which conceives political action rather than learning as the object of academic study.
Since the start of the Gambling Act Review, the Oxford study of Lloyds banking data has been cited eight times in written Parliamentary Questions – more than any other piece of non-official research. It claims to have used transaction data from 102,195 Lloyds account holders to identify a relationship between gambling expenditure and harm. What ought to have been a fascinating piece of research was undone by a slew of methodological errors and the application of the research team’s own value judgements (for example, describing low spending on gym memberships and education as ‘harms’).
The researchers mistakenly assumed that a Pound deposited online or used to buy chips in a casino or tickets in a bingo club is a Pound staked; and that a Pound staked is a Pound lost (by way of illustration, ‘the house’ retains around 15% of the money exchanged for chips in land-based casinos; while margins for bookmakers are lower still). This error (the most glaring of many in the paper) is all the more remarkable when one considers that one of the authors of the paper (Dr Philip Newall of the Gambling Commission’s Advisory Board for Safer Gambling) is a former professional gambler and author of ‘The Intelligent Poker Player’, a book based entirely on the premise that gambling involves winning as well as losing – and that for a very small proportion, it actually provides a livelihood. In this way, a study that claimed to identify links between gambling expenditure (as a percentage of income) and harms, contained no information whatsoever on gambling expenditure and very little (if any) information on related harm.
In a similar vein, a study last year from researchers at the University of Bristol claimed to identify a relationship between the siting of licensed gambling premises and deprivation – essentially the rather unsurprising claim that bingo clubs and betting shops are popular in working class areas. The problem was that they used an incorrect list of gambling premises (overstating the number of casinos by 50%, the number of arcades by 30% and the number of betting shops by 20%). A different Bristol study (Rossi & Nairn, 2021 – the subject of a debate in Parliament at the end of last year), claimed that twitter adverts for gambling appealed more strongly to children than to adults – but was only able to do so by raising the threshold of adulthood from 18 years to 25. It also found that adverts for insurance held greater appeal for actual children than gambling adverts did, although this finding was conveniently suppressed in parliamentary debate.
A skewed evidence base
The minister responsible for the review, Chris Philp, has publicly endorsed both the PHE report and the Oxford-Lloyds study (despite the fact that his department had previously expressed reservations about the latter). The Gambling Commission has also put its name to the PHE estimates, despite the obvious falsities and is credited in the notes to the Muggleton et al. paper on Lloyds Bank data. This presents a very real risk of unsound policy formulation based upon a skewed understanding of the evidence base. Parliament, meanwhile, tends to be lobbied rather than informed – and this is likely to undermine the ability of both Houses to apply proper scrutiny to the forthcoming process of legislative change.
Worse is likely to follow. PHE/OHID also commissioned a Delphi Study to recommend Government interventions to counter the alleged £1.2billion of gambling harms. This involves the canvassing of views of a secret list of academic experts (anyone in receipt of funding from the gambling industry in the prior decade was disbarred from the group but no proscriptions applied to other vested interest, such as Big Pharma or political activism) and people with ‘lived experience’ of harm. Recommendations will include ‘regulation, restrictions, limits, pricing, levies, taxes, duties, and related actions to make gambling products less affordable and less attractive to gambling operators and consumers’ alongside initiatives to reduce the visibility and availability of betting and gaming. This report is understood to have been completed and now awaits publication – just in time for the DCMS White Paper on proposed changes to Britain’s gambling laws.
One of the most dangerous outcomes of an ‘evidence based’ review is one based on evidence that is biased, incomplete or flawed. This guarantees an over-complex and stifling regulatory regime that squashes the industry without actually protecting customers. Unfortunately, this the clear direction of travel for gambling reform in Britain, with too few stakeholders willing or able to engage with evidence in a more robust manner than choosing which ‘side’ to believe.
In the absence of a coherent and suitably independent programme of official research, it is unsurprising that research-as-advocacy should fill the vacuum. The suggestion however that “we would not start from here” is not helpful in the middle of a review. What is critical in the present situation is that Westminster and Whitehall know when they are appraising high quality data that has been properly analysed or being fed propaganda wrapped in scientific language. If politicians want to restrict personal freedoms and undermine a regulated industry simply because they don’t like it then they should be honest enough to say so. If they want to regulate it more effectively using evidence then they need to review that evidence dispassionately and critically. We are approaching the last chance for ‘evidence-based’ decision making to mean what it should in the context of gambling reform in Great Britain. This is also therefore the last chance for politicians to put consumer interests and proportionate regulation above grandstanding and the appeasement of pressure groups.