Insights Statistics and the Eye of the Beholder


The nation has just received a ticking-off from the chief executive of the Gambling Commission of Great Britain on the subject of statistics.[1] The nub of the rebuke, contained in an ‘open letter’ published by the Commission on its website on August 14th, is that ‘The Commission is very concerned at the significant increase in the misuse of statistics around gambling as different parties seek to make persuasive arguments for or against different proposals’.  It is hard to argue with the overall sentiment and, indeed, it is fair to say that there have been examples of industry participants and the anti-gambling lobby doing just that in recent years in an endless cycle of points scoring rather than problem solving. The open letter states that:

“Where Official Statistics have been used inaccurately, the Commission will generally assume the misuse was accidental and ask the party to correct the record. If the party fails to do so, or declines, then we will consider referring the matter to the Office for Statistics Regulation. Something we have done recently. The Commission also reserves the right to publicly challenge the misuse of statistics by any party, if they fail to correct their misstatement.”

Perhaps inevitably, the raw nerve at the Commission relates to the bandying-about of statistics relating to problem gambling and gambling-related harm. The Commission’s concern is that:

‘Multiple individuals and organisations have misused problem gambling statistics to create an inaccurate picture. For example, it has been stated that ‘99.7 percent of people who gamble do so without being harmed’ and variations on this theme such as arguing that ‘only 0.3 percent of gamblers are harmed’. This is not true and misrepresents the statistics.’

However when it comes to the collection, interpretation and exploitation of data, the Commission should consider the beam in its own eye as well as what motes there may be in the eyes of others. As the National Audit Office, in its 2020 report on the way the Commission discharges its functions, witheringly put it:[2]

The Commission undertakes some evaluation of its impact, although these are mostly limited examinations of whether its projects have been delivered as planned, with little assessment of the outcomes achieved, the timeliness of interventions or the effectiveness of industry-led initiatives to protect consumers from harm.’

In other words, the NAO seemed to think that the Commission itself was seriously data – and statistics – deficient in the execution of its core statutory functions, functions the exercise of which are required by basic good public sector practice (and the Commission’s own governance documents) to be ‘evidence-based’. Again, in the words of the NAO:

‘The Commission has translated its statutory objectives and overall aim to make gambling safer into strategic priorities, business plan activities and high-level outcomes it wants to achieve. However, it has not yet developed these high-level outcomes into detailed, measurable success criteria against which to judge progress.’

So the Commission hardly comes to the argument over statistics – data – with clean hands itself. This is particularly tricky when one considers the complexity of the core phenomenon of ‘gambling harm’, including difficulties of identification, quantification and treatment, and when one further considers that the Commission has, at times, seemed to morph itself into a public health regulator.  In the ‘open letter’ there is a rare acknowledgment from the Commission ‘that the problem gambling rate overall is low, compared to the level of participation’. However it is not at all certain that the Commission knows why it is low even though the ‘key conclusions’ of the Commission’s most recent round of telephone surveys state that ‘The period from the end of 2020 saw the proportion of respondents classified as moderate risk and low risk gamblers significantly decrease’.[3] Given its statutory obligation to advise HM Government, as well as its obligation to based the exercise of its functions on evidence, the Commission’s own understanding of core industry data is surprisingly incomplete.

To the writer’s knowledge, at least, the Commission has not published any detailed evaluation of the effect of the guidance on SR Code 3.4 relating to customer interactions, or the additional customer protection guidance issued during the pandemic, or the ban on reverse withdrawals, or new practice in relation to high value customers, or GAMSTOP, or indeed the effect of its own regulatory settlements and sanctions and the general rapid and substantial industry-wide improvements in player protection over the last several years, including the widespread rollout of financial risk checks. The Commission’s web page ‘ABSG Progress Report on the National Strategy to Reduce Gambling Harms – Year Two’ says that ‘The need to evaluate in order to understand what works is a long-standing priority’.  But, for example, the credit card ban, introduced in April 2020, had to wait until November 2021 before an ‘interim evaluation’ was published, and we still apparently await the full evaluation that The National Centre for Social Research was apparently commissioned to provide in May 2021 and which was scheduled for completion in early 2023.[4] A good milestone, by contrast, was the Commission’s publication of an assessment of online games design changes on June 8th 2023.[5] This sort of evaluation needs to be far more widely utilised.

As well as setting a better leadership example in research, statistics and the use of data itself, the Commission may well have bitten off more than it can chew, politically-speaking, with the ‘open letter’. The misuse of statistics that the Commission complains of in its ‘open letter’ seem to be twofold: first, the misrepresentation of problem gambling rates as per-gamblers and not per-population, and secondly the inaccurate representation of trends. Although the ‘open letter’ purports to rebuke ‘All parties seeking to rely on statistics’ and ‘Anyone seeking to compare data sets’ the examples of misuse in fact only castigate those who defend the industry by pointing out that problems are – in all honesty and without any complacency whatever and recognising that there are few who don’t agree with proportionate measures to protect the vulnerable – relatively low. Equally, if not much more misleading misuse of statistics by anti-gambling lobbyists, which enjoy far greater media prominence, are not mentioned.

The ‘open letter’ might have come across as a little better-balanced had the Commission also taken the chance to cite some of the statistics from commentators such as the Bishop of St Albans or, indeed, Public Health England itself, which are in a different league altogether to the examples cited in the ‘open letter’. In the case of Public Health England, FOI responses from the Commission indicated that even the Commission’s own officials didn’t understand PHE’s conclusions, speculating in an internal memo that they were simply designed ‘to ensure gambling is considered as a public health issue’. [6] The entire industry will look forward to the blood-sport that will ensue as the Commission throws itself into the task of imposing statistical objectivity over my Lord Bishop, his parliamentary colleagues and indeed the Commission’s fellow Agencies of State, whose agendas often very clearly require the picking and choosing of statistics to fit.

Mr Rhodes has generally seemed more prepared than his immediate predecessors to seek to strike a balance between creating a sustainable gambling market in Great Britain which gives freedom to the many and protects the vulnerable few with a commitment to base regulatory change on evidence and for that we must be thankful. However, having set out that stall, the Commission must be objective in its own use of statistics (including in current and forthcoming consultations) and look to hold outliers on both side of the fence to account.


[1], (15 Aug. 2023).

[2] See:, paragraph 13, (24 Feb. 2021).

[3], (15 Aug. 2023).

[4], (15 Aug. 2023).

[5], (15 Aug. 2023).

[6] Readers are encourage to visit: for the critique by Regulus Partners of both PHE and OHID reports (15 Aug. 2023).