Regulation by vibes: Gambling Commission’s youth gambling statistics raise questions

The UK’s gambling regulator’s response to questions from iGB raises the prospect of regulation by vibes: a statistic apparently describing something children did is defended as a measure of something they merely felt. Feelings can be worth studying but policymakers need to know which they are being offered, particularly when the distinction emerges only after someone points out that the answers do not add up.

The issue concerns advertising and Regulus Partners’ consultant Dan Waugh has highlighted an anomaly: a majority of children reporting that advertising prompted them to spend money on gambling also reported no gambling expenditure.

The published 2025 report says 7% of those who reported exposure to advertising; or 5% of all children in the survey, which represents a figure of around 200,000 people. The figure has been quoted multiple times in parliament by the anti-gambling lobby, including by Professor Heather Wardle, who now heads up the UK’s Gambling Harms Research Centre. Yet a majority of those who responded to that question were reported elsewhere in the survey never to have gambled.

Asked to reconcile the findings, the Commission told iGB the results were descriptive and should not be interpreted as evidence that advertising directly caused gambling participation.

“The survey asks respondents whether they have ever felt prompted to gamble after seeing gambling advertising or promotions, but it does not seek to establish a causal relationship between exposure and behaviour.”

It added that feeling prompted and subsequently participating were distinct concepts, invoking the “intention-behaviour gap”: someone can experience an urge without acting on it.

Interpreting the question posed to children

That is a perfectly intelligible distinction. Unfortunately, it is not the distinction the question actually asks children to make.

The wording is: “Have adverts or promotion about gambling ever prompted you to spend money on gambling when you were not otherwise planning to?”

There is no reference to an intention left unfulfilled. The obvious conclusion is, therefore, that money was spent, with advertising supplying the prompt. A child might interpret it differently, but that would raise a question about comprehension rather than settle the anomaly.

Mis-character witness

Waugh, who wrote to the Commission on 19 August and received its response on 25 September, calls the explanation a “mischaracterisation”. His subsequent analysis argues that the question implies a causal relationship rather than simply asking what happened after exposure.

There are two separate issues here. A respondent’s attribution cannot, by itself, establish that advertising caused an action. But that methodological limitation does not turn a question about reported spending into a question about an unacted-upon feeling. The Commission’s answer appears to slide between the two.

Its published presentation is also less clear-cut than its defence suggests. The 2025 report describes perceived advertising effects and sometimes uses language about feelings. Elsewhere, however, it describes respondents as having been prompted to spend. That mixture leaves readers to decide whether the statistic concerns perceived motivation, actual expenditure or both.

Feelings over actions

Another objection from Waugh concerns how the Commission has used such findings. His analysis reproduces passages from its 2023 advice to government on the Gambling Act review, including a statement that advertising and sponsorship prompted 7% of children aged between 11-16 to spend on gambling, citing the 2022 survey.

This was presented within the discussion supporting advertising restrictions. As Waugh asks: “Did the Commission make it clear to the government that these statistics referred to feelings rather than action?”

It is a question that deserves some kind of answer. If the measure captures urges, its policy relevance should be explained accordingly. If it captures reported spending, contradictory participation answers require investigation. Describing it one way in policy advice and another when challenged risks making the evidence look conveniently adaptable.

Waugh’s concerns extend beyond advertising. Across the 2023-2025 datasets, he identifies six children who reported spending their own money on all 17 listed gambling activities in the preceding week. The repertoire included lotteries, betting shops, online gambling and casino games. One respondent was 12.

The Commission says the small number of extreme responses did not fail its established quality checks. It argues that researchers should not amend or remove answers simply because they appear unusual, and that self-report surveys inevitably contain some measurement error.

Reasonable points, perhaps, but neither demonstrates that these particular answers are accurate. Waugh says he was not demanding their removal; he wanted their implications for reliability and interpretation acknowledged.

Aggregate score

The Commission also disputes his aggregation of three survey years because of inconsistent weighting. Waugh responds that weighting matters for population estimates, while his purpose was to identify inconsistent individual responses. He says the same issues arise within each year.

The Gambling Commission has raised no such concerns with regard to aggregation elsewhere. For example, Professor Wardle recently published estimates of problem gambling for all local authorities in England, based upon aggregation of data from the Gambling Survey for Great Britain.

On problem gambling, the regulator explains that its youth screening tool measures behaviours, experiences and consequences over 12 months. Relatively limited participation can therefore coexist with a classification indicating problems. Low participation alone does not disprove that classification. Waugh’s argument is that readers need more context to assess what the headline estimates mean.

Understanding that a meaningful proportion of the children identified as “problem gamblers” have not taken part in any age-restricted gambling activities (or that they have solely played lottery draws) casts a different complexion on the statistics than they are usually accorded.

Independent methological review pending

The Commission describes extensive checks for routing errors, rapid completion and disengaged responding. However, its reply did not provide the requested number of exclusions. It says checks undertaken before it receives the data are not published. An independent methodological review is planned for publication in spring 2027, which it describes as routine rather than a response to Waugh.

All this sits awkwardly beside the Commission’s own warning about statistical misuse. In his August 2023 open letter, then-chief executive Andrew Rhodes said “nobody is well-served by statistics being misused to further an argument”. He urged accurate interpretation, proper context and necessary caveats.

The Commission says it has found no evidence that the concerns undermine the survey’s overall integrity, while acknowledging opportunities to improve presentation. The exchanges do not establish deliberate manipulation or invalidate the entire survey. They do reveal an unresolved discrepancy between a question about spending and a defence based on feelings.

A regulator asking everyone else to use statistics responsibly should be able to explain that discrepancy without changing what its question appears to measure.

Until it does, it means we have to consider a less comfortable interpretation: that the Commission’s handling of its own evidence is not as trustworthy as its lecturing of the industry suggests.