UK Gambling Survey Faces Scrutiny Over Participation Rate Discrepancies

Dan Waugh from Regulus Partners presented findings at the UNLV Eadington Conference 2026 that compare the UK Gambling Commission's Gambling Survey for Great Britain against operator records, and these comparisons reveal consistent overstatements in self-reported participation figures across multiple categories. The analysis draws on administrative data from the Betting and Gaming Council to benchmark survey responses, while researchers note that the differences range from several hundred percent in some segments to nearly seven times higher in others. Conference attendees in May 2026 heard how these gaps could influence regulatory decisions that rely on the GSGB as a primary source of behavioral insights.

Survey methodologies often depend on participant recall and voluntary responses, whereas operator data captures actual transactions and visits recorded in real time. Waugh's work isolates specific products where the divergence appears most pronounced, and it flags the potential for systemic bias when survey instruments fail to align with verifiable records. Those who've examined similar survey designs in other jurisdictions observe that response rates and question framing can amplify certain behaviors while undercounting others, yet the UKGC figures show directional overestimation across the board rather than random variation.

Key Areas of Divergence in Reported Activity

Casino table game participation estimates from the GSGB sit between 408 and 628 percent above the actual visit counts compiled by industry operators, according to the side-by-side figures presented at the conference. Football pools participation appears inflated by as much as 694 percent when measured against the same administrative baseline, and betting exchange usage registers 225 percent higher in survey responses than transaction logs indicate. Each of these products represents a distinct player segment, and the pattern of overstatement persists even after standard demographic weighting is applied in the survey process.

Further breakdowns in the research separate online and land-based channels, revealing that the largest gaps occur in activities where players may have limited or infrequent engagement. Observers note that infrequent participants sometimes round up their activity levels in recall-based questions, which can shift aggregate percentages dramatically when base rates are low. The analysis stops short of attributing motives but documents how these inflation effects compound when policymakers extrapolate from survey percentages to population totals.

Implications for Regulatory Decision-Making

Policy development in the UK gambling sector draws on the GSGB to set prevalence benchmarks, assess harm levels, and calibrate interventions. When survey figures exceed operator-verified activity by multiples of two to seven times, the resulting risk models may overstate exposure across the adult population. Waugh's presentation connects these measurement issues directly to the reliability of downstream regulatory outputs, and it urges closer integration between survey instruments and transaction-level datasets maintained by licensed operators.

Regulators already maintain access to granular data through licensing conditions, yet the GSGB continues to operate as a standalone population survey without routine calibration against those records. The conference discussion explored whether hybrid approaches, such as weighting survey responses by verified operator benchmarks or supplementing them with anonymized transaction samples, could reduce the observed gaps. Data from the Analysis comparing GSGB participation estimates to operator data illustrates how unadjusted survey outputs currently feed into public consultations and impact assessments.

Conference Context and Broader Measurement Challenges

The UNLV Eadington Conference has long served as a forum for examining gambling regulation, economics, and data integrity, and the 2026 edition placed renewed emphasis on evidence quality amid evolving licensing frameworks. Waugh's session drew attention because it juxtaposes two data sources that rarely appear together in official reporting, and it quantifies the scale of misalignment using product-level examples rather than aggregate trends alone. Participants discussed parallel issues in other jurisdictions where self-report surveys have diverged from tax or licensing records, underscoring that the UK case is not isolated.

Technical factors such as question wording, sample frame construction, and handling of non-response can each contribute to divergence, while operator data benefits from automated capture that eliminates recall error. The research does not claim one source is inherently superior for all purposes; instead it demonstrates that exclusive reliance on the GSGB introduces measurable upward bias for certain participation metrics. Those responsible for survey design have acknowledged similar limitations in methodological reviews, yet the scale of the discrepancies documented here exceeds previously reported margins of error.

Conclusion

The findings presented by Dan Waugh at the 2026 Eadington Conference supply concrete benchmarks that contrast survey-derived participation rates with operator administrative records, and they quantify overstatements reaching several hundred percent for specific gambling products. These discrepancies carry direct relevance for any regulatory framework that treats the GSGB as an authoritative measure of market activity and exposure. Continued refinement of data collection methods, including tighter linkage between survey responses and transaction data, offers one pathway toward reducing systematic bias in future iterations of the survey.