Report
Phase 4: Illegal Gambling Index – Technical Report
The Gambling Commission’s technical report detailing the development of the Illegal Gambling Index.
Development of the illegal gambling index
All 7 statements and the main intentionality question (Question A1) were put forward to contribute to a pilot composite index score for the Gambling Commission to track via the Gambling Survey for Great Britain (GSGB). The final Illegal Gambling Index approach can be found in the Final illegal gambling index section.
A composite index was chosen as it allows the Commission to aggregate multiple responses into a single metric to give them a holistic view of the online illegal market (websites only). This is particularly pertinent given the broad and complex nature of the indicators of illegal gambling, as demonstrated earlier in this report.
The index is compiled based on strength ratings assigned to each indicator, based on insights and data detailed earlier in this report, and agreed between Yonder and the Commission in a workshop session.
Yonder designed a Composite Index that can flex to reflect new indicators of illegal gambling in the future, or a change in the strength of a code’s indication of illegal gambling (that is, if wording is amended, verbatims in support are more convincing, or additional checks are put in place in the survey).
8 indicators feed into the pilot index produced, with varying strength, ranging from ‘Near Certain’, ‘Strong’, ‘Weaker’. These strength ratings were assigned based on:
- qualitative insight; (both from Phase 1 of the research, where we gained insight into the extent to which consumers understand the behaviours they are identifying, and from verbatims accompanying responses in Phase 2, which in some cases disputed indicator selection (such as, the naming of cryptocurrency trading sites, rather than gambling sites))
- survey design experience (that is, acknowledging where there is scope for misinterpretation or difference within a survey code)
- issues with self-reporting, evidenced by other research (such as, a quarter of British people being unsure on the difference between a credit and debit card).
Of the 8 indicators in this specific project, 3 were given a ‘Near Certain’ rating, 4 were given a ‘Strong’ rating, and one was deemed to be ‘Weaker’. Details of the ratings and their rationale were decided in a workshop between the Commission and Yonder.
Index scores were assigned in relationship to one another, based on strength. To calculate it, we assign scores of ‘3’ to ‘Near Certain’, ‘2’, to ‘Strong’, and ‘1’ to ‘Weaker’. Pilot Index scores are then halved for each indicator if identified in the last 12 months – then halved again if identified further back – that is, ‘ever’, to demonstrate the decreasing confidence in response based on poorer recall.
Table 1: Strength rating given to indicators
| Strength of indicator | A1. Intention | A3. Not on Gamstop | B1_A. VPN usage | B1_B. Paying using cryptocurrency | B1_C. Paying using a credit card | B1_D. Paying using NFTs | B1_E. Avoiding verification | B1_F. Cancelling a withdrawl |
|---|---|---|---|---|---|---|---|---|
| Strength of indicator | Near certain | Strong | Strong | Strong | Weaker | Strong | Near certain | Near certain |
| Yes, in the past 4 weeks | 3 | 2 | 2 | 2 | 1 | 2 | 3 | 3 |
| Yes, in the past 12 months | 1.5 | 1 | 1 | 1 | 0.5 | 1 | 1.5 | 1.5 |
| Yes, ever | 0.75 | 0.5 | 0.5 | 0.5 | 0.25 | 0.5 | 0.75 | 0.75 |
| Never | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
From this initial stage, as you can calculate by totaling the top row of the table, the maximum feasible pilot index score is 18. To create an easy-to-comprehend, single metric for the Commission, we have then scaled up the index from a maximum of 18 to 100 (by multiplying by 5.56) so that the maximum feasible score – that is, if a respondent identified all 8 indicators within the past 4 weeks – is 100.
So, given the strength ratings assigned and the maximum combination of these totalling 100, the pilot index calculations were as shown in Table 2.
Table 2: Index calculations for the indicators
| Strength of indicator | A1. Intention | A3. Not on Gamstop | B1_A. VPN usage | B1_B. Paying using cryptocurrency | B1_C. Paying using a credit card | B1_D. Paying using NFTs | B1_E. Avoiding verification | B1_F. Cancelling a withdrawl |
|---|---|---|---|---|---|---|---|---|
| Strength of indicator | Near certain | Strong | Strong | Strong | Weaker | Strong | Near certain | Near certain |
| Initial rating | 3 | 2 | 2 | 2 | 1 | 2 | 3 | 3 |
| Multiplication factor | 5.56 | 5.56 | 5.56 | 5.56 | 5.56 | 5.56 | 5.56 | 5.56 |
| Yes, in the past 4 weeks | 16.7 | 11 | 11 | 11 | 5.7 | 11 | 16.7 | 16.7 |
| Yes, in the past 12 months | 8.3 | 5.5 | 5.5 | 5.5 | 2.8 | 5.5 | 8.3 | 8.3 |
| Yes, ever | 4.2 | 2.8 | 2.8 | 2.8 | 1.4 | 2.8 | 4.2 | 4.2 |
| Never | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
Yonder suggested respondents that demonstrated one ‘near certain’ indicator, or any combination of 2 other ‘Strong’ indicators, in the past 4 weeks, could confidently be categorised as an ‘Illegal gambling website user in the past 4 weeks. In practice, this means a threshold score of 16.7 when looking at the index based on responses from the past 4 weeks only.
Where this threshold isn’t met – that is, respondents have not demonstrated enough indicators in the past 4 weeks to give us confidence in categorising them as an illegal gambling website user – their index score considering all timeframes is analysed. Again, if the threshold of 16.7 is met here, Yonder suggests they are categorised as ‘Illegal website users further in the past’.
Where this threshold isn’t met – that is, respondents have not demonstrated enough indicators in the past 4 weeks nor further back to give us confidence in categorising them as ever having used gambled on an illegal website, we look at whether they have demonstrated any indicators at all. If they have (for example, they used a credit card to gamble at some point, ever10), they will have a low index score (below 16.7). These respondents are categorised as ‘Minimal indications’ for analysis, but do not feed in to our ‘illegal group’, owing to the lack of confidence we can truly have that they have engaged with illegal sites.
The majority of respondents have an index score of 0, having never demonstrated any indicators of illegal website gambling. For analysis, these respondents are categorised as ‘No indications’. We can use their responses to questions about upcoming regulations in the gambling industry, and their motivations when choosing where to gamble, to determine whether they are at-risk of being pushed into the illegal market in future.
References
10Based on the design of the code used in this report, choosing the ‘ever’ response means we cannot be certain if they used a credit card to gamble while it was banned by the Commission for licensed websites, or previous to the ban.
Introduction to the development of the analysis approach Next section
Flexing the index for the GSGB
Last updated: 5 November 2026
Show updates to this content
No changes to show.