This box is not visible in the printed version.
The Gambling Commission’s technical report detailing the development of the Illegal Gambling Index.
Published: 5 November 2026
Last updated: 5 November 2026
This version was printed or saved on: 5 November 2026
Online version: https://www.gamblingcommission.gov.uk/report/phase-4-illegal-gambling-index-technical-report
This report details the technical approach taken to developing a metric to track consumer usage of illegal gambling websites for the Gambling Commission’s Gambling Survey for Great Britain (GSGB). This report was produced by Yonder Consulting in collaboration with the Commission as part of their Consumer Voice Research Programme.
This report documents the evolution of the question set from initial development and piloting through to the methodology, including refinements informed by subsequent research and testing. The development of the approach represents an important step in strengthening the Commission's evidence base on participation in online illegal gambling in Great Britain. Designed as a flexible and evidence-led measure, the index provides a framework for monitoring consumer engagement with illegal gambling websites and may evolve over time as new evidence, behaviours and market developments emerge.
Through conducting a multi-phased research project aimed at understanding consumer engagement in illegal gambling websites, Yonder was able to develop an Illegal Gambling Index to provide the Commission with a metric through which it can track consumer engagement with illegal gambling websites. The aim of the index is to identify varying degrees of confidence in consumer engagement, rather than to provide a definitive measure of illegal gambling participation.
The Index uses survey responses to 7 ‘indicators’ that reflect behaviours associated with gambling on websites that are not licensed by the Commission, and are therefore considered ‘illegal’ gambling websites. To reflect differing levels of confidence that an indicator represents engagement with an illegal gambling website, each indicator was assigned a strength rating (and corresponding numerical score).
3 indicators were classified as Near Certain (score of 3) because they are closely linked to the Commission’s licensing requirements and thus provide a high degree of confidence that the website is operating illegally:
3 indicators were classified as Strong (score of 2) as they reflect behaviours that are closely associated with illegal gambling websites but do not provide definitive evidence of engagement. A lower degree of certainty was applied to these measures because respondent misunderstanding, misclassification or inaccurate self-reporting may account for a proportion of reported activity:
An indicator was classified as Weak (score of 1) as it may indicate engagement with an illegal gambling website but is subject to a higher degree of uncertainty. This lower confidence primarily stems from the possibility of respondents misreporting payment methods, due to poor recall or due to confusion between credit and debit card use.
Placed a bet and/or played a game on a gambling website using money directly from your credit card (that can be paid off each month) was classified as a weak indicator.
When the survey respondents are asked whether they have ever experienced any of these 7 indicators, they are presented with 6 response options to choose from:
As the gambling landscape changes, new indicators may be added, and existing indicators may be revised or removed. Accordingly, the scoring framework may be recalculated to ensure the index remains robust, relevant, and capable of accurately capturing consumer engagement with illegal websites.
Using the composite index score and the recency of reported behaviours, respondents are classified into 4 groups:
This approach to classifying survey respondents enables reporting of both likely illegal gambling website users and respondents demonstrating limited indications of engagement, while recognising the importance of behavioural recency when assessing confidence in illegal gambling website engagement.
This report details the technical approach taken to developing statements concerning consumer usage of illegal gambling websites for the Gambling Commission’s Gambling Survey for Great Britain (GSGB). This report was produced by Yonder Consulting in collaboration with the Commission as part of their Consumer Voice Research Programme1.
The study was designed to assess various indicators of illegal gambling website usage, to inform the Commission of key areas to focus on to track consumer engagement in the online illegal market. The research was underpinned by the Commission’s Licensing objectives as set out in the Gambling Act 2005 (opens in new tab), its Corporate Strategy, and its Evidence gaps and priorities.
The primary aim was to gain a comprehensive understanding of what behaviours and experiences can be best used to understand whether an individual has engaged with an illegal gambling website.
From a practical sense, this study had 3 output objectives:
This report focuses on the development process for the illegal questions, as well as providing a suite of recommendations for future-proofing the question for future iterations of the GSGB.
1 This report was co-authored by Yonder and the Gambling Commission. Further details on authorship of individual sections are provided in Appendix B.
The full report has a comprehensive description of the multi-phased methodology to meeting this project’s objectives. A short overview has been provided in the following timeline:
Focus group with 11 members of the Gambling Commission’s Lived Experience Advisory Panel (LEAP) to explore consumer pathways into online illegal gambling and test questionnaire content for the quantitative stage.
The quantitative incidence screening questionnaire was conducted with 1,000 individuals who had gambled online in the last 4 weeks, excluding those that had only played National Lottery products2. Within this questionnaire, a suite of indicators of illegal gambling, motivations, factors and experiences were presented to respondents. Data and re-contact consent from this survey was used to shape a subsequent qualitative phase, speaking to those who gamble with a range of experiences and engagement in the online illegal market.
10 online depth interviews with respondents who completed the online survey, to delve into their pathway into illegal gambling, recounting levels of awareness of engagement with illegal websites, the role of regulation, whether they do any research to vet these websites, their motivations for using these websites, and attitudes towards future use of illegal websites.
2 focus groups with a total of 7 individuals recruited from the Commission’s wider lived experience network, peer facilitated by members of LEAP, exploring experiences of using illegal gambling websites while registered to Gamstop.
Following this, the Commission asked Yonder to develop a proposal for Phase 2 of the project, building upon the findings from the mixed-methodology project earlier in the year3. The key objectives were to test and validate the findings from the first phase, as well as drafting questions based on the data that could be taken forward into the Gambling Survey for Great Britain (GSGB).
An online quantitative survey with 2,000 adults who had gambled online at least once in the past 4 weeks, excluding those that had only played National Lottery products, was conducted. The survey was created based on learnings from Phase 1 on the most reliable indicators of illegal websites that could be used to gauge engagement.
Based on the responses received from the survey, as well as extensive survey design experience in both the Commission and Yonder, a workshop session was held to assess the strength of each illegal gambling indicator, how they might be refined for the GSGB, and how they could be combined into a single, trackable metric (a composite index) for GSGB stakeholders.
The newly developed illegal questions were added to the 2025, Wave 2 GSGB survey, fielded in between April 2025 to July 2025. These pilot questions were included on the online version of the GSGB throughout 2025 and have been included in all subsequent waves.
Whilst the GSGB questions were being piloted, Yonder continued to develop the analysis approach for these questions. Yonder delivered a technical report to the Commission outlining the question and index development. This report was reviewed by the Commission and scheduled to be published alongside the first illegal gambling GSGB data release.
The Commission received the first quarterly wave of data from the piloted GSGB questions and analysed the data using Yonder’s original index approach. Within this process, refinements to the index were made, including the removal of questions relating to intentionally using illegal gambling websites and the upgrading of the strength of one of the statements.
The Commission has added a section to Yonder’s original technical report to reflect these changes.
Phase 3 is a mixed-method study building on the findings from previous phases to develop an analytical methodology for quantifying persona composition and the primary 'push factors' contributing to consumer usage of illegal gambling websites.
Note: This phase sits outside the scope of this report, which discusses the development of the Illegal Gambling Index only.
This technical report details the approach taken to developing statements concerning consumer usage of illegal gambling websites for the GSGB, as well as providing a suite of recommendations for future-proofing the question for future iterations of the GSGB.
This report is also published alongside the ‘Initial findings from new data on consumer engagement with illegal gambling websites’ report, which presents initial GSGB findings based on data collected across 4 waves running quarterly between April 2025 and April 2026, with a combined sample of around 14,000 respondents who completed the survey online.
We have used a multi-phase process, with methodologies spanning qualitative depth, quantitative rigour, and data science, to develop this index, using findings from earlier phases to validate and refine the indicators to be taken forward into the the Gambling Survey for Great Britain (GSGB).
Qualitative research: This ensures that all findings are contextualised with a richer, more nuanced understanding. Conducting focus groups with members of the Lived Experience Advisory Panel (LEAP) ensures that those with true lived experience of gambling are represented in the development of the research.
Quantitative research and data science: This ensures that findings have a strong analytical grounding and have enabled the Gambling Commission to develop measurement techniques to understand the illegal gambling market over time.
The development process has consistency engaged with key stakeholders to ensure the index is reflective of the gambling space, through workshops with the Commission’s Lived Experience Advisory Panel (LEAP), their Industry Forum, academic specialists, and researchers within the Commission.
The indicators that form the basis of the index are grounded in real-life behaviours that are associated with illegal gambling websites, rather than solely being based on respondents’ own understanding of the legality of the websites they use.
The index is flexible and can be adjusted to respond to emerging indicators of illegal gambling websites as the industry adapts. It also accounts for varying level of confidence that researchers can have in each indicator, building in context around recall issues through survey design experience and qualitative insights.
The index is not intended to be a definitive measure of the size and scope of the illegal gambling market. The aim instead is to identify degrees of confidence in consumer engagement, with the indicators acting as proxies for engagement, rather than explicit evidence.
Consumer understanding and recall of engagement with the illegal market is limited. Qualitative research has previously found that respondents do not always know if, and when, they have used an illegal gambling website, often due to tactics deployed by the company to disguise their legality. Generally speaking, survey recall of beyond 4 weeks is limited; we know that this holds true when respondents are considering their previous gambling behaviour, too.
Qualitative research: Qualitative sample sizes are by their nature small, so conclusions can only be directional, unless contextualised by quantitative data. Qualitative research heavily relies on the skills and experience of researchers, introducing potential researcher bias although this is mitigated by employing a grounded theory approach, where thematic grid analysis is used to capture qualitative data and notes, and where narratives are developed through rigorous analysis of this data. All insights gathered from this study are based on self-reported behaviour rather than observed behaviour, meaning insight is limited to what participants felt comfortable revealing to Yonder in the research setting. While interviews mentioned in this report took place with a range of participants, the focus groups mentioned above took place with either members of LEAP or via their facilitation, meaning the full scope of gambling engagement could not be fully represented.
Quantitative research: Quantitative research provides a valuable snapshot of attitudes and opinions, offering a broad view of trends and patterns across a specific sample. However, one key limitation is that quantitative results often lack the depth to explain why respondents hold certain views or what specific factors drive their responses. The sample selections for the quantitative research also provides limitations in the interpretation of results. The development survey samples selected from Yonder’s research panel were those who had gambled in the past 4 weeks, excluding National Lottery only players. As such, findings from these surveys should not be treated as representative of all those who gamble, and caution should be taken in assessing the significance of sub-group differences such as gambling frequency and Problem Gambling Severity Index (PGSI) score, as these groups may also not be representative, and may represent a higher than natural proportion within the sample than they would in a survey of the general population. This is mitigated against by the inclusion of the GSGB in the development of these statements.
Phase One began with a scoping exercise, in which Yonder spoke with the Gambling Commission’s Enforcement teams, and their Lived Experience Advisory Panel (LEAP), in order to inform a preliminary understanding of the online illegal gambling market, the features of operators within it, and the motivations of those that gamble with them.
A quantitative survey was then conducted with 1,000 individuals who had gambled online in the last 4 weeks (excluding National Lottery, or in-person, only), in which respondents were asked whether they had engaged in any types of behaviour that may indicate use of an illegal gambling website (such as gambling using a Virtual Private Network (VPN), or gambling on a ‘foreign’ website), and whether they had any experiences that are typical of some illegal gambling websites (such as glitchy websites, or difficulties withdrawing).
Their responses to these questions, coupled with their consent to re-contact, provided Yonder with a sample of individuals who gamble on illegal gambling websites for follow-up qualitative research, in which their identified behaviours and experiences were used to garner a range of opinions across the sample4 .
The qualitative element of Phase 1 consisted of 10 in-depth online interviews with these illegal gambling website users, as well as 2 1.5 hour mini-focus groups with people with lived experience of previously using illegal gambling websites. Findings from the qualitative phase helped to shape the Commission’s understanding of the various individuals who gamble online that are engaged in and/or attracted to the illegal market, and their experiences. This phase of research also served to demonstrate that there is not often a clear understanding among individuals who gamble online of whether they are using an illegal operator.
Although it provided the Commission with a richer understanding of the online illegal gambling market, Phase 1 was limited in that it used a broad stroke approach to categorising and recruiting respondents for qualitative research. Statements used in this phase were not suitable for accurately understanding incidence of consumer engagement in the market, due to the room for legal website usage within some (such as use of a glitchy website, which could have been the case on a website licensed by the Commission, or difficulty withdrawing funds, which is also sometimes the case on licensed websites).
Phase 2 was needed to validate findings from the first phase, and to also develop a core set of indicators that could be taken forward into the Gambling Survey for Great Britain (GSGB), in order to provide an accurate picture of the engagement with illegal gambling websites over time.
4 For more insight on how these individuals were identified as having used illegal gambling websites, see the findings of the Phase 1 project report.
Based on the qualitative findings from Phase 1, and further consultation with key stakeholders, Yonder and the Gambling Commission developed an adapted set of indicators of illegal gambling website usage to enter quantitative fieldwork with.
Following desk research and an initial draft of the indicators, the Commission’s Lived Experience Advisory Panel (LEAP) and Industry Forum, were convened to investigate indicators of engagement with the online illegal gambling market. Members of Yonder’s research team attended a 1-day workshop on illegal gambling at the Commission’s office, with a 1-hour session during the day set aside for an ‘at-scale cognitive test’ with all representatives, from both Industry Forum and LEAP.
The objective of the discussion was to conduct a rigorous and comprehensive analysis of engagement in with illegal gambling websites, and the types of experiences that fall within it, to inform how it can best be identified in a survey environment. By bringing together diverse perspectives, the panel and forum aimed to develop insights that would contribute to more effective and sensitive approaches to understanding illegal gambling. While LEAP were able to speak to their experiences of using illegal gambling websites and understanding from a gameplay, marketing, and engagement perspective, Industry Forum representatives added value by discussing how regulation has a part to play in the shape and scope of the illegal market.
Following the LEAP and Industry Forum workshop from Phase 2, and input from an academic from the University of Glasgow, Yonder finalised a list of indicators of illegal gambling website usage, as well as 4 bespoke questions that would sit outside the indicator list (concerning intentionality of usage and Gamstop circumvention, respectively), to take forward into the quantitative study.
An online survey was developed and conducted via Yonder Data Solutions’ proprietary online panel. The sample for the survey was 2,000 adults in Great Britain aged 18 years and over who had participated in any online gambling activity in the past 4 weeks, to ensure that questions were posed to a relevant audience, and those with timely recall of their behaviours and actions. Given the nature of the statements, it was decided that those who only used lottery products (that is, National Lottery draws or scratchcards) would not be included in the survey sample.
Quotas were set on age, gender, work status, and region to ensure the sample was designed to be broadly representative of the population. Weighting was then applied to the final achieved sample to ensure the key demographic groups are proportionally reflected in the final data, improving its accuracy and representativeness5 .
Phase 2 survey questions relevant to the objective of assessing prevalence of consumer engagement with illegal gambling websites can be found in the following sections. A brief description and rationale for the inclusion of each indicator is explained in the following paragraphs.
Asks all respondents if they have ever intentionally used an illegal gambling website to gauge the level of awareness consumers may have about their engagement with illegal gambling websites.
If respondents report that they have intentionally used an illegal gambling website, asks them to name (at least) of these websites. This question was included to verify individuals’ responses to Question A1.
Ask all respondents whether they have ever signed up to Gamstop.
Asks those respondents who have ever signed up to Gamstop (or are currently signed up) whether Gamstop prevented them from using gambling websites with their own account(s). All gambling websites licensed by the Commission are required to participate in Gamstop. This question aimed to identify respondents who have used gambling websites whilst being registered to Gamstop, indicating that they used illegal gambling websites.
Asks all respondents whether they have engaged in specific behaviours that may indicate their engagement with illegal gambling websites. Phase 1 identified that many consumers are not aware of their use of illegal gambling websites or do not know how to identify such a website. The specific behaviours we asked about can to some extent be considered ‘proxy indicators’ for engagement with an illegal gambling website as they relate to common characteristics or experiences consumers may have on these websites.
All respondents were asked about their intentional usage of illegal gambling websites through a multiple-choice question (see the question in the following paragraphs).
(Ask all respondents. Show intro text on a separate screen, allow respondents to proceed to the question after a 5 second delay.)
In Great Britain, gambling websites are regulated by the Gambling Commission. As the regulator, the Commission provides licences to companies allowing them to operate in Britain if they meet a set of rules and requirements.
Sometimes, however, there are some gambling companies that operate gambling websites in Great Britain without having a licence from the Commission. These websites may have a licence to operate in a different country, or not have one at all, from any country.
(Question shown on next page)
Have you intentionally gambled on a gambling website that did not have a licence to operate in Great Britain?
We want to assure you that the information you provide in this survey is completely confidential, and that individuals who gamble with unlicensed websites are not engaging in illegal activity.
Ask respondents who answer ‘Yes, in the last 4 weeks’
Which gambling website(s) have you used in the last 4 weeks that you believe do not have a licence from the Gambling Commission?
Everything you tell us is completely confidential, we’ll only use this information to check the licensing status of these companies for research purposes. Please enter your answers in the boxes below. You do not need to fill all the boxes to go to the next question.
(Open text box, Respondents can choose to enter a maximum of 5 websites.)
99. Prefer not to say
To ensure respondents’ understanding of the illegal market was appropriate, a definition of illegal gambling websites which highlights the role of licensing by the Gambling Commission was included, which also outlined that some illegal websites in Great Britain could be licensed in other jurisdictions.
The definition was included to ensure a base-level understanding among respondents of what the Commission’s role is within the gambling market, and what it means to be a ‘licensed’ website within this context. The full definition of the Commission’s role, and how licensing is decided on an operator-by-operator basis, was shortened due to survey design principles of not overloading respondents, and keeping the information provided to them concise, engaging and understandable.
Respondents were also assured about the confidentiality of the information they provide during the survey, and this question in particular, to attain the most accurate responses possible.
Where respondents reported they had intentionally used an illegal website in the last 4 weeks, they were prompted to name the website; we used this to verify their responses against the Commission’s list of licensed operators, with two-fold aims: firstly, to gauge the level of understanding among respondents; secondly, to re-code the data where necessary to ensure utmost accuracy.
To determine the usage of Gamstop, respondents were asked if they had ever signed up to it, and the survey outlined how it restricts access to gambling websites based on personal details. The survey assessed whether Gamstop entirely stopped respondents from using gambling websites with their own account. The specific wording was chosen to ensure respondents only report incidents where they were able to use their own personal details to gamble despite Gamstop, serving as a strong indicator they had used an illegal website to do so.
(Ask all respondents)
Have you ever signed up to GAMSTOP, an online tool that restricts access to gambling websites based on personal details you share?
This tool is sometimes called ‘self-exclusion’ and applies across all licensed gambling websites.
(Asked to all those who reported being signed up currently or in the past to Gamstop)
Does or did GAMSTOP entirely prevent you from using gambling websites with your own account?6
We have compiled the following list of indicators of illegal gambling to assess whether respondents have gambled on illegal websites unintentionally. The indicators mark features that would not be available on licensed websites, thus, if engaged in, suggesting a person has gambled on an illegal website.
(Ask all respondents)
(Randomise statements)
Have you ever:
To assess which behaviours are associated with each other, responses to the indicators were plotted against each other in a grid. This allowed us to see which indicators were primary and secondary, by highlighting those that had broad reach across other behaviours, and those that were more distinct (that is, where a respondent had engaged with the specific behaviour, they were unlikely to have engaged in others).
Credit card deposits were determined to be a secondary indicator, as where multiple indicators were chosen, credit card depositing was often among them. Other crossovers that stood out include using cryptocurrency to gamble, which had a high correlation with signing up to gambling websites that didn’t require verification. The most convincing singular indicator of engagement with the illegal gambling market, based on correlation, is the depositing of NFTs: where respondents identified this behaviour, they were often likely to select others in combination.
We asked the respondents who reported intentional use of illegal gambling websites about the specific website(s) they used. The verbatims for the question suggest the understanding of the illegal market is poor, with around half of respondents giving names of websites that are licensed. In addition, understanding of the mechanics of cryptocurrency and/or NFT gambling seemed poor, with some respondents naming cryptocurrency trading websites, which are out of scope for the Commission.
One of the limitations to this research is people’s understanding of the types of payment cards. As noted in research conducted by MoneySuperMarket (opens in new tab), as much as a quarter of British people do not know the difference between a credit and debit card. This could potentially lead to respondents misreporting their credit card depositing behaviours, decreasing the reliability of credit card usage as an indicator of illegal gambling. In addition, credit card deposits on licensed gambling websites was only prohibited in 2021 by the Commission; respondents may have issues with recall of when they exhibited this behaviour, meaning it could have fell either in or outside of the window in which the behaviour is specifically indicative of usage of an illegal website.
Legislation around reversing withdrawals before they reach banks and/or wallets has also changed relatively recently (2021). As with credit card depositing, those who have reported making a reverse withdrawal, could have done so before the feature was prohibited by the Gambling Commission in 2021, which at the time would not have been an indicator of an illegal gambling website.
The Commission required a distinct, battery list of statements that they could include on their GSGB survey. While the initial question around intentional usage of illegal sites was kept separate to this list of statements, owing to the need to include a detailed description of illegal sites before assessing intentionality, the 2 questions about Gamstop usage and Gamstop circumvention needed to be condensed into 1 statement in order to be included in the battery list of statements, guided by the space available on the GSGB survey. The statement list used is included the next section.
A final draft list of statements was shared with a specialist in gambling research at the University of Glasgow. Her feedback led to the inclusion of some example licensed operators in the question concerning intentional usage, and reframed the language around codes concerning ‘depositing’ cryptocurrency and/or NFTs and/or with a credit card to more commonplace language (‘placed a bet' and/or ‘played a game'). An academic from the Bristol Hub for Gambling Harms also provided an academic review, which led to the addition of a clarification to the credit card statement.
LEAP and the Industry Forum were then invited to comment on the finalised list of statements, to suggest any final tweaks that they saw fit from the consumer and industry angles, respectively. Their feedback focused mostly on avoiding prejudicing responses and creating an environment in which respondents felt comfortable in giving truthful answers. Based on this feedback, the indicator list was moved before the question on intentional usage, to avoid desirability bias, and the word ‘deliberately’ was removed from a code concerning Gamstop usage and circumvention to reduce stigma.
Commission colleagues who had little to no prior knowledge of the research were also invited to feedback on the updated set of draft statements. Their feedback focused largely on respondent comprehension of the questions. As a result of their feedback, the response option of ‘yes, ever’, was changed to ‘yes, more than 12 months ago’. Feedback suggested that this would help to avoid ambiguity in what was meant by ‘yes, ever’ and to promote accurate reporting in timeframes7.
5A full list of quotas and weighting schemes are provided in the appendices.
6 'With your own account' was underlined when presented to respondents to emphasise this element of the survey question.
7 This feedback was also provided by LEAP and Industry Forum.
Following the post-fieldwork analysis phase, 7 statements have been finalised8 and selected for piloting in the GSGB survey, as shown in this section.
(Asked to all respondents)
Randomise statements
Have you ever…
(Drop down definition)
Gamstop is a free online tool that, if you sign up, prevents you from using gambling websites and apps run by companies licensed in Great Britain, for a period of your choosing.
This statement list preceded a question concerning intentional illegal gambling website usage, shown in the next section.
(Asked to all respondents)
In Great Britain, gambling websites are regulated by the Gambling Commission. As the regulator, the Commission provides licences to companies allowing them to operate in Britain if they meet a set of rules and requirements. Examples of some gambling companies licensed by the Gambling Commission include well-known companies such as William Hill, Bet365, Ladbrokes, and Coral.
Sometimes, however, there are some gambling companies that operate gambling websites in Great Britain without having a licence from the Commission. These websites may have a licence to operate in a different country, or not have one at all, from any country.
Question A1. Have you intentionally9 gambled on a gambling website that did not have a licence to operate in Great Britain?
We want to assure you that the information you provide in this survey is completely confidential, and that individuals who gamble with unlicensed websites are not engaging in illegal activity.
(Asked to all respondents who report intentionally using an illegal gambling website in the past 4 weeks.)
Question A2. Which gambling website(s) have you used in the last 4 weeks that you believe do not have a licence from the Gambling Commission?
Everything you tell us is completely confidential, we’ll only use this information to check the licensing status of these companies for research purposes.
8As outlined in later sections of this report, although the analytical approach applied to these questions was modified during piloting, the questions themselves were not revised after the pilot phase and have been retained as the final GSGB questions for ongoing use. The exception to this, is Question A2 which has been removed from the GSGB following piloting, the explanation for which is documented later in this report.
9The word ‘intentionally’ was underlined when presented to respondents, to emphasis this part of the survey question.
To support finalisation of these questions and analysis approach, Yonder hosted a refinement workshop with Gambling Commission colleagues. This involved reviewing the Phase 2 quantitative survey responses, and utilised Yonder’s and the Commission’s extensive survey design experience to assess the strength of each illegal gambling statement, how they might be refined for the Gambing Survey for Great Britain (GSGB), and how they could be combined into a single, trackable metric for GSGB stakeholders.
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:
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.
| 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.
| 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.
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.
This section describes the methodology for maintaining and updating the Gambling Survey for Great Britain (GSGB) Index, including the incorporation of new statements, the removal of existing statements, and revisions to statement strength ratings.
An example of how a statement being upgraded in its strength rating is shown in the following section (the bold figures in the tables signify differences to the original Index). The multiplication factor is scaled down slightly to 5.26 as to reflect the renewed strength of the ‘Not on Gamstop’ code brings the total initial highest score to 19, with 100 over 19, equalling the new multiplication factor of 5.26.
| 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 withdrawal |
|---|---|---|---|---|---|---|---|---|
| Strength of indicator | Near certain | Near certain | Strong | Strong | Weaker | Strong | Near certain | Near certain |
| Initial rating | 3 | 3
| 2 |
2 |
1 |
2 |
3 |
3 |
|
| Multiplication factor | 5.26 | 5.26 | 5.26 | 5.26 | 5.26 | 5.26 | 5.26 | 5.26 |
| In the last 4 weeks | 15.78 | 15.78 | 10.52 | 10.52 | 5.26 | 10.52 | 15.78 | 15.78 |
| In the last 12 months | 7.89 | 7.89 | 5.26 | 5.26 | 2.63 | 5.26 | 7.89 | 7.89 |
| More than 12 months ago | 3.94 | 3.94 | 2.63 | 2.63 | 1.32 | 2.63 | 3.94 | 3.94 |
| Never | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
An example of how you might flex the index to incorporate a new statement, with a ‘near certain’ strength rating, is below. The multiplication factor to 100 is scaled down to 4.55 reflect the now 9 possible indicators feeding in, with a total initial highest score of 22.
| Strength of indicator | A1. Intention | A3S. Not on Gamstop | New indictator | 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 withdrawal |
|---|---|---|---|---|---|---|---|---|---|
| Strength of indicator | Near certain | Near certain | Near certain | Strong | Strong | Weaker | Strong | Near certain | Near certain |
| Initial rating | 3 | 3 | 3
| 2 |
2 |
1 |
2 |
3 |
3 |
|
| Multiplication factor | 4.55 | 4.55 | 4.55 | 4.55 | 4.55 | 4.55 | 4.55 | 4.55 | 4.55 |
| In the last 4 weeks | 13.65 | 13.65 | 13.65 | 9.10 | 9.10 | 4.55 | 9.10 | 13.65 | 13.65 |
| In the last 12 months | 6.83 | 6.83 | 6.83 | 4.55 | 4.55 | 2.23 | 4.55 | 6.83 | 6.83 |
| More than 12 months ago | 3.42 | 3.42 | 3.42 | 2.27 | 2.27 | 1.12 | 2.27 | 3.42 | 3.42 |
| Never | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
And finally, an example of how you might flex the index if one of the ingoing statements is dropped. Here, we’ve dropped the ‘intention’ input from the index calculation. The multiplication factor to 100 is scaled up to 6.25 to reflect the now 7 indicators feeding into the index, with a total initial highest score of 16.
| Strength of indicator | A3S. 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 withdrawal |
|---|---|---|---|---|---|---|---|
| Strength of indicator | Near certain | Strong | Strong | Weaker | Strong | Near certain | Near certain |
| Initial rating | 3 | 2 | 2 | 1 | 2 | 3 | 3 |
| Multiplication factor | 6.25 | 6.25 | 6.25 | 6.25 | 6.25 | 6.25 | 6.25 |
| In the last 4 weeks | 18.75 | 12.5 | 12.5 | 6.25 | 12.5 | 18.75 | 18.75 |
| In the last 12 months | 9.38 | 6.25 | 6.25 | 3.13 | 6.25 | 9.38 | 9.38 |
| More than 12 months ago | 4.69 | 3.13 | 3.13 | 1.57 | 3.13 | 4.69 | 4.69 |
| Never | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
While Phase 2, its findings, and its outputs for the Gambling Survey for Great Britain (GSGB), have been subject to rigorous academic and stakeholder input and analysis, there is scope for more in-depth analytical approaches to strengthening findings.
A factor analysis could be employed to distil the indicators into a set of themes, while also allowing us to see the greatest correlation with a pre-set construct of ‘illegal usage’. This technique derives attributes from the data that are similar to underlying imaginary constructs called factors and assigns a score to each attribute which explains the importance towards that factor. Factor analysis gives us additional layer of information in order for us to make the most statistically robust decision.
In conjunction with importance and performance scores, a factor analysis gives further credence to the scores and strengths assigned to the variables for index creation.
The primary aim of this study was to give an overview of the most pertinent indicators of illegal gambling. Future use of the data could deep dive into how the index score and individual statements within it differ for different consumer audiences. For example, there may be value in an assessment of the most prevalent indicator of illegal gambling for those with a higher propensity to gamble, or with a higher Problem Gambling Severity Index (PGSI) score. Data from the GSGB survey will provide robust sample sizes to analyse different audiences with greater confidence.
While altering the core statements too frequently risks introducing noise into the data, which can undermine the ability to track trends effectively, Yonder have designed this index so it can flex somewhat, given the volatile nature of the illegal market and the need for the Gambling Commission to be flexible in responding to emergent changes in legislation and play behaviours. Changes to the index should therefore be made cautiously and only when there is strong evidence that new themes are both critical and enduring.
When changes are necessary, statements could be added to the list of indicators and sit outside of the index, to begin with, in order for the Commission to take the same data-led approach, supplemented by academic and stakeholder review, to assessing how emergent indicators should feed into the composite index. Once this process has taken place, the index calculation is designed to flex. Overleaf are some demonstrations on how this flexing can take place – either to incorporate a new indicator, to remove one, or to reflect the increased strength of confidence one indicator gives as to a respondent’s engagement with the illegal market.
Yonder delivered a technical report to the Gambling Commission, outlining the question development process and index calculation instructions. This report was reviewed and scheduled to be published alongside the first illegal gambling Gambling Survey for Great Britain (GSGB) data release of 4 waves of data. This section of the report details the refinements the Commission made either after delivery of the technical report and/or analyses of the first waves of piloted GSGB data using Yonder’s original Illegal Gambling Index.
The ‘Not on Gamstop’ indicator was originally coded as ‘Strong’ for the Phase 2 quantitative survey, but Yonder and the Commission agreed that it should be upgraded to ‘Near Certain’ for the Illegal Gambling Index. Individuals using a gambling website they knew was not on Gamstop is considered a valuable indicator of consumer engagement with illegal gambling websites, given that all licensed gambling websites are required to participate in Gamstop.
One of the questions piloted in the GSGB asked respondents directly whether they had intentionally used illegal gambling websites. Respondents who reported intentionally gambling on illegal websites in the past 4 weeks were also asked to name the websites they believed to be illegal. The rationale was that this would help to validate whether they had, in fact, used an illegal gambling website. Both questions were tested with key stakeholders during the Phase 2 research and were considered suitable for inclusion in the GSGB.
However, following initial analysis of the first two waves of the piloted GSGB questions, only a small number of respondents reported intentionally gambling on illegal websites. When those who reported doing so in the past 4 weeks were asked to name the illegal gambling website they had used, an even smaller number provided the names of websites they believed to be illegal, with the vast majority selecting ‘prefer not to say’.
For the handful who did provide a website name, analysis showed that many of the websites named were, in fact, licensed gambling operators. This finding is consistent with previous phases of this research, which shows that consumers often lack clarity about what constitutes a licensed versus an illegal gambling website. Furthermore, respondents who selected ‘prefer not to say’ were still able to be identified as potential users of illegal gambling websites through the remaining 7 indicators, meaning the 7 indicators alone provided a sufficiently robust basis for constructing the index, and the intentionality questions were not incorporated into the final index scoring methodology.
Furthermore, the follow-up question (Question A2) asking respondents to name a website was also removed from the question listed due to its limited additional value and redundancy, as all of the individuals identified as having used illegal gambling websites using this question were also captured through the index approach using the remaining 7 indicators.
The removal of the main intentionality question (Question A1) from the index (meaning that the index is now calculated using 7 statements rather than 8) resulted in the threshold for a respondent to be considered as having engaged with illegal gambling websites in the past 4 weeks from 16.7 to 18.75, as is demonstrated in the earlier section Flexing the index for the GSGB.
It should be noted that although Question A1 was removed from the index calculation, the question itself was retained in the GSGB.
Of the 7 indicators taken forward for the final Illegal gambling Index, 3 have been assigned a ‘Near Certain’ rating, 3 have a ‘Strong’ rating, and 1 is deemed to be ‘Weaker’.
Index scores are assigned in relationship to one another, based on strength. For calculating the final Gambling Survey for Great Britain (GSGB) illegal gambling index:
The following indicators were assigned a ‘Near Certain’ strength rating.
Used a gambling website that was not on Gamstop – This was deemed a near certain indicator because all online licensed gambling operators in Great Britain are required to participate in Gamstop. If a gambling website is not registered with Gamstop, it is operating illegally.
Signed up to a gambling website because you knew you wouldn’t need to verify your age or identity - Using a website that does not verify age or identity is a near certain indication that the website is not being ran by a licensed operator, as age and identity verification are fundamental requirements for Gambling Commission licensed operators.
Cancelling a withdrawal before it reached their bank and/or wallet on a gambling website – This is a feature that can be available on illegal gambling websites. Remote Technical Standards (RTS) 14B, introduced in 2021 prohibited gambling operators from allowing customers to cancel withdrawal requests following evidence that this practice was associated with increased gambling-related harm. As the requirement has been in place for several years now, this was considered a near-certain indicator of engagement with an illegal gambling website.
The indicators assigned a ‘strong’ strength rating were as follows.
Used a Virtual Private Network (VPN) specifically to access a gambling website that was not otherwise available in Great Britain – This was considered a strong indicator because the use of a VPN to access a website that would otherwise be blocked in Great Britain, whether a website licensed by another country’s gambling regulator or an entirely illegal website, provides a strong indication of illegal gambling. It was classified as strong, rather than near certain, due to the possibility of inaccurate self-reporting by respondents who had used VPNs to access websites licensed within Great Britain (for example, to bypass workplace internet restrictions).
Placed a bet and/or played a game on a gambling website using cryptocurrency (for example, Bitcoin) and Placed a bet and/or played a game on a gambling website using Non-Fungible Tokens (NFTs) or virtual assets – These indicators were considered strong, rather than near certain because respondents confuse NFT or cryptocurrency trading with gambling, as evidence by one verbatim response in earlier research.
The indicator assigned this strength rating was Placed a bet and/or played a game on a gambling website using money directly from your credit card (that can be paid off each month). It is important to consider the legislative context for this indicator. In 2020, the Commission banned operators from allowing customers to deposit on a gambling website using a credit card, meaning that respondents reporting using a credit card to gamble may be doing so on illegal gambling websites. However, it is possible that respondents could also be misunderstanding credit card versus debit card use.
Index scores are halved for each indicator if identified in the last 12 months (rather than in the ‘past 4 weeks’ ago) – then halved again if identified further back – that is, ‘more than 12 months ago’, to demonstrate the decreasing confidence in response based on poorer recall.
| Indicator | A3S. 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 withdrawal |
|---|---|---|---|---|---|---|---|
| Strength of indicator | Near certain | Strong | Strong | Weaker | Strong | Near certain | Near certain |
| Yes, in the past 4 weeks | 3 | 2 | 2 | 1 | 2 | 3 | 3 |
| Yes, in the past 12 months | 1.5 | 1 | 1 | 0.5 | 1 | 1.5 | 1.5 |
| More than 12 months ago | 0.75 | 0.5 | 0.5 | 0.25 | 0.5 | 0.75 | 0.75 |
| Never | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
From this initial stage, as you can calculate by summing the top row, the maximum feasible index score when summing all the most recent indicators is 16. To create an easy-to-comprehend, single metric for the Gambling Commission with a range of 0 to 100, we have then scaled up the index from a maximum of 18 to 100 (by multiplying 16 – the maximum score - by 6.2511 ) so that the maximum feasible score – that is, if a respondent identified all seven indicators within the past 4 weeks – is 100.
So, given the strength ratings assigned and the maximum combination of these totalling 100, the index calculations are as follows.
| Indicator | A3S. 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 withdrawal |
|---|---|---|---|---|---|---|---|
| Strength of indicator | Near certain | Strong | Strong | Weaker | Strong | Near certain | Near certain |
| Multiplication factor | 6.25 | 6.25 | 6.25 | 6.25 | 6.25 | 6.25 | 6.25 |
| Yes, in the past 4 weeks | 18.75 | 12.50 | 12.50 | 6.25 | 12.50 | 18.75 | 18.75 |
| Yes, in the past 12 months | 9.38 | 6.25 | 6.25 | 3.13 | 6.25 | 9.38 | 9.38 |
| More than 12 months ago | 4.69 | 3.13 | 3.13 | 1.57 | 3.13 | 4.69 | 4.69 |
| Never | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
11 This multiplication factor may change to respond to new indicators being built into the Index in future, or the upgrading and/or downgrading of the strength of the current indicators in use.
The Gambling Commission will discuss the illegal gambling findings from the Gambling Survey for Great Britain (GSGB) using 4 respondent categories.
Respondents who demonstrate at least 1 ‘Near certain’ indicator or 2 ‘Strong’ indicators within the past 4 weeks. This corresponds to an index score of 18.75 when considering the past 4 weeks’ responses only.
Where the above threshold isn’t met, indicators reported within the past 12 months are incorporated at a progressively lower strength rating to reflect the reduced recency of the reported behaviour. Respondents whose score then reaches the index threshold of 18.75 are assigned to this qualifying timeframe.
This includes respondents who report 1 or more indicators, but whose combination of indicator strength and recency does not reach the qualifying criteria for being categorised as an illegal gambling website user in the past 4 weeks or past 12 months. This may reflect either a score of below the 18.75 index threshold, or indicators that are historic (more than 12 months ago).
Within the GSGB, respondents were offered a ‘More than 12 months ago’ response option to ensure that those whose behaviour occurred several years earlier had an appropriate response available. However, because these responses relate to less recent activity and are more susceptible to recall error, they should be treated with caution.
Additionally, some indicators related to regulatory changes for licensed gambling websites that occurred years ago, such as the banning of credit card gambling and customers being able to cancelling the withdrawals. These changes were introduced in 2020 and 2021 respectively, so some respondents may report engaging in these behaviours prior to the bans occurring. For example, a respondent who recalls using a credit card on a gambling website 10 years ago may select ‘More than 12 months ago’, but this behaviour would not necessarily indicate illegal gambling website use as it occurred before this capability was banned on licensed gambling websites.
Some respondents in this category may achieve the 18.75 index score threshold, even after accounting for the reduced weighting and timeframe of the indicators they reported. However, where all relevant indicators relate solely to behaviours that occurred more than 12 months ago, these respondents remain classified within this category and are not considered illegal gambling website users
Important: Note that although the Commission will publish findings relating to those demonstrating limited indications of engagement with illegal gambling websites, this is primarily for completeness of reporting. Interpretation and discussion should focus on respondents who meet the threshold for likely illegal gambling website use, rather than those in the limited indications category.
Respondents with a score of 0 who demonstrate no indication of engaging with illegal gambling websites.
This report signifies the conclusion of a multi-phase series of Consumer Voice research projects conducted by Yonder, each phase of which has been instrumental in helping to build a complete picture of consumer engagement with illegal gambling websites. With the addition of this workstream, the Commission has moved from a position of limited knowledge of the nature of the market to one of more robust statistical measurement of the consumer engagement within it.
This journey represents a significant milestone in adding to the Commission’s Illegal gambling and crime Evidence Roadmap, supporting the development of a robust evidence base on participation in online illegal gambling in Great Britain. This was only made possible by the pursual of a robust, evidence-led understanding of this complex and often hidden market. The decision to develop the Illegal Gambling Index was one that came out of requiring a comprehensive and adaptable metric through which the Commission can monitor consumer engagement with illegal gambling websites. The Illegal Gambling Index has therefore been intentionally designed to be flexible and may therefore evolve over time as further evidence and data becomes available.
As the gambling landscape changes, new indicators may be added, and existing indicators may be revised or removed. Accordingly, the scoring framework may be recalculated to ensure the index remains robust, relevant, and capable of accurately capturing consumer engagement with illegal websites.
In addition to having the capability to monitor overall consumer engagement with illegal gambling websites, the Commission aims to track the different persona groups within this population. These personas were identified and explored in depth through previous phases of this strand of research. Most recently, Phase 3 provided clear metrics through which prevalence of some of the persona groups can be identified using the Gambling Survey for Great Britain (GSGB). Building on the existing GSGB illegal gambling questions, an additional suite of questions developed within Phase 3 will be incorporated into the GSGB 2027, Wave 1.
| Group | Target (percentage) |
|---|---|
| Male | 53% |
| Female | 47% |
| Aged 18 to 24 years | 10% |
| Aged 25 to 34 years | 19% |
| Aged 35 to 44 years | 17% |
| Aged 45 to 54 years | 18% |
| Aged 55 to 64 years | 17% |
| Aged 65 to 74 years | 13% |
| Aged over 75 years | 7% |
| Working | 70% |
| Non-working | 30% |
| Scotland | 9% |
| North East | 5% |
| North West | 5% |
| Yorkshire and Humber | 9% |
| West Midlands | 9% |
| East Midlands | 8% |
| Wales | 5% |
| East of England | 10% |
| London | 10% |
| South East | 13% |
| South West | 13% |