Problem Gambling Prevalence by Country

Methodology & Why These Numbers Don't Line Up

Most gambling stats sites publish a ranked list of countries by "problem gambling rate" without saying what was actually measured. That list is more misleading than useful, so this page does the opposite: same data, but every row is labeled with its instrument, threshold, and period, and nothing is force-ranked against anything it isn't compatible with.

• ToolThe screening instrument varies. Great Britain, Sweden, Denmark, Norway, Finland, Ireland, Italy, the Netherlands, Australia, New Zealand, Japan, and Canada use PGSI or CPGI (a closely related, earlier instrument), and France's ICJE is the French version of the CPGI. All of these are validated questionnaires that produce a numeric score. Germany's survey applies DSM-5 diagnostic criteria instead, Hong Kong uses DSM-5 criteria, Singapore uses a DSM-based screen, Spain's health survey uses the two-item Lie/Bet screen, and Switzerland's health survey uses a four-question screen. Australia's headline "15% at some risk" counts everyone scoring 1 or more, so it is not a problem gambling rate; its PGSI 8+ figure is 2.6%. The US survey (NCPG/Ipsos) asks about behavior frequency directly, without a validated scoring instrument, and NCPG itself says it is not a study of the prevalence of gambling addiction.
• ThresholdEven the same tool gets cut off differently. Great Britain's official "problem gambling" figure uses a PGSI score of 8 or higher. The most-cited Swedish figure uses a PGSI score of 3 or higher, which on Great Britain's own bands would be classified as "moderate risk," not "problem." Comparing the two headline numbers directly overstates how close they actually are. Finland and New Zealand also headline a combined PGSI 3+ group, while Denmark, Norway, Ireland, Italy, the Netherlands, Australia, Japan, and Great Britain report a PGSI 8+ figure.
• PeriodSurvey years span eight years. From Italy's 2017-18 survey to Great Britain's 2025 survey. Problem gambling prevalence can move with market conditions (online expansion, new product types), so an old figure isn't just less precise. It may no longer be accurate. Statistics Canada notes that its 2018 data pre-date the 2021 legalization of single-event sports betting.
• ModeHow people are surveyed matters, even with the same tool and cutoff. Ireland's 2019-20 national survey (HRB) interviewed people in person, in their homes, and found 0.3% of people aged 15+ at the PGSI 8+ threshold. ESRI's 2023 survey used the same tool and cutoff but was completed anonymously online by adults aged 18+, and found 3.3%. ESRI attributes most of the gap to survey method and concludes its figure is the better estimate, while noting that all methods have drawbacks and that its own figure may still be too low. The two surveys also differ in sample: a household probability sample versus quota-sampled online panels. Great Britain's Gambling Commission reports that a good part of the gap between online and in-person surveys is due to downward bias from interviewer administration, and it says direct comparisons between its survey and surveys using different methods are not appropriate.

Sourced Prevalence Data

Countries are grouped by region. Each row is shown as its publisher reports it, with the tool, survey mode, and age group in the fourth column.

*1.6% of past-year gamblers (~304,400 people), not of the total adult population. Irish percentages are shares of all respondents (HRB: ages 15+; ESRI: ages 18+), and the people counts are each publisher's own rounded estimates. ESRI calls the PGSI 1-2 and 3-7 bands "some" and "moderate" evidence rather than low and moderate risk, because it says risk labels can mislead.

The four Nordic sources all use the PGSI, but they cover different ages (Denmark 18-79, Finland 15-74, Norway 16-74, Sweden 16-84), years, and collection methods, and Finland's figure is a combined PGSI 3+ group. The Danish Gambling Authority's own report says comparisons between the Nordic countries are subject to reservations. Germany's figure uses DSM-5 diagnostic criteria rather than a PGSI score.

Same Tool, Different Method: Ireland

Ireland has two national surveys, and both measured problem gambling with the PGSI at the same cutoff (8 or higher). They still found a roughly tenfold gap. ESRI, which ran the 2023 survey, says most of the difference is likely down to survey method: the earlier survey interviewed people in person, while ESRI's was completed anonymously online. Its reasoning is that gambling is a sensitive subject and people may be reluctant to be candid with an interviewer.

This is not a measured rise in problem gambling. The two surveys also differ in sample and age group, and each has stated limits: HRB's fieldwork stopped in March 2020 because of COVID-19, at 88% of its target sample, and ESRI's online panels may not represent people without internet access. ESRI's estimate was 3.3% (95% confidence interval 2.5% to 4.0%), and its three panels ranged from 1.2% to 4.4%. As of 28 September 2026, this page has found no newer national prevalence survey for Ireland.

Change Over Time: Repeated Surveys

A trend is only meaningful when the same survey is repeated with the same method. Thirteen sources here have been: Australia (through the HILDA survey), Denmark, Finland, Germany, Great Britain, Hong Kong, Japan, the Netherlands, Norway, Singapore, Spain, Sweden, and the US. Ireland's two surveys use different methods, so they are compared separately above.

Each row compares like with like within one survey; levels are not comparable across rows. AIFS's 2024 pilot found 15% of Australian adults scoring 1 or more on the PGSI, against 10.9% in the 2019 Interactive Gambling Study. AIFS reads this as an increase, but they are separate surveys.

Worth treating cautiously: the source citing this trend is a gambling-industry trade association (BOS), not Sweden's Public Health Agency directly; the underlying data originates from the agency's Swelogs studies (the latest wave is 2021, with about 7,300 respondents aged 16-84). The Danish Gambling Authority's report cites the Agency's 2021 figures and gives the same 1.3% at PGSI 3+ (0.8% moderate plus 0.5% problem), which supports the BOS number. This page still couldn't access the original Swedish report directly. The same BOS report says severe problem gambling (PGSI 8+ or equivalent) stayed between 0.3% and 0.6% over the period. See sources.

Why the Headline Number Understates the At-Risk Population

Great Britain's full PGSI breakdown. "Problem gambling" alone (2.4%) is less than a fifth of everyone showing some level of risk. Australia's pilot survey shows a similar shape: 2.6% at PGSI 8+ within the 15% scoring 1 or more.

Frequently Asked

Sources

1. UK Gambling Commission, Gambling Survey for Great Britain (GSGB): Annual Report 2025, published 16 Jul 2026. The 2023 and 2024 figures come from the annual reports for those years, published 25 Jul 2024 and 2 Oct 2025 (2023, 2024). gamblingcommission.gov.uk
2. BOS (Swedish Trade Association for Online Gambling), Sweden problem gambling rate report, published 31 Mar 2026, citing Sweden's public-health gambling studies. igamingbusiness.com
3. Statistics Canada, Who gambles and who experiences gambling problems in Canada, published 9 Aug 2022, 2018 Canadian Community Health Survey, Gambling Rapid Response module. statcan.gc.ca. Sports betting note: The Daily.
4. Australian Institute of Family Studies / Australian Gambling Research Centre, National Gambling Prevalence Study Pilot 2024: Key findings, published 24 Sep 2025. aifs.gov.au
5. Australian Institute of Health and Welfare, Gambling (Australia's welfare topic summary), released 16 Oct 2025, reporting the HILDA Survey (Wilkins et al. 2024). aihw.gov.au
6. National Council on Problem Gambling, NGAGE 3.0 survey (with Ipsos), fielded Jan-Mar 2024, published 17 Jul 2025. ncpgambling.org Survey page: ncpgambling.org.
7. Health Research Board (Ireland), Gambling in the Republic of Ireland: Results from the 2019-20 National Drug and Alcohol Survey, published 9 Feb 2022. hrb.ie
8. Economic and Social Research Institute (ESRI), Measures of Problem Gambling, Gambling Behaviours and Perceptions of Gambling in Ireland (Research Series 169), published 5 Oct 2023. esri.ie
9. Danish Gambling Authority (Spillemyndigheden), Survey of the Prevalence of Gambling and Gambling Problems in Denmark 2021 (Rambøll Management Consulting), published May 2022, unofficial English translation. It also gives the 2015 to 2021 Swedish and Norwegian PGSI comparison figures used here, taken from the Public Health Agency of Sweden (2022) and Pallesen et al. (2020). spillemyndigheden.dk
10. Pallesen et al. (2023), Norwegian Gaming Authority prevalence survey for 2022, as reported in Journal of Gambling Studies (2025). This page cites the peer-reviewed article rather than the original Norwegian report. springer.com
11. Finnish Institute for Health and Welfare (THL), Gambling among Finns 2023 (Statistical report 15/2024), published 16 Apr 2024. thl.fi
12. Observatoire français des drogues et des tendances addictives (OFDT), Les jeux d'argent et de hasard en France en 2024, published Aug 2025, reporting the EROPP 2023 survey. ofdt.fr
13. Autorité nationale des jeux (ANJ), L'algorithme développé par l'ANJ révèle un nombre de joueurs excessifs inquiétant et en croissance, 13 May 2026. anj.fr
14. Buth S, Meyer G, Rosenkranz M, Kalke J, Glücksspielteilnahme und glücksspielbezogene Probleme in der Bevölkerung: Ergebnisse des Glücksspiel-Survey 2025 (ISD Hamburg and University of Bremen), published Mar 2026. The survey is funded by the Deutscher Lotto- und Totoblock, and the institutes hold the publication rights. isd-hamburg.de
15. Ministry of Health / Health New Zealand, New Zealand Gambling Survey 2023/24, results published on the Kupe data explorer (PGSI 3+ and PGSI 1+ indicators). The Ministry notes the survey does not form a time series with the earlier Health and Lifestyles Survey. kupe.healthpromotion.govt.nz
16. Kurihama Medical and Addiction Center, national surveys on gambling disorder (2020 and 2023), as summarized in the UMIN clinical trial registry, record R000067975. This is a registry summary, not the full Japanese report. umin.ac.jp
17. Ipsos I&O for the Wetenschappelijk Onderzoek- en Datacentrum (WODC), Deelname aan kansspelen in Nederland: meting 2025 (report 2025/102), June 2025. Fieldwork 7 to 22 March 2025 in the I&O Research Panel (n=7,608, ages 16+, weighted). ipsos-publiek.nl
18. Spain: Ministry of Health (Plan Nacional sobre Drogas), report on behavioral addictions from the EDADES 2024 survey, published 27 Jan 2025 (press release), with the 1.4% figure as reported by Infocop and the 2022 comparison of 1.7% as cited by the Ministry of Consumption. Regulator DGOJ's 2022-23 prevalence study is described in the sidenote. This page cites the secondary reports rather than the ministry's full document.
19. National Council on Problem Gambling (Singapore), Report on Survey on Participation in Gambling Activities Among Singapore Residents 2023, released 28 Nov 2024 (3,007 residents aged 18+, July 2023 to March 2024). ncpg.org.sg
20. Istituto Superiore di Sanità (ISS), Gioco d'azzardo nella popolazione adulta: studio epidemiologico trasversale di tipo osservazionale (Rapporto ISTISAN 19/28), published 2019; fieldwork Aug 2017 to Jan 2018 by face-to-face computer-assisted interviews. publ.iss.it and usciredalgioco.iss.it
21. Federal Office of Public Health (FOPH) and Obsan, At-risk or problematic gambling behaviour (age: 15+), MonAM indicator from the Swiss Health Survey 2022 (about 22,000 respondents), updated 28 Mar 2025. ind.obsan.admin.ch. The federal gambling authorities' 2022 study is described in this summary of the 31 Oct 2024 release.
22. Home and Youth Affairs Bureau (Hong Kong), press release on the report of the Ping Wo Fund study by the Social Sciences Research Centre of the University of Hong Kong (2021 data), 30 Dec 2022. The 2016 comparison comes from the Ping Wo Fund's 2017 release. info.gov.hk