The unhappiness of problem gamblers

Abstract

In 2011, the UN General Assembly passed a resolution inviting member states to gather data to capture the importance of the pursuit of happiness, with a view to these measures guiding their public policies. The latest edition of the British Gambling Prevalence Survey is the first large-scale social survey in the World to have included a question on happiness alongside detailed questions about gambling behaviour; and it also administered to respondents two of the most commonly employed screens for problem gambling. In the present paper, data from the Survey are analysed with an eye to implications for public policy. Statistical modelling reveals that problem gamblers typically exhibit exceptionally low wellbeing, comparably low with those reporting very bad physical health, even in the presence of a rich variety of controls for life circumstances. In the case of the DSM-IV screen, there proves to be a similarly strong association between low wellbeing and ‘at risk’ problem gambling (but only for women). In the case of the PGSI screen, low wellbeing is evident for both men and women amongst those who score any points on the screen, even when their total score is well below the threshold for classification as a problem gambler. Further, respondents who report having a relative with a gambling problem also tend to exhibit significantly lower wellbeing than would be expected from their personal characteristics and situation. Together all these findings suggest that significant expenditure on therapy may be justified for individuals who test positive for ‘problem gambling’. They further suggest that the thresholds used to estimate the number of ‘problem gamblers’ may be conservatively high, leading to underestimation of the number of people affected by gambling issues. The paper goes on to explore the association between low wellbeing and endorsements of individual items on the DSM-IV screen and finds that the patterns of which questions predict unhappiness are very different between men and women. Together, these extra findings are relevant to development of short screens and also raise issues about whether interpretation of screen results should be different according to gender. Finally, the analysis finds, particularly for men, that gamblers who have no signs of a gambling problem tend to report modestly higher levels of wellbeing than people who do not gamble at all, again controlling for life circumstances.