Understanding the relationship between regional gambling behaviour and regional differences in personality
- Start date: 1 January 2022
- End date: 31 August 2023
- Principal investigator: Joshua Weller
- Co-investigators: Barbara Summers; Xingjie Wei
Examining the degree to which problem gambling behaviour is associated with regional differences in personality using social media language use.
Worldwide, gambling is a popular activity in which individuals engage in multiple ways, including lottery participation, casino and slots gaming, and sportsbooks. Although gambling is an activity that many people enjoy without problems, for many gambling behaviour becomes disordered, which may lead to adverse financial social, and psychological consequences. For instance, problem gambling is associated with increased debts, dissolution of interpersonal relationships, and losses may exacerbate mood disorders and other externalising behaviours like alcohol and substance abuse.
Regional differences in dispositions are important because they may indicate variations of important political, financial and health outcomes. For instance, regional differences in personality have been associated with individuals’ voting behaviour for Brexit, lower reported health outcomes, and life expectancy. Researchers suggest that these regional differences emerge as a result of numerous processes including social norms and local culture, environmental and industrial influences, as well as selective migratory patterns.
Although numerous studies exist regarding predictors of gambling behaviour , the proposed project goes beyond the state-of-the-art to test the feasibility of extracting language patterns through online gambling communities via social media. Whereas prior studies investigating the associations between personality and gambling behaviour have solely utilised self-report, extracting social media has several advantages. Social media platforms generate millions of messages daily, providing a rich, geo-located dataset that is ideal for studying regional variations in behaviour. This point contrasts with past research on regional personality differences that have utilised large existing datasets, often taking years to develop, and are limited by the measures that were initially administered. Social media language has been found to be a reliable source of dispositional cues. Language use is objective and measurable behavioural data and, unlike surveys and questionnaires, social media language allows us to observe individuals as they freely express themselves in their own words. Online social forums such as Reddit ProblemGambling or GamblingAddiction6, have gathered people who suffered from gambling addiction to talk about their experience and emotions, making these platforms ideal for examining language patterns associated with problem gambling. Creating such a lexicon of language patterns can enhance our understanding on gambling behaviours, and may inform the development of an early-risk detection system, which defines a novel application to prevent problem gambling behaviours.
By understanding regional variations in gambling behaviour, this research has the potential to provide substantial impact in several ways.
Firstly, though regional summary data on gambling behaviour may provide information about the prevalence of gambling in a certain area, the potential mechanisms leading to these differences remain opaque. This is of particular concern given that recent research shows that 21% of gambling premises were based in the most deprived decile of areas of the country, ten times as many as in the least deprived decile. Exploring the associations between regional personality differences, sociodemographics, and gambling behaviour will assist in developing evidence-based targeted interventions that give individuals a mechanism to control their level of problem gambling.
Secondly, these problem gambling language patterns could suggest new hypotheses and give detailed insights, and lead to the building of an early risk detection / problem gambling prevention system for players.
Thirdly, the modelling techniques that we propose for extracting social media language could be extended and modified to understand other problem behaviours, such as depression and substance abuse.