LAUSR

To examine the underlying dimensionality and structure of problem gambling using a comprehensive range of problem gambling assessments from an international online survey of gamblers. A total of 12,521 gamblers from 105 countries were recruited through banner advertising placed on a popular online gambling portal to take an online survey. Although participants were recruited online, the majority of the sample (71.6%) gambled only at land-based venues in the past 12 months. A total of 5081 individuals completed all items from the four problem gambling assessments. Participants were allocated to answer one of the four problem gambling assessments and the remaining unique items from the three other problem gambling assessments. The order of assessments were counterbalanced. Two optimal scaling procedures were independently employed to estimate the number of dimensions within the data: exploratory categorical principal component bootstrap analysis and multidimensional scaling. Nonlinear canonical correlation was then used to establish how well each of the four assessment instruments captured the identified dimensions. A final confirmatory principal component analysis was run to understand and characterise the nature of the dimensions that were identified. Both the categorical principal component bootstrap analysis and multidimensional scaling indicated the data was multidimensional, with four dimensions (including a single dominant dimension) providing the best characterisation of the data. The nonlinear canonical correlation analysis found that the Problem and Pathological Gambling Measure and the National Opinion Research Center DSM-IV Screen for Gambling Problems operationalization of the Diagnostic and Statistical Manual of Mental Disorders Four (DSM-IV) criteria best captured these multiple dimensions. Confirmatory principal component analysis suggest a core experience of generic problem gambling symptomology and three other components: “financial problems”, “health and relationship issues”, and “difficulty controlling gambling”. Problem gambling symptomology appears to be multi-dimensional. Certain assessments capture this heterogeneity better than others and thereby provide a more complete and accurate assessment of this construct.