Thresholds of probable problematic gambling involvement for the German population: Results of the Pathological...

Consumption measures in gambling research may help to establish thresholds of low-risk gambling as 1 part of evidence-based responsible gambling strategies. The aim of this study is to replicate existing Canadian thresholds of probable low-risk gambling (Currie et al., 2006) in a representative dataset of German gambling behavior (Pathological Gambling and Epidemiology [PAGE]; N = 15,023). Receiver-operating characteristic curves applied in a training dataset (60%) extracted robust thresholds of low-risk gambling across 4 nonexclusive definitions of gambling problems (1 + to 4 + Diagnostic and Statistical Manual for Mental Disorders-Fifth Edition [DSM-5] Composite International Diagnostic Interview [CIDI] symptoms), different indicators of gambling involvement (across all game types; form-specific) and different timeframes (lifetime; last year). Logistic regressions applied in a test dataset (40%) to cross-validate the heuristics of probable low-risk gambling incorporated confounding covariates (age, gender, education, migration, and unemployment) and confirmed the strong concurrent validity of the thresholds. Moreover, it was possible to establish robust form-specific thresholds of low-risk gambling (only for gaming machines and poker). Possible implications for early detection of problem gamblers in offline or online environments are discussed. Results substantiate international knowledge about problem gambling prevention and contribute to a German discussion about empirically based guidelines of low-risk gambling.