Dichotomous and Weighted Scoring of the Problem Gambling Severity Index Converge on Predictors of Problem Gambling

In its original development, the Problem Gambling Severity Index (PGSI) compared frequency weighted and dichotomous scoring of items. Although dichotomous scoring yielded a higher correlation with clinical assessments, it also yielded higher estimates of problem gambling prevalence. The aim of the present study was to compare the two scoring methods as potential moderators in the identification of important psychological predictors of problem gambling. Sports (n = 581) and non-sports gamblers (n = 670) completed an online survey containing the PGSI and standardized measures of impulsivity, gambling motivation, and gambling cognitions. Psychometric analyses were run on the weighted and dichotomous scores of PGSI scale. The two types of scoring systems were regressed on cognitive, motivational, and personality factors. Both scoring systems had high internal consistency for weighted and dichotomous scoring, respectively. Important predictors were determined with stochastic search variable selection (SSVS) and dominance analyses. Weighted and dichotomous scoring converged on identifying four important predictors of problem gambling in both sports and non-sports gamblers: positive urgency; amotivation; luck/perseverance; and gambling identity. The results suggest that dichotomous and weighted scoring systems are similar in their psychometric properties and in the identification of important predictors of problem gambling. One practical advantage of dichotomous scoring is the absence of a reliance on the accuracy of judgments of the frequency of gambling behaviors and adverse consequences.