Assessing the factor structure of the Problem and Pathological Gambling Measure (PPGM)

Abstract

The Problem and Pathological Gambling Measure (PPGM) is widely used to assess problem gambling, showing strong correspondence between population surveys and clinical interviews. However, the literature on its factor structure is limited. This study used confirmatory factor analysis to identify the optimal structural model of the PPGM. Data from the Finnish Gambling Harm Survey (2016, n = 3218; 2017, n = 1250) were used. Five models were compared on fit indices, reliability, and criterion validity, including gambling frequency, expenditure, and harm. The models included one-factor, two-factor, three-factor, two-factor with correlated residuals, and bifactor models. Results supported a two-factor model comprising dependence and harm, with adjustments for two items with correlated residuals. The bifactor model had similar fit levels but poorer reliability and replicability. Permutation tests did not support distinguishing between impaired control and other addictive issues. The models showed stronger associations with gambling harm than gambling behavior (expenditure, intensity, breadth). The study concludes that the two-factor model of the PPGM, measuring harm and dependence, is valid, and scores on these factors can serve as quantitative indices of general severity, dependence, and harm.