The Gambling Urge Scale: Reliability and validity in a clinical population

The aim of this study was to establish reliability and validity of the Gambling Urge Scale (GUS) in a clinical population of problem gamblers. This cohort study was conducted in South Australia between March 2008 and March 2009. Participants were problem gamblers aged >= 18 years (n = 158) who were seeking treatment from a range of gambling help services. Measures included gambling urge, problem gambling screening, gambling behaviour and problems caused by gambling, such as personal health and relationships. The psychometric properties investigated were internal reliability, criterion-related validity, concurrent validity and construct validity. Results showed high internal consistency for GUS (alpha = 0.93) and significant item-rest correlations ranging from 0.72 to 0.86. For criterion-related validity, a GUS cut score of three correctly classified 81.13% of participants as problem gambling with sensitivity 84.75% and specificity 76.6%. Concurrent validity was significant with a number of gambling-related symptoms and problems including psychological disturbance, work and social functioning and gambling-related cognitions (p < 0.001). An insignificant correlation was found between gambling urge and sensation seeking traits (p = 0.663). When controlling for gender and age the instrument was shown to have significant predictive properties for different levels of gambling severity (p < 0.001). A principal component analysis for the one component showed an overall explained variance of 75.54%. These findings indicate that GUS is a valid and reliable instrument for problem gambling screening, to measure treatment outcomes and may predict relapse in problem gambling.