This cross-sectional research study sought to explore how outcome expectancies can be integrated into broader models of youth gambling risk, by evaluating their relationships with two known common determinants of gambling problems: impulsivity and depression. A modified version of the Gambling Expectancy Questionnaire (Gillespie, Derevensky, & Gupta, 2007a) was used to measure the salience of positive and negative expectancy constructs among a large sample of youth (ages 16-21; n = 1,123; males = 488, females = 635). In addition, to further examine the relationship between impulsivity and gambling behaviour among youth, a multidimensional impulsivity construct (Lynam, Smith, Whiteside, & Cyders, 2006) was employed for this research. As it was important to explore the relationships between outcome expectancies, impulsivity and depression from different perspectives of gambling involvement, gambling behaviour was operationalized in two different ways-as gambling participation (non-gambler vs. gambler) and subsequently as gambling severity (social gambler vs. problem gambler). Sequential logistic regression analyses were performed separately for males and females with age and substance use being employed as covariates. Due to a lack of female problem gamblers, the relationships between variables in the prediction of gambling severity were only explored for males. The results indicate that Enjoyment/Arousal expectancies are significant predictors of male and female gambling participation. For females, these expectancies appear to mediate the relationship between Sensation-Avoidance and gambling participation. In contrast, Self-Enhancement expectancies appear to partially mediate the relationship between Positive Urgency impulsivity and gambling severity for males. Thus, the findings suggest that the outcome expectancies involved in decisions to initiate gambling may differ from those that maintain the behaviour. Moreover, the facet of impulsivity that serves as a risk factor for gambling participation may not be the same as the facet of impulsivity that serves as a risk factor for gambling over-involvement. The theoretical and practical implications of these findings are discussed.