Although there are many cross-sectional studies of adolescent gambling, very few longitudinal investigations have been undertaken. As a result, little is known about the individual stability of gambling behaviour and the extent to which behaviour measured during adolescence is related to adult behaviour. In this paper, we report the results of a 4-wave longitudinal investigation of gambling behaviour in a probability sample of 256 young people (50 % male, 50 % female) who were interviewed in 2005 at the age of 16–18 years and then followed through to the age of 20–21 years. The results indicated that young people showed little stability in their gambling. Relatively few reported gambling on the same individual activities consistently over time. Gambling participation rates increased rapidly as young people made the transition from adolescence to adulthood and then were generally more stable. Gambling at 15–16 years was generally not associated with gambling at age 20–21 years. These results highlight the importance of individual-level analyses when examining gambling patterns over time.