This study examined the strength of tracking sport participation from childhood to early adulthood among the Dunedin Multidisciplinary Health and Development Study cohort. Participation in sport, dance, or gymnastics as part of a club or group (outside of school) was assessed at ages 7, 9, 15, 18, and 21 years. In addition to the traditionally used correlation coefficients, summary statistics (intraclass correlations; ICC) from random effect models and stability coefficients from generalized estimating equations (GEE) were calculated using all the longitudinal data and controlling for the influence of covariates on tracking strength. Correlation coefficients revealed statistically significant tracking of club sport participation (7-21 years) at low levels (r = 0.07-0.28). The ICC summary statistic (0.23) was consistent with this, while the GEE suggested moderate tracking (0.59). The results of this study suggest that encouraging sport participation during childhood and adolescence may result in a modest increase in the likelihood of participation later in life. However, the substantial movement into and out of sport participation observed here and in other studies cautions against relying solely on sport promotion among youth as a strategy to promote lifelong participation. (Contains 3 tables.)