Existing literature suggests that Blacks and to some extent Hispanics are less likely than Whites to use substances in their youth. However, by age 35 their use rates are higher than those of Whites. This has been referred to as the race/ethnic age crossover effect in substance use. However, more research is needed to clarify whether this is true for Blacks and Hispanics, males and females, and for drugs and alcohol. It is also empirically unclear why this crossover effect occurs. The present study explores these issues. The National Household Survey of Drug Abuse is merged for four years (1999-2002) to study the race/ethnic age crossover effect (n = 208,878). The results suggest that the race/ethnic age crossover effect cannot be used to describe Hispanic relative to White substance use. However, it is applicable to Black substance use, in particular illegal drug use for Black males and heavy drinking for Black females. Additional analyses reveal that the crossover effect for Blacks is eliminated when socio-demographic controls are included. With controls, drug use and heavy drinking for Blacks are lower than that of Whites for all age groups and for both males and females. Process differences are also revealed in the effects of the socio-demographic correlates. More specifically, these analyses of composition and process suggest that drug availability/exposure is a particularly important factor which likely contributes to the crossover effect in substance use and deserves further study.