We use unique survey data linked to nearly a decade of administrative welfare data to examine the relationship between early marijuana use (at age 14 or younger) and young people's educational outcomes. We find evidence that early marijuana use is related to educational penalties that are...