In recent years there has been a growing interest in applying social network models to the problem of adolescent substance use. However, there has been little research conducted examining the reliability of social network information within this population. The current study attempts to address this gap, specifically by examining social network test-retest reliability over a two week period among a sample of adolescent substance users. The results of the current study suggest that for social network variables dealing with substance use, reliabilities are at least moderate with correlations of.6 or above. However, there is a large degree of turnover with regards to the specific individuals being named in the network with only 62 % of alters mentioned at Time 1 being mentioned at Time 2.