This exploratory study investigates whether associations between social network measures and substance use differ according to type of substance and social context. The analyses use data obtained from 13 and 15 year olds (N=3146) in a school-based survey and focus on three social network measures: sociometric position (e.g. group, dyad, isolate); popularity (friendship nominations received); and expansiveness (nominations made). The study aims to investigate: (i) the patterning of these social network measures by age, gender and school socioeconomic status (SES); and (ii) their relationships with substance use (current smoking, experience of drugs, alcohol in the previous week). Females and those from higher SES schools were more likely to be in groups, more popular and more expansive. Analyses including all three network measures, together with age, gender and school SES found main effects of sociometric position on smoking (lower than average rates among group members, higher than average for pupils in dyads and among isolates), of popularity on drugs and alcohol (highest levels among the most popular pupils) and of expansiveness on alcohol (highest rates among the most expansive). Interactions between the network measures, age, gender, school SES and substance use were also examined. We discuss our results concerning the use of different substances in relation to the mechanisms of selection and influence (as suggested by sociometric position, popularity and expansiveness) in differing social contexts represented by gender and school SES. (Contains 1 figure and 4 tables.)