Effects of peer network interactions on adolescent cannabis use

Research output: Contribution to journalArticlepeer-review

4 Citations (Scopus)
389 Downloads (Pure)


This study capitalises on three waves of longitudinal data from a cohort of 4351 secondary school pupils to examine the effects on individuals’ cannabis use uptake of both peer cannabis use and position within a peer network.

Both cross-sectional and individual fixed effects models are used to estimate the effect on cannabis use of nominated friends’ cannabis use, of reciprocity and transitivity of nominations across the friendship cluster, and of interactions between these nominated friends. Post hoc analyses parsed the behaviour of reciprocating and non-reciprocating friends.

Cannabis use varied depending on the stability of friendship network and the degree of reciprocity and interconnectedness within the group. Behavioural influence was strong, but interaction effects were observed between the prevalence of cannabis use among friends, the structure of the friendship group and ego’s proximity to group members. These interactions demonstrate that behavioural influence is more salient in more cohesive groups. When reciprocating and non-reciprocating friends’ mean cannabis use were separated, influence from reciprocating friends was estimated at twice the magnitude of other friends.

While preventing any one individual from using cannabis is likely to have a multiplier effect on classmates, the bonds and interactions between classmates will determine which classmates are affected by this multiplier and the salience of that effect.
Original languageEnglish
Pages (from-to)75-91
Number of pages17
JournalJournal of Criminal Psychology
Issue number2
Publication statusPublished - 2015

Bibliographical note

Special Issue: Contemporary multivariate statistics


  • Peer Influence
  • Cannabis
  • Social Networks
  • Adolescent


Dive into the research topics of 'Effects of peer network interactions on adolescent cannabis use'. Together they form a unique fingerprint.

Cite this