Abstract The present investigation undertook a large-scale replication of an apparently successful group cohesion manipulation reported by Etringer, Gregory, and Lando (1984). A 2 × 2 factorial design was initiated in which subjects were to be randomly assigned to enriched or standard cohesion and to satiation or nicotine fading. However, subject resistance to the satiation technique necessitated the abandoning of this part of the experimental procedure. Subjects assigned to nicotine fading (enriched or standard cohesion) included 184 women and 169 men. In contrast to the Etringer et al. study, there was virtually no evidence of a treatment effect on either group cohesion scores or abstinence outcome. Future work might attempt to establish more powerful manipulations of cohesion, which could then be applied especially to groups that fail to coalesce on their own.