Group climate refers to the atmosphere that is both created and experienced by a group of people. Theoretically, the "climate" of a group should therefore be related to the behaviors of its members during group sessions; however, this has not been empirically tested. A longitudinal design was used to examine the relationship between group member perceptions of engagement and conflict and group member intimate behaviors during the group session. Data came from 73 incarcerated women participating in 6 manualized (Trauma Recovery and Empowerment Model) therapy groups. At each session, group members completed measures of their perceptions of the group's engagement and conflict (the Group Climate Questionnaire), and the group leader completed a behavioral count of each member's intimate behaviors. Group climate was examined both for the group member (actor) and aggregated across the other group members, excluding the actor (partner). In addition, these actor and partner scores were decomposed into between-member and within-member components. None of the between-member differences in engaged or conflict perceptions were significantly related to intimate behaviors. There was a significant positive relationship between actor's and partner's engagement ratings and partner's conflict ratings in a session and the actor's intimate behaviors in that session. Study results highlight the importance of examining changes that may be occurring from session to session in group climate perceptions and the need to examine partner's perceptions in order to best understand how an individual group member will behave.