The first purpose of this report was to develop distinct construct measures of Cohesion and Conflict that evaluated affective dimensions of parenting style. Constructs were formed from multiple method, agent, and setting perspectives in a bootstrapping approach. The second purpose was to test the hypotheses that Cohesion and Conflict in the family would contribute to changing levels of boys’ Wandering from Grades 4 to 6. The sample was 206 parents and a focal boy participating in the Oregon Youth Study, a longitudinal study designed to study the development and maintenance of antisocial behavior and delinquency in boys. Latent variable path models demonstrated convergent and discriminant validity of the Cohesion and Conflict constructs. The hypotheses were partially supported. Cohesion and Conflict were related to Wandering in the expected directions. In latent variable growth models Cohesion significantly reduced growth in Wandering controlling for parental and child antisocial behaviors.