Using a longitudinal mediation framework and a low-income sample, this study had 2 aims: (a) to model bidirectional associations between parent academic expectations and child academic outcomes from first through fifth grade, and (b) to explore 3 mediators of parental influence: parent involvement in child schooling, child learning behaviors, and child perceived academic competence. Participants included 356 children and their caregivers (89% mothers) recruited from Head Start centers (58% European American, 25% African American, 17% Latino). At each time point (grades 1, 2, 3, 5), parents rated their academic expectations, teachers rated parent involvement and child learning behaviors, and children rated their self-perceptions of their academic competence. Bidirectional longitudinal associations emerged between parent academic expectations and child academic outcomes. Child learning behaviors mediated this association from first to third grade, whereas child perceived academic competence mediated from second to fifth grade. Parallel cross-lagged models replicated these findings with child academic outcomes assessed using a test of reading achievement and teacher ratings of academic performance.