This prospective study used a conceptually based risk and protective framework to investigate whether parental influences exert a protective effect on the robust association between peer influences and college alcohol involvement. Participants were incoming freshmen in the control condition of a randomized clinical trial, N = 256, 57.0% female, baseline age: M = 18.36 years (SD = 0.41). Participants completed telephone surveys in the summer before matriculation (baseline) and in the spring of the freshman (10-month) and sophomore years (22-month) with 85.6% retention at 22-months. Latent growth models were estimated for heavy episodic drinking and alcohol-related consequences. Descriptive norms and social modeling among peers were positively associated with initial heavy drinking and consequences; parental drinking permissiveness was positively associated with initial heavy drinking. Greater social modeling among peers was associated with less growth in consequences. Parental monitoring was not significantly associated with alcohol involvement. Prematriculation social modeling exhibited a weaker positive association with initial heavy drinking and consequences at low prematriculation parental drinking permissiveness compared to high. Similarly, prematriculation descriptive norms exhibited a weaker positive association with initial heavy drinking at low prematriculation parental drinking permissiveness compared to high. Prematriculation descriptive norms were not significantly associated with growth in heavy drinking at low parental drinking permissiveness; in contrast, higher prematriculation descriptive norms were associated with less growth in heavy drinking at high parental drinking permissiveness. Findings provide support for a protective parental influence on peer-alcohol relations extending into college. Parental drinking permissiveness may be an important target for parent-based interventions.