The aim of this study was to examine the effect of a theory-based in-home family intervention (In control: No alcohol!) on adolescent alcohol cognitions via its putative mediators using a randomized controlled design. In the South Holland region of the Netherlands, a total of 213 children (11-12 years) and their mothers were randomly assigned to the prevention program (108 dyads) and the control condition (105 dyads). Mediation effects were analyzed using pretest and two follow-up measurements (5 and 12 months after baseline). A path model was estimated (using Mplus) to examine the effect of the intervention on the putative mediators (frequency- and quality of mother-child communication, rules about alcohol, establishing a nondrinking agreement, and parental monitoring of the child's whereabouts). Outcomes were adolescents' perceived harmfulness of drinking and intention to drink. Multigroup analyses were performed to examine potential differences across gender. The program led to an increase in frequency of alcohol-specific communication, nondrinking agreements, and parental monitoring. Moreover, adolescents in the experimental condition perceived drinking to be more harmful and had less intention to drink compared to adolescents in the control condition. The effect of the program on adolescent alcohol cognitions was significantly mediated through having more frequent conversations about alcohol, yet only among boys. Although results on actual drinking need to be added, findings indicate that this relatively inexpensive, easy-to-administer home intervention is promising.