An experimental trial was used to assess the effectiveness of a self-help book and a personalized assessment-feedback intervention, both separately and in combination with each other, in a general population survey. Participants (N = 86; 66.3% male) were recruited through a random digit dialing telephone survey conducted by the Centre for Addiction and Mental Health, Toronto, Ontario, Canada. Respondents were randomly assigned to one of four conditions in a two-by-two factorial design: "no-intervention" control group, "personalized feedback only," "self-help book only" and "both personalized feedback and self-help book." Respondents were followed up in 6 months' time, and differences in drinking status were compared between experimental conditions using a multivariate analysis of covariance with baseline drinking severity as the covariate. Support was provided for an interaction hypothesis in which respondents who received both interventions reported significantly improved drinking outcomes at 6-month follow-up, compared with respondents who received just one of the interventions or who received no intervention. Because respondents were recruited from a representative sample of the general population into a randomized trial with a no-intervention control group, this research design maximized both external and internal validity in examining the effectiveness of self-help interventions.