Possible conclusions regarding the effectiveness of alcohol education programs are that: (1) implementation of unevaluated programs is better than doing nothing; (2) unevaluated alcohol education runs the risk of increasing alcohol use; (3) current alcohol education programs benefit some recipients, some of the time, with respect to some outcomes; and (4) the impact of alcohol education programs is unknown. The inconclusiveness of empirical evidence concerning program effectiveness leads to the adoption of this last, agnostic position, according to which the state of empirical ignorance is compensated by giving greater attention to the deductive logic implied in the assumptions underlying alcohol education. Common flaws in the deductive logic of alcohol education are identified; a series of increasingly sound underlying syllogisms is proposed, leading to the establishment of (inductively and deductively) logically sound objectives and targets for alcohol education programs. A review of existing empirical evidence concerning effectiveness of alcohol education is combined with the outcome of the deductive-logical analysis, resulting in a series of recommendations for effective alcohol programming, including targeting programs to specific subpopulations, and incorporating programs with as much evaluation research as is feasible. 64 references. (Author abstract modified)