Abstract
A review of outcome evaluations published after 1970 dealing with school based interventions on alcohol and tobacco reveals four major theoretical bases supporting the program intervention models employed: rational, social reinforcement, social norms and developmental. The review found this research still in need of methodological improvements prominently involving the assessment of experimental attrition and the validity of the dependent variable measure. For the projects reviewed here, program interventions based upon predominantly rational models produced fewer positive behavior and attitude effects but more positive knowledge effects than those for which the rational model was not dominant. Behavior effects for published school based smoking outcome evaluations were generally positive for both short and long term effects. Behavior effects for published alcohol evaluations were mixed, positive and negative. Knowledge effects were consistently positive over all studies reviewed here, but attitude effects were mixed, some positive and some negative. The sign, positive or negative, of the attitude effect was always directly associated with the sign of the associated behavior effect over all projects reviewed here. No such relation existed between knowledge and behavior effects. An interpretation of this finding is developed using the value-expectancy theory.