Predicting Success in a Smoking Control Program

Abstract
Using 10 independent variables, several of which were previously related to smoking behavior, a regression equation was derived to predict success in a smoking control program. Human subjects [31] were divided into the main and cross-validation treatment groups. Both groups participated in 4 2 h sessions. Taken individually, none of the predictor variables discriminated between successful and unsucessful subjects. The regression equation, based on the 5 best predictors, accurately predicted direction of change of smoking rate 87% of the time.