Predicting future smoking of adolescent experimental smokers

Abstract
This prospective study examined the usefulness of seven selected variables from Jessor and Jessor's problem behavior theory in predicting the future smoking levels of adolescent experimental smokers. Experimental smokers were identified through a survey of 2,550 seventh- and eighth-grade students assessing subjects' current smoking levels as well as subjects' responses to the selected variables. Twelve months later the entire sample was resurveyed to assess subsequent smoking levels of subjects initially reporting experimental smoking. Experimental smokers who responded to both surveys (n=225) were subjects of the present study. The full model discriminant function was significant in discriminating between the two groups; 68% of the subjects were correctly classified. This moderately successful hit rate was 24% better than what could be expected by chance. Increased smokers had lower expectations for academic achievement, had stronger beliefs about the positive functions of smoking, and perceived more smoking among their classmates and close friends. Educational interventions are suggested.