This study evaluated cigarette smoking prevention methods with a sample of adolescent females and males. Arranged by school, 331 informed and consenting sixth graders were randomly divided into four groups: 1) pretest, skills-building methods, posttest; 2) pretest, discussion methods, posttest; 3) pretest and posttest; and 4) posttest only. All subjects were followed for 6 and 15 months after the posttest. Outcome results on measures of non-smoking intentions, attitudes, predictions, problem-solving abilities, and peer interactions favored subjects in the skills-building group when compared with subjects in the discussion and pretest-posttest control groups. Smoking rates at posttest and at both follow-ups were lower in the skills-building group than in the other three groups. Results from posttest-only subjects did not support pretest reactivity. The study's strengths and limits are discussed along with directions for future smoking prevention research.