Modeling effects in student drinking and smoking, revisited after 24 years

A systematic replication is reported of a campus survey carried out in 1978. The present data (N = 100) indicate increased peer modeling effects for drinking, in which a respondent's intake of alcohol and drunkenness are both best predicted from the drinking of his/her friends. Smoking, however, today shows no correlation with the smoking of either friends or parents, unlike the previous survey. Alcohol consumption, measured as drinks per week, has remained constant over 24 years for female students, but has doubled for males, reaching four times the female level. The frequency of drinking "to excess" also increased greatly over this interval. Smoking has decreased to minimal levels, and now shows no sign of modeling effects, but self-serving bias now occurs, since respondents today report themselves as smoking fewer cigarettes than their friends.