Introduction: We observed an increase in prevalence of former smokers in Long Beach between 1998 and 2000, the third year of the law that banned smoking in California bars. To contribute to a generalizable understanding of those quitters, we fitted a model to distinguish between quitters and smokers and applied the model to predict the proportion of quitters in 2000.

Methods: Data were obtained from two telephone surveys conducted at the ends of 1998 (N=1502) and 2000 (N=1506). With variables measuring Demographics, Tobacco-Control Attitudes, and Exposure to Location-Specific Smoking Bans, we used step-wise logistic regression and fitted a model to predict quitters. Using the 1998 independent-variable means, we calibrated the model to match precisely the observed 1998 quitter proportions; we then applied the model to predict the year 2000 quitter proportions from that year's independent-variable means.

Results: Among the 1998 ever smokers, 56.3% were quitters. For 2000, the model predicted 62.9% quitters, just slightly more than the observed proportion of 60.8%. Three variables that contributed the highest percentage to the predicted increase in quitters were the Need for Protection from Secondhand-Smoke Scale (34.1%), Personal Approval for the 1998 law that extended a workplace-smoking ban to workers in bars (22.6%), and the presence of a Home Smoking Ban (26.3%).

Conclusions: Successful variables combined predictive strength with sufficient change between 1998 and 2000. These variables primarily comprised direct measures of attitudes toward tobacco control in the form of bans on smoking, along with exposure to a ban on smoking in the home.