A population survey on legislative measures to restrict smoking in Ontario: 4. Variables related to knowledge of active and passive smoking health effects and to predicted behavior of smokers and nonsmokers.

Results from the further analysis of a population survey on legislative measures to restrict smoking revealed that identification of subgroups of smokers is more reliable than identification of subgroups of nonsmokers when a variety of attitudes were the measures of interest. A similar pattern emerged when analyses were carried out on knowledge of active and passive smoking health effects and on predicted personal and general compliance. Because distinct sets of variables were found to be related to distinct outcomes, program planning for changes in knowledge and behavior might, of necessity, have to be different. Media messages might be useful for changes in knowledge, while actual experience might be more important for attitude and behavior change.