This study undertakes an empirical test of the "bandwagon effect"—individuals rallying to the majority opinion. The study is done outside the electoral context on two issues: abortion and the constitutional future of Quebec. A panel is used, as well as an experimental design in which respondents are told the state and direction of public opinion. Three methodological criteria are used as minimal requirements for a satisfac tory test of the bandwagon thesis. "Underdogging," as well as opinion movement due to factors outside the experiment, are both accounted for. Though the reasons for a bandwagon remain unclear, the authors demon strate that a bandwagon effect of 5-7 percent existed on both issues.