Where does social change come from?

Many discussions, academic and otherwise, implicitly assume that public opinion changes because opinions change. This ignores the possibility that public opinion changes because publics change. In this paper I show how, by modifying existing component-difference methods, the proximate sources of societal change - actual individual change versus change in publics (turnover) - can be separated using repeated survey data. The method is applied to change in gender role attitudes in the United States, using 1972–1988 data from the General Social Surveys. Both components have contributed substantially to the trend away from traditional attitudes. Yet there is an important difference: Population turnover has contributed steadily to the trend, whereas the contribution of individual change has been erratic from survey to survey.