Changeable weather in a cooling climate atop the liberal plateau: Conversion and replacement in forty-two general social survey items, 1972-1989

Abstract

This article tracks trends (early 1970s to late 1980s) in U.S. opinion for 42 General Social Survey items with liberal/ conservative overtones. The broad question is whether the great "liberal" shift since World War II has ended; the narrow issue is the relative importance of cohort succession and intracohort shifts. Despite common impressions, the overall trend is more liberal than conservative, but it conceals opposing "weather" and "climate" processes. Within cohorts ("weather") I find a conservative trend between the early and late 1970s and a liberal "rebound" in the 1980s. Between cohorts virtually all items show small but cumulative liberalizing produced by cohort succession. These cohort effects are declining in magnitude because the association between year of birth and liberalism is nonlinear. I find a curvilinearity such that Americans born after World War II are not consistently more liberal than their predecessors. This shift is not explained by the lesser schooling of youngest adults or by ceiling effects. Consequently, I predict lessening of the liberalizing "climate" produced by cohort succession. All these propositions are qualified, depending on the topic, and the analysis takes heed of the notorious age/period/cohort identification problem. © 1992 by the American Association for Public Opinion Research.