Recent research into the structure of public attitudes suggests that levels of "constraint" in the mass public increased substantially between 1956 and 1964, largely in response to the ideological nature of the 1964 presidential election. This article examines the validity of this interpretation. It suggests that constraint in the mass public probably did not increase very much between 1956 and 1972 but that reported changes were due instead to changes in the survey items used to measure constraint. Some implications of this view are then considered.