Abstract: | Public opinion research in general and the study of opinionchange in particular have frequently been challenged as unreliable.Attitudinal measurements are characterized as merely reflectingnon-attitudes, labile moods, and/or methodological artefacts.Given this basis, the measurement of opinion change is seenas a product of random variations and systematic biases ratherthan as reliably measuring true change. However, most opinionchange is (1) not chaotic, but slow and steady and (2) largelyexplicable. Even opinion change that is rapid and/or multi-directionalcan be plausibly explained. Measurement variation does howeveroften distort time series and greatly complicates the reliableassessment of true change. |