The Impact of Age and Education on Response Order Effects in Attitude Measurement

Abstract

This article addresses the impact of age and education on the emergence of response order effects. Previous research has suggested that response order effects are stronger for respondents with less formal education. Education, however, is negatively correlated with age. Therefore, the apparent educational differences in response order effects might in fact reflect differences in age. A series of meta-analyses of experiments originally conducted by Howard Schuman and Stanley Presser and subsequent hierarchical logistic regression analyses demonstrate that there is a statistically reliable relationship between response order effects and age and that this relationship is not explained by education. Age-related differences in response order effects may undermine substantive conclusions about age or generational differences in attitudes