Cognitive style and the survey response

According to democratic theory, citizens should have a set of well-defined, valenced opinions. Yet evidence suggests that individuals vary in chronic evaluative tendencies, with some processing information in an online fashion, spontaneously becoming opinionated about whatever they encounter and recalling summary evaluations easily. Individuals lower in evaluative tendencies do not form attitudes in this fashion and instead must construct and report attitudes based upon recalled considerations. The implications of these individual differences for survey practice are largely unexamined. The present research shows, across multiple surveys and question formats, that individuals low in need to evaluate (NE) offer greater numbers of non-informative responses to opinion questions than those higher in NE, an effect not attributable to interest, knowledge, or lack of cognitive effort. And when they do supply answers to opinion questions, low-NE individuals report attitudes less extreme than individuals higher in NE. The results suggest that opinion surveys may systematically misrepresent the variability of citizens’ beliefs and the extremity of aggregate public opinion by relegating some of those low in NE to non-informative response options.