Analysis of the tendency to select the “neither nor” option in agreement/disagreement scales on a low – salience topic: The contribution of individual differences

Abstract

This paper examines the frequency of midpoint responses in agree/disagree scales in a survey measuring attitudes toward science and scientific policy, a low-salience and difficult topic. It also examines the contribution of individual differences to explain the tendency to select this option. It is assumed that the use of the midpoint “Neither agree nor disagree” (NA/ND) in Agree/Disagree (A/D) scales is to some extent an indication of satisficing. It is also assumed that there are individual differences in respondents’ tendency to select the NA/ND response. Using a Generalized Linear Mixed Model we include Krosnick’s regulators of satisficing, socio-demographics and individual differences as predictors. We find that the contribution of the regulators of satisficing identified by Krosnick is small. In turn, factors associated with individual differences explain a great amount of variance on the number of NA/ND responses. We conclude that the presence of this option in a survey of a low-salience topic increases satisficing as strategy of the respondents to deal with the cognitive burden of both the A/D scale and the difficulty of the topic. Furthermore, the number of NA/ND responses may be understood as an indicator of individual differences in the propensity of respondents to satisficing when answering a survey on a low-salience or difficult topic.