Abstract
How do survey respondents arrive at answers to survey questions? How do they fit different kinds of subtle or even contradictory ideas into “moderate” or “don’t know” answers? Public opinion research practice typically treats attitudes as discrete thought processes that can be accurately captured with appropriately designed survey instruments. Critics have challenged conventional practice of public opinion research by highlighting the ambiguity of attitudes and their dependence on social context. These issues are most salient in the case of “neither” and mid-scale response selections, which have been characterized by survey researchers as a “dumping ground” for indifferent, unsure, refused, ambivalent, and other response types. Drawing on novel data from 20 cognitive interviews conducted for the 2006 ANES Pilot Study, we examine the distinct meanings respondents assign to “neither” and middle response categories, along with the various strategies they use to match their own reasoning with these categories. We propose a simple analytical framework that distinguishes between the meanings respondents intend to communicate and the logics they use to select “neither” and middle responses. Our analysis of this data from a common validation method empirically demonstrates how the ambiguity of “neither” and middle responses emerges from the complex cognitive work respondents undertake in relation to the flexible meanings of these categories and social expectations generated within the survey interaction. In addition, we show how qualitative cognitive interview data provides valuable insight into the response process that can be extended beyond question validation and used to triangulate interpretation of survey results.