"The Words of Things Entangle and Confuse"

Abstract

The validity of social science research based on interviews and questionnaires is threatened by the inclusion of ambiguous concepts which cannot be assumed to have the same connotations for all respondents. Recent developments in cognitive psychology offer the first satisfactory procedures for reducing these ambiguities. Using "the government in Washington" as the illustrative concept, we here show in two experiments how the ambiguities can be identified and reduced. The importance of using these techniques in preference to common survey approaches to ambiguity, such as the use of open-ended questions, is discussed.