Reconstruction of Relationship Memories: A Mental Models Approach

Abstract

As survey researchers, we must often rely on the introspection of strangers. If we wish to know a person’s opinion on some complex political issue, we have little recourse but to ask him or her. We can take pains to present the questions in as simple, comprehensible, and unbiased a manner as possible; in the end, however, the difficult task of remembering, organizing, and synthesizing disparate information into a single coherent attitude, and then mapping that attitude onto the available response alternatives, must fall to the respondent alone. Research by Converse (1970) and others suggests that this task may often be too much for respondents: They may end up responding with transitory and inconsistent “nonattitudes” that provide little information about their true beliefs.