Can we assume that responses to poll questions actually reflect respondents' opinions, attitudes, thoughts, or feelings? Several variables can confound actual opinions and our interpretations of them. A great deal of ‘noise,’ including communication breakdowns from dicult‐to‐follow questions, may hamper the typical opinion poll. Researchers have explored poll questions regarding question order, response order, context eects of preceding questions, even gender of the questioner or respondent. Only a few studies, however, have looked at the actual complexity or diculty of the poll questions themselves.

You do not currently have access to this article.