The Effect of Response Categories on Questionnaire Answers

Abstract

This article reports the effect that the ranges presented in answer categories for survey questions can have on respondent answers. Response categories were manipulated in a split-ballot survey conducted in both telephone and mail modes. These categories, presented in the separate ballots, overlapped in one category; the other categories were unique to each ballot. The experiment was conducted on four questions: two frequent and mundane and two rare and salient. It was found that the response categories significantly affected the response for frequent and mundane questions. One question demonstrated a significant difference in response between the mail and telephone modes. For this question, a response scale with a limited number of socially desirable alternatives resulted in a social desirability effect in the telephone mode. Alternatively, the telephone mode demonstrated an extremeness effect when the response scale comprised a greater number of socially desirable alternatives.