Open versus closed

The expression of attitudes is very sensitive to variations in survey structure. With a survey with both open and closed questions of a single sample of respondents, an analysis was done on attitudes about public transportation of peak-hour transit users in Arequipa, Peru. It was found that combining both question types helped identify and compensate for the limitations of each. The results of the open and closed questions were dramatically different, even though the same individuals answered both questions. In response to nondirective, open-ended questioning, respondents emphasized experiential concerns such as comfort and travel times. However, when asked to rank transit-related issues in order of importance, respondents overwhelmingly chose social concerns such as pollution and traffic accidents over the experiential concerns they themselves previously mentioned. The psychological forces that may have led to these apparently contradictory results were explored. The paper drew on the belief-sampling model of survey response and social desirability bias to explain why open questioning may have understated the true value respondents placed on social issues, whereas the closed question may have underemphasized the importance of users' quotidian travel experience. The participants' willingness to change answers between the two question forms suggests that confirmation bias may not be a significant impediment to the use of mixed question types.