Survey Response Bias: Effects of Introductory Approaches and Feedback on Honesty of Response.

摘要

Response bias, is the reason a respondent may answer a survey question in a way which does not reflect true thoughts or feelings. As a consequence, survey data may reflect various degrees of response bias which render survey findings suspect or even useless. This thesis tests the hypotheses that approaching respondents humanistically and promising feedback in the form of results of the survey would significantly reduce response bias. Four sets of instructions to a standard military survey were developed. The sets consisted of 2 sets of authoritarian instructions (with and without the promise of feedback) and 2 sets of humanistic instructions (with and without the promise of feedback). The research design was a 2x2 factorial design using two-way analysis of variance to test the research hypothesis. In a postsurvey interview, responses indicated the degree of honesty (indicator of response bias) of response to the survey. Results of the statistical analysis were unclear. There was no significance for the hypothesis that a humanistic approach would decrease response bias and only one of the three interview questions showed significance for the reduction of response bias through offering feedback. (Author)