The Shirking Model – A Theory of How People Answer Survey Questions

This thesis explores survey responses in an attempt to uncover behavior that might indicate that respondents are not exerting sufficient effort to ensure reliable responses. By developing a rational choice theory about survey response behavior, new tests for examining how respondents behave when ans...

Full description

Bibliographic Details
Main Author: VanBeselaere, Carla Emily
Format: Others
Language:en
Published: 2004
Online Access:https://thesis.library.caltech.edu/2197/1/ThesisDraftFinal20040528.pdf
VanBeselaere, Carla Emily (2004) The Shirking Model – A Theory of How People Answer Survey Questions. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/VZ6S-MM10. https://resolver.caltech.edu/CaltechETD:etd-05282004-161831 <https://resolver.caltech.edu/CaltechETD:etd-05282004-161831>
Description
Summary:This thesis explores survey responses in an attempt to uncover behavior that might indicate that respondents are not exerting sufficient effort to ensure reliable responses. By developing a rational choice theory about survey response behavior, new tests for examining how respondents behave when answering opinion questions are established. The first and most important problem addressed in this work is the identification of respondents who shirk by providing responses that are not fully thought-out. Assuming that shirking behavior is indicated by "no opinion" responses, this work explores the frequency and sources of this behavior. The results from this analysis provide useful recommendations for survey design. However, to further explore the implications of shirking behavior a new measure of shirking is proposed. This measure, based on survey response times, enables an examination of the implications of shirking behavior and proposes possible ways by which shirking behavior might be accounted for in the analysis of survey data. The results from this work suggest that if analysts want to ensure that accurate results drawn from survey data, consideration should be given to the likelihood that respondents shirk in answering survey questions.