The study demonstrates that higher perceived burden of panel survey participants, especially in earlier waves, is a much stronger predictor of attrition in subsequent waves than item nonresponse, providing support for existing theories of respondent burden in the longitudinal survey context.
