Reducing Panel Attrition: A Search for Effective Policy Instruments

In this paper we develop a theory of the survey response decision process and apply it to the analysis of field office policy measures in an attempt to see which of these are effective in reducing panel attrition. We use data from the Health and Retirement Study (HRS) to assess the effectiveness of 1) reducing the length of the interview and 2) assigning the same initial interviewer wave after wave. There is virtually no evidence in the data that interview length affects subsequent wave response. Assigning the same interviewer wave after wave, however, has a strong positive effect on response rates.