A Comparative Analysis of Attrition in Household Panel Studies

This paper examines aspects of the use of cross-national comparative research methodologies to review characteristics of respondents who cease to participate after the first wave of a longitudinal household panel study. In general terms, this kind of non-response in survey research is known as panel attrition. The first stages of the paper look at non-response generally as a source of bias in survey interviews and then focus on panel studies. This is followed by a discussion of cross-national research and the opportunities it presents for the study of panel attrition. The Panel Comparative data base (PACO) was identified as a suitable resource and in terms of methodological findings, proved to be an excellent tool for our purposes. From the PACO data base, similar panel studies from France, Germany, Luxembourg, and the UK were analysed for the characteristics of respondents not re-interviewing in the second or third wave of the study, using the Pearson's Chi Square test for significance. In general terms, our findings were in line with other studies suggesting distinct characteristics for respondents who are subject to panel attrition in the second and third wave. There is also a review of the demographic changes through which respondents pass during the first three waves. The findings imply that for all the four countries, such demographic changes are consistent which further strengthens the assumption that respondents have similar patterns in terms of response across nations.