Demographic and psychological predictors of panel attrition: evidence from the New Zealand attitudes and values study.

This study examines attrition rates over the first four years of the New Zealand Attitudes and Values Study, a longitudinal national panel sample of New Zealand adults. We report the base rate and covariates for the following four distinct classes of respondents: explicit withdrawals, lost respondents, intermittent respondents and constant respondents. A multinomial logistic regression examined an extensive range of demographic and socio-psychological covariates (among them the Big-Six personality traits) associated with membership in these classes (N = 5,814). Results indicated that men, Māori and Asian peoples were less likely to be constant respondents. Conscientiousness and Honesty-Humility were also positively associated with membership in the constant respondent class. Notably, the effect sizes for the socio-psychological covariates of panel attrition tended to match or exceed those of standard demographic covariates. This investigation broadens the focus of research on panel attrition beyond demographics by including a comprehensive set of socio-psychological covariates. Our findings show that core psychological covariates convey important information about panel attrition, and are practically important to the management of longitudinal panel samples like the New Zealand Attitudes and Values Study.