Respondents for Nearly Three Decades: How Do Loyal Sample Members Differ From Others?

Abstract

This paper aims to understand panel attrition by exploring response behaviour in longitudinal social surveys using a latent class framework and incorporating measures to account for unknown eligibility. From this, the characteristics of loyal sample members are identified and how they differ from others in the panel are highlighted. Understanding Society is a household panel survey that began in 2009, following its predecessor, the British Household Panel Study (BHPS). The Understanding Society harmonised BHPS project facilitates this research as it combines both studies such that there are 26 waves of data available. Theexisting literature on panel attrition is extensive but focuses on attritors that leave the panel completely, despite most panel studies allowing sample members to intermittently respond. Latent class analysis allows the study of atypical patterns of response by classifying respondents into groups based on similar response patterns. The key characteristics for loyal respondents are being older respondents (particularly pensioners), educated, and those from smaller household sizes, and lower reported household moves which is consistent with current attrition research.