Understanding the patterns of modes switch in longitudinal studies
|Day:||Tue 2 Jul|
The research proposed here aims to better understand one of the key characteristics of a mixed mode design: how people transition from one mode of interview to another in a longitudinal survey. This is essential as it can be used both for targeting respondents and for dealing with non-response bias. Additionally, it can potentially explain differences in mode effects on estimates of change. In this paper we investigate the process of changing modes by looking at waves 5-10 of the UKHLS Innovation Panel. Based on the mode of participation in the mixed mode group and the likelihood of participation in waves 5 to 9 of the Innovation Panel we have identified five distinct classes of respondents. Furthermore, these classes were used to predict the likelihood of participation and the mode in which they do so in wave 10. Our results show that the latent classes significantly contribute to our prediction of future wave participation. We discuss possible strategies to use this information for targeting and non-response.