Matthias Schonlau, Rand Corporation
Martin Kroh, DIW Berlin
Abstract
The first wave of a household panel survey captures a cross section of the population at one point in time. However, target population and household composition change over time. Following rules are introduced to decide which household members are still surveyed after they leave the household. We compare following rules of several large household panels and show how different following rules would have affected sample size in the SOEP. Changing household compositions also raise questions about appropriate sampling weights. We compare various weighting approaches employed by large household panels to address this issue. We find that after 25 panel years most respondents still live in a household with a wave 1 respondent. Therefore, following rules have relatively little impact on sample size.