The implementation of cross-sectional weights in household panel surveys

Abstract

While household panel surveys are longitudinal in nature crosssectional sampling weights are also of interest. The computation of crosssectional weights is challenging because household compositions change over time. Sampling probabilities of household entrants after wave 1 are generally not known and assigning them zero weight is not satisfying. Two common approaches to cross-sectional weighting address this issue: (1) “shared weights” and (2) modeling or estimating unobserved sampling probabilities based on person-level characteristics. We survey how several well-known national household panels address cross-sectional weights for different groups of respondents (including immigrants and births) and in different situations (including household mergers and splits). When a new person moves into a household, both “shared weights” and “modeling” lead to reduced individual weights of pre-existing household members, but differences due to the approach arise elsewhere. The implementation of “shared weights” is problematic when the panel contains households without a household member already present in wave 1. Panels also differ in the treatment of immigrants, household merges, and sometimes on how weights are assigned to children born to wave 1 panel members.