Hello. I'm sorry if this question has an obvious answer. It was my understanding that revised cross-sectional weights for Wave 3 would be included in the Wave 4 release. (For example, I think this issue was mentioned in Support #315.) I was just wondering if you could tell me which file these revised weights are included in?
#2 Updated by Noah Carl almost 6 years ago
Many thanks for the quick response. If I am not mistaken, there are a still a considerable number of zeros on the weight variable c_indinub_xw. For example, several thousand individuals originating in the 2009-10 ukhls sample seem to have a zero weight on this variable. Similarly, in the Wave 4 data file, several thousand individuals originating in the 2009-10 ukhls sample seem to have a zero weight on d_indinub_xw. I was just wondering if you could tell me what is the explanation?
#3 Updated by Peter Lynn almost 6 years ago
As described in the user guide (section 3.7.3), cross-sectional weights are non-zero for persons in households where at least one person has a non-zero analogous longitudinal weight. We do not believe that appropriate methods exist for giving non-zero weights to others. In the case of c_indinub_xw, it is non-zero for 86.4% of wave 3 respondents. This proportion is higher for the GPS (88.2%) than for the EMBS (74.3%), with BHPS in-between (86.6%), reflecting the differences in wave response rates between the samples. Amongst OSMs, the only sample members eligible to have responded at each wave, the proportion with non-zero weights is higher (88.7% for GPS; 79.6% for EMBS).