Support #2141
Time-varying weights within the same ID
10%
Description
Hello,
We are analyzing subjects participating in all waves of BHPS and UKHLS from all the boosts, using the time-varying longitudinal weights.
We are also in the presence of censoring. Neither R or Stata allow doing both a fixed effects, random effects or survival analysis using
weights changing within the same subject.
I have a few questions regarding this
1) Should each subject use longitudinal weights up to the last wave available? What about the 0 or missing weights? If we don't want to lose these observations, should we use the last valid weight?
2) Do we need to change the types of weights once a boost is added (e.g. in wave ten or so, change from indin91_lw to indin99_lw?)?
3) Given that we are using UKHLS and BHPS, do we still need to rescale the weights? Should the rescaling consider the weights of all person-years as the base or only the ones selected in 1)?
Thank you and best,
Lydia