Understanding Society User Support

Genevieve,

Thanks for your question. There is no one way of conducting longitudinal analysis of youth data. The main reason is that one can be in youth data for a maximum 6 years (ages 10-15), but this group is usually very small and may not be sufficient for analysis, especially in BHPS. There are many alternatives, like using 2,3,4 or 5 consecutive years. Because of the variety of options, we do not provide special weights for the longitudinal analysis of youth.
We suggest that for any longitudinal analysis of youth (either alone or with a combination of adult questionnaire) you use enumeration weights, e.g. w_psnen91_lw. These will be suboptimal, but will take into account everything expect for nonresponse to youth questionnaire conditional on household response.

Hope this helps,
Olena