Support #506

weights and design variables query Waves 2-5

Added by Orla McBride almost 5 years ago. Updated over 4 years ago.

Start date:
% Done:




I was wondering if you could answer a question I have about weights and psu/strata variables for analysing Understanding Society data.

If I want to analyse data from the BPHS, GPS, and EMB samples from Wave 2-5, which comes from both the self-completion questionnaire and the main survey, should I use:

e_indscus_lw: Longitudinal adult self-completion questionnaire weight
e_strata: Sampling strata
e_psu: Primary sampling unit

Many thanks for your time.

Kind regards


Updated by Victoria Nolan almost 5 years ago

  • Assignee changed from Orla McBride to Olena Kaminska

Updated by Olena Kaminska almost 5 years ago

Yes, the variables are correct. Note, the BHPS sample will be excluded (as it is not in wave 1 of UKHLS). But you still will represent the UK population - so you should be fine.
Best, Olena


Updated by Orla McBride almost 5 years ago

Great thanks. I'm not using data from W1 for that reason. I thought the BHPS would be included in the longitudinal weight from W5 because they were included from W2 onwards?

Kind regards


Updated by Olena Kaminska almost 5 years ago

There is almost unlimited number of sets of weights we could provide. Yet providing w2+ longitudinal sc weights can be confusing for users. You can use w1 in your analysis as e_indscus_lw includes it. By dropping BHPS you loose a very small sample size, so your analysis results should not be affected.
Again your aim should not be to use a particular sample. Your aim should be to represent a population - and this can be perfectly achieved with e_indscus_lw weights.
Hope this helps, Olena


Updated by Orla McBride almost 5 years ago

Very helpful - thanks a lot!


Updated by Victoria Nolan over 4 years ago

  • Status changed from New to Closed
  • % Done changed from 0 to 100

Also available in: Atom PDF