Household weighting for longitudinal analysis
Dear Understanding Society Team,
I'm running an analysis at the household level with averaged individual-level attributes by household (dependent variable is a household attribute). I would like to clarify the following:
1) Can household weights be used irrespective of the individual-level attributes?
2) For a longitudinal analysis at the household level, has for each wave a different cross-sectional household weight to be used (n_hhdenub_xw) since there is not longitudinal weight for household-level analyses?
Thank you for your help.
#1 Updated by Stephanie Auty about 1 year ago
- Category set to Weights
- Assignee set to Olena Kaminska
- Target version set to X M
- Private changed from Yes to No
Many thanks for your enquiry. The Understanding Society team is looking into it and we will get back to you as soon as we can.
Stephanie Auty - Understanding Society User Support Officer
#3 Updated by Olena Kaminska about 1 year ago
1) You can use household weight if you have information for everyone in the household (if the information comes from hhresp or indall) and this information is averaged. Essentially you will have one value per household then - so household weight is appropriate. Nevertheless if the information comes from not all individuals in the household (e.g. from indresp file) then you must use individual level weights.
2) We believe that households are not longitudinal entities as they change over time.
Hope this helps,
#5 Updated by Ante B about 1 year ago
Thank you for the clarification.
I have one more question related to weights. In the User Manual it writes that weights "... adjust for unequal selection probabilities(1), differential nonresponse(2), and potential sampling error(3). A weighted analysis will adjust for the higher sampling fraction in Northern Ireland and for different probabilities of selection in the EMB sample, as well as for response rate differences between subgroups of the sample."
I understand that
"higher sampling fraction in Northern Ireland" refers to (1),
"different probabilities of selection in the EMB sample" refers also to (1),
"response rate differences between subgroups of the sample" refers to (2).
Is this understanding correct? If so, what does sampling error (3) exactly refer to since I understand that not accounting for (1) and (2) can lead to (3)?
#6 Updated by Olena Kaminska about 1 year ago
Yes, your understanding is correct. And the sampling error refers to the point that there is a small chance that the selected sample may not be representative of the population - so the weights correct for this through post stratification.
Hope this helps,