Dear users,
My Center uses national health surveys a lot. There is documentation on creating level 1 and level 2 weights for contextual analysis using surveys (https://www.bristol.ac.uk/cmm/software/ ... ights.html). But my problem is that I want to pool multiple surveys from different countries, adding a third level to the model. My question is how to extend the proposed weights to this situation? How should the survey weights (level 3) be calculated?
Thanks a lot!
Aluisio
Weighting in pooled survey analysis - 3 levels
-
- Posts: 2
- Joined: Wed Dec 13, 2023 4:58 am
Re: Weighting in pooled survey analysis - 3 levels
Once the samples are aligned, you can calculate the weights at the third level. Start by calculating the overall weight for each survey by combiningmoto x3m the weights from the first and second levels. Then, adjust the weights proportionally to align with the reference population at the third level.
-
- Posts: 6
- Joined: Wed Dec 20, 2023 8:55 am
Re: Weighting in pooled survey analysis - 3 levels
The third level of weight calculation may be performed once the samples have been aligned. Gather the first two levels' weights and combine them to get the total survey weight moto x3m. The next step is to make proportionate weight adjustments such that the third level reference population is satisfied.