I'm fairly new to weights because the databases I used previously were not samples but actual populations I wanted to study, so apologies if all of this sounds very noobish.

The way I understand it, weights give certain elements in a database more "weight" in order to make the whole database more representative. If your survey has, for example, 10 countries, but you have 1,000 observations for each country, it won't be representative because all those countries don't have the same population. So the more populated ones should "weight" more. Is this correct?

I'm using the European Working Conditions Survey, a survey that includes observations from all EU countries (and some more). My analysis, however, wants to focus on only three countries of all of them, but use two waves, the 2010 and the 2015 in order to conduct a cross-sectional analysis at two points. As I understand, there are different weights included in the survey, one for when you use all the countries, when you use just one...

1) Do you know what weight should I use if I want to analyze three countries from two different waves (2010 and 2015 from the EWCS) and how to do so? Should I stratify the analysis per country? Is the weight applied in the same way if my database has both waves combined? The weight is just applied as" svyset[pw=x]" (with X being the weight I would use)?

2) In this question (https://www.researchgate.net/post/How_to_apply_survey_weights_in_Stata_for_the_European_Working_Conditions_Survey) there is snippet explaining the different types of weights used in the EWCS. Where is that text bit coming from, the EWCS weighting report? I can't seem to find it.

Thanks a lot!


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.