Conducting ttest for pretest and posttest survey using svyset

[Sonam Rinchen]

Hello experts,
I would like to request for your expert help in conducting ttest for pretest and posttest survey using svyset. I have two datasets fro two different surveys - pretest and posttest survey. Prestest with 13000 samples and posttest with 2000 samples; these 2000 samples match with pretest sample. For these, I also have two different weights for presets and posttest. I have merged this two datasets together. I want to do ttest for these two datasets (unpaired) using svyset. For example, considering household size, I have two variables - hhmem (pretest) and b1 (posttest). How to perform ttest or compare mean using svyset? Your support is highly appreciated. Thanking you.

[Rich Goldstein]

the t-test command itself cannot be directly used; however, since a t-test is the same thing as a regression (in your case the only predictor will be the indicator variable for pre v post), you can use "svy: regress"; note that this question has been discussed a number of time on statalist and a search via good search engine will find these discussions

[Sonam Rinchen]

Thanks a lot. I will explore the discussion. Much appreciated.