February 12, 2021
I have a prof using the 2017 SHS. They have a question about the bootstrap weights :
"I have a question you might be able to answer about these bootstrap weights. In order to use them in Stata, I need to know whether they were produced using the 'mean bootstrap' method and if so, I would need to specify how many samples were used to produce each weight in order to adjust the variance estimates to account for mean bootstrap weights (see for example https://www150.statcan.gc.ca/n1/pub/12-002-x/2014001/article/11901-eng.htm under “STATA 12”; the example refers to the GSS, but I’m assuming the procedure applies as well to the SHS PUMF)."
Maybe there is some literature we missed?
The bootstrap weights created for the SHS2017 PUMF should be treated as regular bootstrap weights, and therefore, you are right that the mean bootstrap method is not used, and that the number of samples should be set to 1 (bsn=1).