"How can I estimate correlations and their level of significance with survey data?" Is the recommendation from this post still valid?
Hello, I am running Stata/SE 14.2 on Windows 10. I copied the post title from an old (I think) post -- I'd like to know whether the recommendation from that post is still current.
I need to estimate pairwise correlations and their p-values for a subset of observations from a complex survey. For other analyses based on these data, I use the svy command with the subpop option.
I have not found a command to estimate correlations that works with survey. I have considered using corr or pwcorr, but I know I need to use the pweights option to indicate the sampling weights and corr and pwcorr do not take pweights as an option.
Searching statalist I found posts discussing the features and limitations of the pwcorr option, and I found Stata documentation on the functionality of the survey command, but I didn't see an answer for what I need; that may be because some of the discussions of statistics and how Stata works are rather over my head.
I did find a post using a web search, but it may be old. I don't have the posting date, but it was written for Stat 11. Here is the link:
https://www.stata.com/support/faqs/statistics/estimate-correlations-with-survey-data/
Can someone please tell me whether the recommendation from the post is still valid?
Thank you,
Ximena Leroux
I need to estimate pairwise correlations and their p-values for a subset of observations from a complex survey. For other analyses based on these data, I use the svy command with the subpop option.
I have not found a command to estimate correlations that works with survey. I have considered using corr or pwcorr, but I know I need to use the pweights option to indicate the sampling weights and corr and pwcorr do not take pweights as an option.
Searching statalist I found posts discussing the features and limitations of the pwcorr option, and I found Stata documentation on the functionality of the survey command, but I didn't see an answer for what I need; that may be because some of the discussions of statistics and how Stata works are rather over my head.
I did find a post using a web search, but it may be old. I don't have the posting date, but it was written for Stat 11. Here is the link:
https://www.stata.com/support/faqs/statistics/estimate-correlations-with-survey-data/
Can someone please tell me whether the recommendation from the post is still valid?
Thank you,
Ximena Leroux
Comment