Hi guys,
I'm trying to compare the percentages of satisfaction of gp practices in different waves (year). I averaged the number of respondents per practice over the sample period to use as my weight, then I got say 130 responses in the year 2012 for all the gp practices, and 140 responses in the year 2013 and so on. However, when I used xtdidregress (q18_1) (gmpost), group(practice_code) time(wave),[aw=average_base], I got an error 'weight must be constant within practice_code'. The thing is, different practice_code in different waves could not be the same. I wonder how can I weigh it properly, or how can I compare the percentages using Stata? Thanks a lot!
I'm trying to compare the percentages of satisfaction of gp practices in different waves (year). I averaged the number of respondents per practice over the sample period to use as my weight, then I got say 130 responses in the year 2012 for all the gp practices, and 140 responses in the year 2013 and so on. However, when I used xtdidregress (q18_1) (gmpost), group(practice_code) time(wave),[aw=average_base], I got an error 'weight must be constant within practice_code'. The thing is, different practice_code in different waves could not be the same. I wonder how can I weigh it properly, or how can I compare the percentages using Stata? Thanks a lot!
Comment