Correlations from complex survey designs
in [SAS]
|
Prev: Siebel Business Analyst -- MI
Next: A message to anyone posting to SAS-L via usenet, Google or anything other than the listserv
From: BruceBrad on 23 Apr 2010 00:05 Any suggestions on the best method to estimate correlation coefficients and their standard error from stratified and clustered survey data? That is, a replacement for the non-existent proc surveycorr. Is it possible to take the replicate weights generated by surveyreg and use them to get multiple estimates and hence estimate of parameter variances estimated using non-survey procedures? Any example documentation on this? |