In interested in doing a matched sets regression analysis. Conceptually, what I'd like to do is similar to using matched sets in a stratified conditional logistic regression. Except instead of a binary response variable I have a continuous variable. Currenly my matching is one-to-one. That is for each treatment I have one matched control. Each such observation tuple is identified via a common identifier value.
There is no meaningful way to dichotomize my dependent variable.
In PROC LOGISTIC this is accomplished using the STRATA statement.
I noticed PROC SURVEYREG has a STRATA statement and conducts a GLS regression. Are the STRATA statements between these two procedure comparable in any sense?