At first I wondered if it is because your data are degenerate, in the sense that yearprimary is throwing df=0, meaning all obs have the same value. However, that is not the problem--you need to include a CI= option in the PROC QUANTREG statement. To get standard errors, t values and p values, you have to put CI=SPARSITY or CI=RESAMPLING on as options. Steve Denham
... View more