Thanks for all of your help and suggestions. Since I already had all of my "proc sgplot" code written out, and since sgplot apparently does NOT have a way to generate the slope, I just did what I needed in a brute force manner as shown here: For each plot of interest, I used proc sgplot to get my plot with confidence intervals: title "Scatter Plot with Confidence Interval - PMC Non-Hispanics, Dx = Normal"; proc sgplot data=allpmc3; where short_dx = "Normal"; reg y=mmsetotal x=frstotal / cli ; yaxis values= (0 to 30 by 5); xaxis values = (0 to 55 by 5); run; quit; Then I did a Proc Rec using 'noprint', but with the OUTEST keyword to create a dataset containing the needed data: proc reg data=allpmc3 noprint outest=PMC_norm; where short_dx = "Normal"; model mmsetotal = frstotal; run; quit; Then I modified the dataset created by OUTEST so that it provided the info in the format I needed: data PMC_norm2; set PMC_norm; rename frstotal = slope; group = "PMC Normal"; keep group intercept frstotal; run; Finally, once I had repeated the above steps for multiple plots and groupings, I just put all my datasets together so that I could print a nice neat table of intercepts and slopes: data eq_params; length group $12; set metab_ad2 metab_mci2 metab_norm2 pmc_ad2 pmc_mci2 pmc_norm2; run; proc print data=eq_params; run;
... View more
