How do I get the point estimates for the reference value in the following? proc logistic data =WORK.shihab ; class PainSites (ref ='2') model Success_trials (event='1') = PainSites /clodds=wald; run; quit;   I'm getting the following: Pain sites 1 vs. 2 Pain sites 3 vs. 2   But I also want he point estimates for Painsites 2 as well. Any help much appreciated. ... View more

  I would like to change order of the box plots to be: control OA and LSS. I am not sure why it is flipped the opposite way.   Also how do I override the automatic title. looks like the code I put in for the title does not work. Appreciate all the help!   See attached graph. here is the code I used: proc format; value group 0="control" 1="OA" 2= "LSS" ; run; proc glm data=my data; class group; title "Gait Speed Symmetry Index Distribution"; model speed_SymIdx_iqr = group/effectsize alpha=0.05; Format group group.; LABEL speed_SymIdx_iqr='Gait Speed Symmetry Index'; run; quit; ... View more

Hi for the attached graph, my axis labels are cut off for the middle and third rows. how do I prevent this? thank you.   here is the code I'm using:  proc sgscatter data=mydata; title "Spatial Gait Symmetry Parameters vs. SF-36 Functional Scores"; compare x=(PF VT GH) y=(speed_SymIdx_iqr cadenceSymIdx_iqr peakswing_SymIdx_iqr maxTC2_SymIdx_iqr) /loess=(nogroup) group=group markerattrs=(symbol=CircleFilled size=14) filledoutlinedmarkers colormodel=(CX3288BD CX99D594 CXE6F598 CXFEE08B CXFC8D59 CXD53E4F); LABEL PF='Physical Function' VT='Vitality' GH='General Health' speed_SymIdx_iqr='Strike Speed(IQR)' peakswing_SymIdx_iqr='Peak Angular Velocity_IQR' maxTC2_SymIdx_iqr='2nd Max-Toe Clearance_IQR'; format group group.; run; quit;   ... View more

For the attached graph, I want to change the Y-axis labels from the coding variable to reflect the full names. For the Y-axis minTC_ave-iqr = Minimum Toe Clearance (IQR) maxTC1_ave_iqr=Maximum Toe Clearance (IQR)   Same thing for the X-Axis. PF= Physical Function VT= Vitality/Energy GH= General Health   this is the code I used for the graph:   proc sgscatter data=mydata; title "Spatial Gait Parameters vs. SF-36 Functional Scores"; compare x=(PF VT GH) y=(minTC_ave_iqr maxTC1_ave_iqr ) /loess=(nogroup) group=group markerattrs=(symbol=CircleFilled size=14) filledoutlinedmarkers colormodel=(CX3288BD CX99D594 CXE6F598 CXFEE08B CXFC8D59 CXD53E4F); format group group.; run; quit;  How do I modify it so the final graph reflects the full names for the corresponding X and Y axis as indicated above ... View more

So I have these 4 graphs that I want to fit into one panel.   I created these graphs with the codes below: proc format; value group 0="control" 1="OA" 2= "LSS"; run; title2 "Double Support vs.Physical Function with loess fit"; proc sgplot data=mydata; loess x=PF y=DS_median /alpha=0.05 interpolation=linear degree=1 group=group; scatter x=PF y=DS_median / group=group markerattrs=(symbol=CircleFilled size=14) filledoutlinedmarkers colormodel=(CX3288BD CX99D594 CXE6F598 CXFEE08B CXFC8D59 CXD53E4F); xaxis grid; yaxis grid; xaxis label='Physical Function Score'; yaxis label='Percent Double Support'; format group group.; run; quit; proc format; value group 0="control" 1="OA" 2= "LSS"; run; title2 "#cycles_total vs. physical function with loess fit"; proc sgplot data=mydata; loess x=PF y=NrCycles /interpolation=linear degree=1 group=group; scatter x=PF y=NrCycles / group=group markerattrs=(symbol=CircleFilled size=14) filledoutlinedmarkers colormodel=(CX3288BD CX99D594 CXE6F598 CXFEE08B CXFC8D59 CXD53E4F); xaxis grid; yaxis grid; xaxis label='Physical Function Score'; yaxis label='Number of Total Gait Cycles'; format group group.; run; quit; title2 "Straight Cycles vs. physical function with loess fit"; proc sgplot data=mydata; loess x=PF y=NrCycleStraight /interpolation=linear degree=1 group=group; scatter x=PF y=NrCycleStraight / group=group markerattrs=(symbol=CircleFilled size=14) filledoutlinedmarkers colormodel=(CX3288BD CX99D594 CXE6F598 CXFEE08B CXFC8D59 CXD53E4F); xaxis grid; yaxis grid; xaxis label='Physical Function Score'; yaxis label='Number of Straight Gait Cycles'; format group group.; run; quit; title2 "Curved Cycles vs. physical function with loess fit"; proc sgplot data=mydata; loess x=PF y=NrCycleTurn /interpolation=linear degree=1 group=group; scatter x=PF y=NrCycleTurn / group=group markerattrs=(symbol=CircleFilled size=14) filledoutlinedmarkers colormodel=(CX3288BD CX99D594 CXE6F598 CXFEE08B CXFC8D59 CXD53E4F); xaxis grid; yaxis grid; xaxis label='Physical Function Score'; yaxis label='Number of Curved Gait Cycles'; format group group.; run; quit; Attached is the graphs that the above code produces. I would like all 4 graphs in one view/panel/page as opposed to how they are now on separate pages. I had considered doing a proc sgpanel data=mydata; but not sure how to proceed to combine the 4 separate codes as above. Any help much appreciated. thanks!  ... View more

Hi I need help plotting multilevel correlation in one plot. I want to plot correlation of speed vs. height for group 0,1, 2 and 3 all in one plot.  I tried the code below, but it gives me 3 separate plots.  proc corr data=mydata plots=matrix(nvar=all nwith=all); VAR height with Gaitspeed ; BY group; run; quit;   but I want just one plot with different color codes for the separate groups. any help much appreciated. thanks~ ... View more

PROC NPAR1WAY data=mydata wilcoxon; Class group; Var gaitspeed angle; label group='0'='control' '1'='OA' '2'='LSS' ; run; quit; when i run this code the output gives KS=0.37 Ksa=2.04 but there is no p-value given.  how do I get the p-value? how do I interpret the KS where class=3 levels as indicated by label 0 (control), 1(OA), 2(LSS)? ... View more