I was thinking that it could be easily done in the data step, is all. After reading up on the topic further, it now appears that perhaps the best answer is to do Proc BoxPlot with all groups, and include the option of creating OUTBOX or OUTHISTORY datafile. Then in a subsequent run, use one of these new datafiles as the new input. Still not sure, though, of the exact coding for keeping the top 20 groups. See: SAS/STAT(R) 9.2 User's Guide, Second Edition proc boxplot history=Summary;   plot (Weight Yieldstrength) * Batch;   run; Since another datafile is being used in this subsequent run (history), and that datafile contains a column for, say, Q3, one would use a new data statement to tailor the history data file -- sort on the Q3 column, descending order, and use OBS=20, so that only the top 20 'groups' will be included, and then used.  Result should be ONE panel of box plots of the top 20 groups. ... View more

Thanks Anca.  Correct you are.  However, there are aspects to the ODS graphs that are quite nice, thus I'd like to keep ODS on. ... View more

Okay, tried that, simply adding title;, and no difference.  Title still there. By the way, the reason I'd like to get rid of the title is have more room for the graph. Here is the code that produces the boxplots: proc boxplot data=nicholas.boxplots; plot Change*Single / boxstyle=skeletal nohlabel boxconnect=q3 cconnect=blue boxwidthscale=1 clipfactor=1.5 clipsymbol=dot ; insetgroup n; run; ... View more

Once again with SAS, not as simple as "No title, please." Good grief, why do they make things so difficult. ... View more