Perfect! You're an angel!   Thanks! (sorry for all the exclamation marks!)  ... View more

Another guess: when you use PROC TRANSPOSE, the new variables are given a default label. If you don't like the label on the axes, you can use the XAXIS and YAXIS statements to change them:   proc sgplot; scatter x=x y=y; xaxis label="My X label"; yaxis label="My Y label"; run;   ... View more

Hi @Rick_SAS,   I'm actually exploring ways to visualize the presence/absence of heteroskedasticity of residuals in a linear model. It seemed a good idea to do that using boxplots of bins of equal counts across the range of the linear predictors (with smoothed connecting curves to emphasize the general overall pattern). I'll have a look at quantile regressions.    Thanks for the tip! ... View more