Sorry, I forgot to get with you on this. I've put together a little example below. The tricky part is not the graph template -- it's getting the data correct to draw the error bars in the correct location. In the example below, the stderr is computed in PROC SUMMARY, but is later modified to position the error bar correctly at the top of the bar. Let me know if you have any questions about it.
Thanks!
Dan
/* First, all data used for the overlay must be summarized */
proc summary data=sashelp.class nway;
class age;
var height;
output out=limits sum=limitsum stderr=stderr;
run;
proc summary data=sashelp.class nway;
class age sex;
var height;
output out=subgroups sum=barsum;
run;
data limits2;
rename age=age2;
set limits (keep=age stderr limitsum);
ustderror = limitsum + stderr;
lstderror = limitsum - stderr;
run;
data merged;
merge subgroups limits2;
run;
/* The graph template */
proc template;
define statgraph barlimits;
begingraph;
layout overlay;
barchartparm x=age y=barsum / group=sex primary=true;
scatterplot x=age2 y=limitsum / yerrorupper=ustderror yerrorlower=lstderror markerattrs=(size=0);
endlayout;
endgraph;
end;
run;
proc sgrender data=merged template=barlimits; run;
