GPLOT is pretty old and has some esoteric behavior. SGPLOT is, generally easier to use and is where many new features have been added.
This is likely where I would start. First, note the data step. That is the preferred way to provide data on this forum so we do not have to guess as to variable names and properties may be. Most of places that allow you to set an error bar in the Sgplot code want an explicit upper and lower value for the plot. So the data step also allows us to add those.
The Sgplot is using a lot of default information such as the colors and markers of your current ODS Style. The X= and Y= are the axis that a variable is plotted against. Everything after the / are options. Group is how to specify a different plot for each level of another variable. The Scatter statement allow display of error-bars using the calculated variable.
I also show use of SGPANEL to create a separate graph for the levels of the BHB_level because your error bars overlap.
data have;
input Dam_DIM BHB_level Estimate SE;
SeUpper = estimate+SE;
SeLower = estimate-SE;
datalines;
40 0 2.3476 0.4174
40 1 2.171 0.2936
40 2 2.5354 0.854
40 3 1.11 1.7042
41 0 2.2681 0.3165
41 1 2.1614 0.2604
41 2 1.7039 0.9852
41 3 2.5988 1.2062
42 0 2.9415 0.3173
42 1 1.9979 0.2374
42 2 1.0917 0.7625
42 3 1.647 0.9816
43 0 1.8898 0.2728
43 1 2.4978 0.2094
43 2 1.1158 0.9858
43 3 2.7079 0.7616
44 0 2.5648 0.2558
44 1 2.0557 0.211
44 2 1.8988 0.6973
44 3 1.5242 0.5717
45 0 2.1302 0.2377
45 1 2.1301 0.176
45 2 2.3225 0.542
45 3 3.0647 0.8544
;
proc sgplot data=have;
/* if you want the points connected use this*/
series x=Dam_DIM y=estimate / group=BHB_level
;
scatter x=Dam_DIM y=estimate / group=BHB_level
yerrorlower=SeLower yerrorupper=SeUpper
;
run;
OR
proc sgpanel data=have;
/* if you want the points connected use this*/
panelby bhb_level/columns=1;
series x=Dam_DIM y=estimate /
;
scatter x=Dam_DIM y=estimate /
yerrorlower=SeLower yerrorupper=SeUpper
;
run;
