topic Re: sgplot vline jitter standard error bars in Graphics Programming

sgplot vline jitter standard error bars

vstorme — Wed, 25 Apr 2018 14:56:52 GMT

I would like to have non-overlapping standard error or confidence limits. This is my code

proc sgplot data = class4.dental;
vline age / response = y stat = mean limitstat = clm group=gender curvelabel;
run;

Thank you for any sugggestions

Re: sgplot vline jitter standard error bars

PeterClemmensen — Wed, 25 Apr 2018 16:44:58 GMT

Are you sure you want to present your data like this? See this short example, which I assume resemble yours

proc sgplot data=sashelp.class;
vline age / response=height group=sex stat=mean limitstat=clm;
run;

produces this

Already now, there are a lot of lines to grasp. And you want to offset the clm limits, adding more lines to the chart. I think you want to present your data in a more suitable manner? 🙂

Re: sgplot vline jitter standard error bars

ballardw — Wed, 25 Apr 2018 18:00:03 GMT

You might provide an example data set that will let us see what you are currently getting or include an image. Then describe which "overlapping" error bars or confidence limits are not desired.

"non-overlapping" could have a number of interpretations. For example with the image shown by @PeterClemmensen perhaps you would only want the widest confidence limit shown, or the narrowest for each age.

Another possible interpretation could be to not have male and female in the same graph.

Re: sgplot vline jitter standard error bars

Jay54 — Wed, 25 Apr 2018 18:27:21 GMT

Extending on Draycut's reply, use GROUPDISPLAY=CLUSTER and CLUSTERWIDTH=0.2 to get this result.

proc sgplot data=sashelp.class;
vline age / response=height group=sex stat=mean limitstat=clm markers
 groupdisplay=cluster clusterwidth=0.1 arkerattrs=(size=5 symbol=circlefilled);
run;

Re: sgplot vline jitter standard error bars

vstorme — Wed, 25 Apr 2018 21:31:50 GMT

Thank you, this is what I want!

Re: sgplot vline jitter standard error bars

Leo9 — Wed, 25 Apr 2018 22:48:43 GMT

Thank you