Looking for confirmation that I am correctly generating the CI for a mean

[sasguy299]

Hi, I am using the "output" statement in "proc means" in  SAS 9.4 to generate means and 95% CI's for a particular numeric variable for each level of a "by" variable. I have written the following "output" statement:

output out=X N=N mean=mean lclm=lclm uclm=uclm;

My question: Does this code correctly produce the appropriate two-sided confidence intervals for each mean (ie. two-sided and using the appropriate t-value corresponding to the N for each "by" group, as some of the "by" groups are quite small)? I 'd like to confirm that it does because, unlike the documentation for other procedures in 9.4 that explicitly state that "lclm" and "uclm" used in conjunction with one another does in fact do this (while producing only one-sided CI's if only one of the two is specified), I did not notice such a comment in the 9.4 documentation for "proc means".

Thanks!

[ballardw · Accepted solution]

Short answer: Yes.

Depending on which documentation the Usage section has a bit more on Confidence Limits to reassure.

[sasguy299]

thanks!