PROC ANOVA MEANS to get p-values for comparison

[lizzy28]

Hi all,

Thank you for your attention first of all!

I am running PROC ANOVA with MEANS statement to get group means as below:

proc anova data=indat;

class classvar;

model cost = classvar;

means classvar /tukey;

run;

The Tukey test generates the table of difference between means along with 95% CI. However, I need to have specific p-values per difference between means.

Is there a way to list out the specific p-values?

Thanks a lot!

Lizi

[Rick_SAS · Accepted solution]

It sounds like you want an all pairwise comparison of the mean differences. You can use the LSMEANS statement in PROC GLM and the PDIFF=ALL option, as shown in the article "The diffogram and other graphs for multiple comparisons of means." The article includes an example, links to the documentation, and even a discussion of how to visualize the mean differences.

[lizzy28]

Thanks a lot!