Posthoc testing after ANOVA

[Sateesh Shankaranarayana]

Dear Statalist team,

I am interested in doing Posthoc testing after ANOVA.

I only have aggregate data from the journal article. I have used "aovsum" package to do the ANOVA test. How can I do either Tukey or Dunnette's method to compare 3 different means?.

Any help is greatly appreciated.

With warm regards,
Sateesh

[Joseph Coveney]

I assume that -aovsum- is a user-written command. If it doesn't post the return information that is required for -pwcompare-, then you'll probably have to compute the ranges yourself. Stata has functions for both.

help density_functions##dunnett
help density_functions##tukey

In the context of all-pairwise comparisons, see this post and the immediately following one (#7) in that thread for points to consider when you have only three comparisons to make.