Anova and pairwise comparisons

[Adrienne Wold]

Hello,

I am running the following:

reg y i.x z, r

and x has four levels including the base level. The levels are not of equal sizes. Along with testing the effect of each level relative to the base-level, I want to test whether the coefficients for the remaining levels are different from each other.
I run

anova y x

rather than t-tests given multiple comparisons, but what is the appropriate way to check for pairwise comparisons here?

[Jeff Pitblado (StataCorp)]

See docs for pwcompare.
https://www.stata.com/manuals/rpwcompare.pdf