Hi all,

I’ve been asked to run an ANOVA with a Holm correction. I need to compare Stigma (continuous outcome) across five different survey versions (categorical independent variable).

I understand that in this case, I can employ the -oneway- command:
oneway Stigma Survey
And then use an option like -bonferroni- to do a correction.
But it seems like in Stata, in order to get the Holm correction, I need to employ the -anova- command:
anova Stigma Survey
And then I need to use the -test- command with the -mtest(holm)- option. Here is where I get a bit confused. It seems like I need to make a matrix to refer to when I use -test-. I just don’t understand how to make this matrix. I would like to compare Stigma across the five survey versions. How can I do this with a Holm correction?