Effect size (delta) for three group experiment

[INES ALEJO FERNANDEZ]

I want to conduct a three group experiment, and I need to know the expected effect size. I have explored the power option but cant find the way to adapt it to the 3 group design.

My desired power is 0.8
Number of participants 400
Alpha 0.05
Groups 3 --> 3 (2 treatment groups and 1 control)

I am trying to analyse to what extent two kind of messages impact people's attitudes toward migrants, so comparing the two treatments against the control, no interest in seeing how they interact.

Can someone help me maybe?

[Rich Goldstein]

I think you want:

h power oneway

where you will find a section on computing effect size; here is a snippet:

Compute effect size and target between-group variance

power oneway, n(numlist) power(numlist) ngroups(#) [varerror(
numlist) options]

I note however that this may not be what you want; in some ways it sounds like you want two tests, each against the control, and therefor you could get by with

h power twomeans

but of course, you might need to do it twice if the 2 treatments differ and you might want to adjust the alpha level for the multiple tests