Sample size for an interaction effect

[Aron Sen]

Hello!
I am designing a cluster randomised trial and want to estimate sample size for an interaction effect between two binary variables. The interaction term is the treatment effect of interest. Could anyone help me with the best way to do this in Stata please?

The trial outcome is a continuous measure and the effect size for sample size calculation is Cohen's d = 0.2. Other info is below:
Cluster size = 10
ICC = 0.05
Power = 80%
alpha = 5%

I would appreciate some advice.

Thanks
Aron

[Phil Bromiley]

Stata 15 documentation has a full book on power and sample size. Have you checked it?

[Aron Sen]

Thanks, Phil. I did but couldn't find anything specific examples in relation to interaction effect. Any advice would be appreciated. Thanks.

[Phil Bromiley]

Interactions are just variables, although they are usually correlated with the other variables which will increase your needed sample size.