Hi listers,

I was asked to estimate the sample necessary to detect a difference in the average number of online counselling sessions between 2 groups so the outcome is a count and I plan to use Poisson regression to analyse the data and also estimate the sample size.

The -power twomeans- command does not seem appropriate here so I am wondering whether I can simulate the data in Stata. Can I use rnormal to simulate the data (I would need to truncate any negative value to be 0) or is there any way to create a Poisson distribution?

Any input is welcome!