Hi there,
I was going through some Stata work in my MSc and was recommended to use the 'clustersampsi' command for clustered RCTs to calculate sample size per arm and number of clusters per arm. When I try to use this similarly for cluster calculation for something like a survey where clusters may be 100-250 households, it gives a low number of clusters and high sample size per arm when I put, for example, 278 average cluster size. Is this command not necessarily useful in these cases? Am I misinterpreting the inputs required?
Thanks,
-B
I was going through some Stata work in my MSc and was recommended to use the 'clustersampsi' command for clustered RCTs to calculate sample size per arm and number of clusters per arm. When I try to use this similarly for cluster calculation for something like a survey where clusters may be 100-250 households, it gives a low number of clusters and high sample size per arm when I put, for example, 278 average cluster size. Is this command not necessarily useful in these cases? Am I misinterpreting the inputs required?
Thanks,
-B
