Ok, I'm doing this particular study on A, and A consists of 3 subgroups; lets call these subgroups as A_1, A_2 and A_3.

The [B]total population[/B] for A_1, A_2 and A_3 is [B]300[/B].
However, the distribution of each of this group is not equal, e.g. A_1 is [B]100[/B], A_2 is [B]25[/B] and A_3 is [B]175[/B].

Is it possible to use simple random sampling to calculate a sample size for each subgroups or should the sample size be calculated as whole?

*bump*