sample size for control and treatment groups



hello everyone,

I am in the process of designing an agricultural experiment in which I want to compare a series of variables between a group of manufacturers who use a mobile telephone app (n=15,000) is my treatment group and another group that does not (n=100,000) which is my control group.

What would be adequate sample sizes for both control and treatment? I have looked at my stats books, but haven't succeeded to find any conclusive answer. Any help is most appreciated.
Firstly, well done on stopping by here before you start your experiment, it's much easier (and more meaningful) to do statistics with an experiment that has been designed with the tests and hypotheses in mind.

Of the top of my head I think this provides a better summary on the matter than I could!

Hope that helps.

[edit] If you need something more detailed I can personally vouch for the (UK) Open Univeristy's book "Medical Statistics" (M249 Book 1) covering the matter quite well (And does more to cover experiemental design) - don't be put off by the title, although it discusses the matter in a medical context the theory behind it is the same no matter what field it's applied to.
Last edited: