1
$\begingroup$

I'm preparing a test of proportions regarding a campaign redemption of several groups which have all common "demographics" but have been stimulated into different ways.

At the same time I have a big group of people with the same demographics but which have not been stimulated into any way.

My question is, does it make sense to use the whole control group several times as control for each different test group, or instead should I split the control group into different sub-groups and then use each one in a one-to-one way with each test group?

$\endgroup$
1
$\begingroup$

If you use the whole control group each time then your tests will not be independent. This will almost certainly violate the assumptions underlying the tests. Splitting the control group in the way you suggest will overcome that problem. It is a technique which is often used in meta-analysis although there are nowadays usually better ways in meta-analysis of dealing with the issue.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.