analysis of correlated (?) or clustered (?) binary data

Hello NG,

I am reflecting the following problem:

given a sample of e.g. humans, one can be interested in the rate and
the corresponding confidence interval of a certain event.

The difficulty is, that this event may occur several times, one time or
not at all within one person.

That means that events within one person are correlated, whereas
between persons they are independent.

I heard about GEE, but isn't that technique like breaking a butterfly
on a wheel?

Does anybody has hints or references on how to handle that?

Thanks in advance

Uli

.

Loading