The goal of power analysis is to decide (a) how large a sample is needed to enable statistical judgments that are accurate and reliable, and (b) how likely your statistical test will be able to detect the effect of a given size in a particular situation.
In the literature, there is no consensus on the approach to compute the power and sample size with logistic regression. In this talk, I am going to review three test statistics used in the asymptotic way. For the implementation, I follow the procedure advocated by Demidenko(2007). As part of the WebPower project, online software is developed and will be demonstrated at the end of the talk.
For more information about WebPower, please visit http://webpower.psychstat.org/wiki/.