# Power/Sample Size Calculation for Logistic Regression with Binary Covariate(s)

This program computes power, sample size, or minimum detectable odds ratio (OR) for logistic regression with a single binary covariate or two covariates and their interaction. The Wald test is used as the basis for computations. We emphasize that the Wald test should be used to match a typically used coefficient significance testing. You need to fill in two fields and the third leave blank. For example, if you provide values for sample size and detectable OR the power will be computed. This program can be used for case-control studies. Then Pry is simply means the proportion of cases in the total sample.

These algorithms are described in Demidenko E. (2007). "Sample size determination for logistic regression revisited." Statistics in Medicine 26:3385-3397 and Demidenko E. (2008) "Sample size and optimal design for logistic regression with binary interaction." Statistics in Medicine, 27:36-46

Significance level =

What to compute:
Power =
Sample size =
Detectable/alternative OR =

One variable with exposure, x:

Prx=Pr(x=1)= Pry=Pr(y=1|x=0)=

Two variables with exposure, x and confounder, z:

Prx=Pr(x=1)= Prz=Pr(z=1)= ORyz= ORxz=
Disease prevalence rate Pry=Pr(y=1|x=0,z=0)=

Two binary variables, x and z, with their interaction, x*z:

Prx=Pr(x=1)= ORyx= Prz=Pr(z=1)= ORyz= ORxz=
Disease prevalence rate Pry=Pr(y=1|x=0,z=0)=