I'm trying to calculate the minimum detectable effect in an experiment after n samples.
I'm able to use the
pwr package like this to compute the minimum detectable effect.
library(pwr) pwr.2p.test(n = 10000, sig.level = 0.05, power = 0.8, alternative = "two.sided") pwr.2p.test(n = 50000, sig.level = 0.05, power = 0.8, alternative = "two.sided")
However, this returns the MDE in terms of
Difference of proportion power calculation for binomial distribution (arcsine transformation) h = 0.0177296 n = 50000 sig.level = 0.05 power = 0.8 alternative = two.sided
The base rate in the control group of the experiment is 0.10 or 10%. How can I convert the
h MDE to a percent difference in the rates between the groups?