McNemar's test post-hoc power analysis problem when odds ratio=0

#1
I need to do post-hoc power analysis for a McNemar's test (using G*power), and am having trouble getting results. One of the discordant pairs is 0, therefore the odds ratio is zero, and it seems that power cannot be calculated using an odds ratio of 0. Is there any way around this?

Nevermind responding with suggestions to not do post-hoc analysis - I realize there are arguements against it - I have been instructed to do this.

Thank you for your time!
 

hlsmith

Not a robit
#2
Sometimes a continuity correction of 0.5 can be added to all cell counts. This can be used to get the proxy calculation.