Can we determine the appropriate sample size for logistic regression using GPower even before we start to collect the data? If yes, how can we set the odd ratios (illustration in the figure)
With G*Power you need to have some hypothesis about the odds ratio, I mean, probably there is some evidence (or some kind of gold standard) in other research about the odds ratio that can be used when performing sample size calculations.
While G*Power is a good software, I recommend you to check this link:
http://powerandsamplesize.com/Calculators/
Here you can find some procedures with R, with formulae, explanations and details (even the R syntax!)
thank you for the help Dear Merce. Did you mean that I should use the odd ratios from other research with similar topic/model as mine? I am little bit confuse ho to set the odd ratios. Is there any rule of thumb that was commonly used derived from such a statistical book?
It is recommendable to check what is published and if there is some evidence about the odds ratio that must be fixed when computing sample size. This is an important issue in clinical trials, in other settings may be important and may be asked by reviewers.
I think there is no "rule of thumb", because it depends on the kind of population, the kind of study, expected results...