https://www.statmethods.net/stats/power.html

Power analyses

Just as a reminder sample sizes and power are all calculated the same way(just using basic algebra and statistics): eg see: https://www.google.com/search?q=Basic+sample+size+calculation&rlz=1C1CHBF_enUS874US874&oq=Basic+sample+size+calculation&aqs=chrome..69i57.46284j0j4&sourceid=chrome&ie=UTF-8

That is stnd error(se) is a function of n and se is proportional to l= the desired length of a confidence interval. a CI is often statistic +/- const (se) with const depending on the accuracy desired. Thus const*se=l, now solve for n.example for the basic binomial case is given at( https://www.google.com/search?q=Basic+sample+size+calculation&rlz=1C1CHBF_enUS874US874&oq=Basic+sample+size+calculation&aqs=chrome..69i57.46284j0j4&sourceid=chrome&ie=UTF-8 ). Then from the definition of power one can calculate n required for a particular power. just use the se for the test you want to apply. eg see ( https://www.google.com/search?rlz=1C1CHBF_enUS874US874&ei=cE4qXquhC9qmtQaUr6P4Bg&q=Basic+statistical+power+calculation&oq=Basic+statistical+power+calculation&gs_l=psy-ab.3...14313.20625..22280...0.2..0.315.1733.7j0j4j1......0....1..gws-wiz.......0i71j0i67j0i30j0i8i30j0i8i7i30j0i8i7i10i30j0i333j33i10.We3aDcNMEoQ&ved=0ahUKEwirmf3akJvnAhVaU80KHZTXCG8Q4dUDCAs&uact=5 )

Thus one can do the sample size calculation even without having a program to do it. this eactly how the program works. Know the se of your test and the accuracy or power desired and anyone can compute desired sample size. HTH D.E. Booth

PS if there is more than one test to do calc all of the n's and take the largest.

Rahel Yaekob

you have questioned the same thing twice.

https://www.researchgate.net/post/I_planned_to_conduct_case_control_study_and_I_find_difficulties_in_calculating_samples_size