pmid:30365010 23
Basically, by knowing how large an effect would be meaningful (e.g., how much improvement would a drug need to give to be worth taking) and what likelihood of making a type I error is acceptable, one can determine how many subjects need to be in a study
sisters: Linda (Steve) York of Independence, Missouri, Paula Griffith of Arthur, Tina Stanley of Louisville, Mindy (Rick) Green of Flora, Paula (Terry) Young of Flora, Cindy (Alan) Hefter of Independence, Missouri, and Sam (Kenny) Sanders of Olney
Evans, W, M C Farrelly, and E Montgomery (1999), Do Workplace Smoking Bans Reduce Smoking? American Economic Review 89(4): 728-747