I want to take an undertake an experiment for number of elbow fractures in our hospital
than what sample size should I take ?
there are huge number of fracture cases coming to hospital so how should I select sample size from that and make study for it?
It's not clear what you're trying to do here. Do you just want to know how many cases are coming in to your hospital? Aren't there records for that? What exactly is it you want to "study"? In any case, you're going to need to know -- or at least have an educated guess for -- further information. Have a look here for starters http://powerandsamplesize.com/Calculators/