Sample size calculations…they are the bane of my existence! Technically it should be so easy. Choose the alpha, the power, punch in some other relevant numbers to determine effect size and waza! A number is spit out that is the number I need to recruit! However, I find that the actual process isn’t that easy, particularly when you are doing research using paradigms that haven’t been tested before (or in your area with your specified outcome) and when choosing the clinically important difference as the guide to effect size isn’t an option.
One of the things that I have found very helpful has been to use the program called GPower which you can download for free here: http://www.psycho.uni-duesseldorf.de/abteilungen/aap/gpower3/ There is also a paper guiding how to use this program which is a bit technical but helpful.1
What I like about this program is that is allows you to choose which statistical test you will be using (ie, F test, t test, exact test, etc…) and also the specific type of that test (ie, paired t-test, MANOVA, repeated measures within, etc..). Then it guides you through the information that you need to input in order to calculate the needed sample size. This also includes a side bar that allows you to calculate the effect size (what I think is the hardest part) by filling in the appropriate boxes. This has been a real help for me when I have to calculate the sample size I need for analyses that I am not quite clear on as it forces me to find the relevant information. Last, the program provides you with a print copy (similar to SPSS syntax) that you can save so that you actually remember how you got the sample size that you did!
GPower also allows you to perform different types of power analyses – for example, a priori analyses like I described above where the aim to determine the sample size. It also provides the opportunity to do posthoc power analyses when you know the effect size, alpha, and sample size. There are other options as well, but I can handily confess that I’ve never used these!
Regardless, I thought that I would share this resource with people in case you find yourself also wanting to bang your head against the wall when attempting power calculations!
1. Faul F, Erdfelder E, Lang A-G, Buchner A. G*Power3: A flexible statistical power analysis program for the social, behavioural, and biomedical sciences. Behavior Research Methods 2007;39:175-191.