How to compare two frequency distributions - (Nov/14/2013 )
I would like to compare two frequency distributions, to know whether they are statistically different. I've seen a number of tests that have been used in papers
(e.g. Fisher's exact, Chi squared, Kolmogorov-Smirnov,...), but I have some doubts on their validity.
Could anyone of you please help me out here?
Try looking here: http://udel.edu/~mcdonald/statintro.html
As far as I know all mentioned tests are good and reliable (and accepted by journals when you publish). Today Fisher's exact test seems to be preferable especially for smaller sample numbers, with larger sample numbers there is no difference to Chi squared test.
Another test which is highly recommended now is the G-test, which I'd use, if the software offers this one (for R a package is available) and if you have not too small sample numbers (too).