Sorry if this question is poorly composed due to lack of stats knowledge. Any advice to point me in the right direction would be greatly appreciated:
I'm hoping to detect if a dataset originated from the same or different populations as another dataset. (If we assume different subjects behave very differently, while the same subject always behaves somewhat consistently). Each dataset had the same test conditions, which consist of 10 repetitions of 100 conditions.
The raw data is in the form of time-stamps of discrete events, the frequency of which varies with test condition, and is extracted for analysis. The histogram of each dataset for all 10x100 conditions, binned by frequencies, fit the gamma distribution (I just tested distributions that seemed relevant with Matlab's chi2gof function). The population size of each dataset is about 1-4 (exact number unknown) and their outputs are mixed together and cannot be reliably separated.
I'm wondering if I could use distribution shape to detect population changes, but which of the fitted parameters to use and how do I compare them?
