All I have two sets of data. One where people bought and another where they did not. For each sample in the two sets, I have ~3000 binary independent variables. Each dataset has about 1000 samples. The range of positives in the independent variables is 5% through 80%.

The question is: how do I figure out which of the 3000 binary variables has a statistically significant effect on the outcome. I tried using Fisher Exact Test and that does not seem to be giving intuitive answers.

I am using Python and Scipy for the tests. Any and all help most welcome. Thanks

  • $\begingroup$ Have you tried Barnard's test? $\endgroup$ – Arthur B. Nov 13 '14 at 1:43
  • $\begingroup$ Have not. Will try it. For the Fisher test, I picked each independent variable separately and ran a Fisher test. Is that the same I'd do with Barnard's? Based on a quick search online, it seems to be so. $\endgroup$ – ppwt Nov 13 '14 at 5:17

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.