That's because your interview process is already biasing the result. To get a fairer comparison you should take a random sample from both groups (out of all the

That's because your interview process is already biasing the result. To get a fairer comparison you should take a random sample from both groups (out of all the applicants) and see (although self selection bias is still in play here).