What would be interesting is if a random set of the unaccepted got the same opportunities as the accepted. But that's not possible...

Instead what yc needs to do is accept a control group and see if there is any correlation between being chosen and eventual success. Then again...