Well, you still have a self-selecting group. Unless he randomly accepts some of the applicants and compares them to the other applicants, it would still be a fl

Well, you still have a self-selecting group. Unless he randomly accepts some of the applicants and compares them to the other applicants, it would still be a flawed experiment (assuming he wants it to be a proper experiment, which I don't think is the point).