Stereotypes to indeed exist for a reason. However, the populations being looked at have a very large sample size. A deviation of just 1% would refer to hundreds of thousands, if not millions of people. It would be unfair to treat those deviants according to the rest of the 99%.