Statistical over/under representation of features (comparison of 'study set' against 'population set')