Defining the p-factor: an empirical test of five leading theories