An approximately representative national sample of 2,972 male students at 32 U.S. institutions of higher education was surveyed regarding their use of several degrees of verbal coercion and physical force to obtain sexual intimacy with women without consent. The most severe form of sexual aggression each man reported was used to classify him into one of five groups: sexually nonaggressive, sexual coercion, sexual contact, attempted rape, or rape. Respondents also provided data that was grouped into three blocks of variables: early experiences (family violence exposure, childhood sexual abuse, age of sexual initiation), psychological characteristics (MMPI Scale 4, Hostility Toward Women, rape supportive beliefs, gender role orientation), and current behavior (alcohol use, pornography use, male bonding, sexual values and activity, conflict tactics). Data were analysed via blockwise discriminant function analysis. Variables were entered following a suggested development sequence. Specifically, all early experience variables were entered first as a block. Then the entire set of psychological characteristics were entered stepwise followed by all the current behavior variables. Variables from all three blocks entered the model. The classification rates have been discussed and the implications of the analyses for future causal models of male sexual aggression considered.