Sex Biases in the Measurement of Moral Judgment

Abstract

The most commonly used method for measuring levels of moral reasoning has been the Kohlberg/Rest technique of presenting subjects wirh a story in which the main character is faced with a moral dilemma (2, 3). Holstein (1) has argued that this method may be biased in favor of men, since all of the main story characters are males. The present study is designed to determine whether altering the sex of the story characters on Rest's Defining Issues Test differentially affects male and female subjects' test scores. A total of 84 freshman men and women completed either the standard short form of the test (N = 38) or a sex-reversed form (N = 46). For this latter form, each of the main story characters was a woman. Subjects also responded to questions asking them to report how believable and realistic the story characters' acrions were, and how easily the subjects could identify wirh the main story character. In scoring the protocols, a P-score is computed for each subject. The P-score is the percentage of items representing principled reasoning (Kohlberg's Stages 5 and 6) rated as important considerations by the subject. P-scores for the present sample 'ranged from 3 to 70. The mean score for the regular form was 33.6 (SD = 18.1), while [he mean score for the sex-reversed form was 32.0 (SD = 16.2). Application of a 2 (sex of subject) X 2 (form of test) analysis of variance yielded non-significant main effects for sex and form. - The interaction, however, was significant (F,,so = 4.05, ? = .05). Examination of the means showed differences between groups in the non-predicted dlrectlon. Males who took the form with female characters (sex-reversed form) had higher P-scores than males who took the regular rest with male characters (134 = 36.2, SD = 18.6 for the sex-reversed form, and M = 30.2, SD = 19.9 for the regular form). However, females who took the regular version had higher P-scores (M = 37.0. SD = 16.3) than females who took the test with female story characters (M = 27.9, SD = 13.9). Thus, individuals who read dilemmas with main characters who were the opposite sex had higher moral judgment scores than those who read stories in which the main characters were the same sex as themselves. The analysis of the subjects' responses to the believabiliry and identifiabiAity items showed no significant differences associated with subject sex or story character sex. The present results do nor support Holstein's (1) contention that tests such as the Defining Issues Test are biased in favor of males. The results do suggest, however, that the sex of the story characters can be a confounding variable in measuring moral judgment on this test. Research should attempt to replicate and extend these preliminary findings.