RETURNING MATERIALS: P1ace in book drop to remove this checkout from your record. FINES wil] be charged if book is returned after the date stamped below. MSU LIBRARXES . ‘ d ' . D #111." “-4‘.‘ ‘ L :12" ' ./ FEBZ . ‘. oz 22902 ‘210 SEXISM IN WOMEN'S JUDGMENTS OF ARGUMENTS BETWEEN WOMEN AND MEN By Sandra Kay Pinches A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Psychology 1978 ABSTRACT SEXISM IN WOMEN'S JUDGMENTS 0F ARGUMENTS BETWEEN WOMEN AND MEN By Sandra Kay Pinches Naturalistic observations, feminist psychological theories and a substantial body of research suggested that women are more loyal to men than to other women. In the case of an argument between a woman and a man, women with traditional sex-role ideologies were expected to side with the man, while pro-feminist women were predicted to side with the woman. Groups of female college students received one of three pro- jective leads briefly describing a lovers' quarrel. One lead included only the information that the argument was occurring, one lead spe- cified that the argument outcome was negative in terms of the couples' relationship, and the third lead specified that the argument outcome was positive. The subjects wrote a TAT-type story in response to the projective lead, and subsequently rated the story characters on a Likert-type Story Character Evaluation Scale (SCES) designed for this study. (Reliability coefficients for SCES subscales were high.) The subjects also completed a demographic questionnaire and the Spence Attitudes Toward Women Scale (AWS). Sandra Kay Pinches Subjects were classified as pro-feminist liberals, moderates or traditionals, based on their AWS scores. A multivariate analysis of variance was performed, using the SCES scores as dependent vari- ables. The results were that women favored the woman over the man re- gardless of subjects' sex-role ideology and regardless of what they were told about the outcome of the argument. Groups of subjects did not differ significantly from each other in degree of bias. Pro- female, anti-male bias was found in attributions of credit and blame for the argument. In ratings on positive and negative personality traits subjects rated both the woman and the man favorably, but rated the woman more favorably than the man. Many subjects regarded the argument itself as an event which should have been prevented, irrespective of the outcome specified in the projective leads. The belief that the argument should have been prevented was correlated positively with attribution of blame to the woman, but was not related to ratings of the man. These correlations provided modest evidence that women are blamed for arguments, despite the pro-female favoritism found in the rest of the data. The results were consistent with those of a small number of studies showing pro-female favoritism among women. A greater number of studies, however, show that women have pro-male and anti-female attitudes. Possible explanations for this inconsistency in reported data include cultural changes, variations in the stimuli eliciting prejudice, variations in sample characteristics, and differences between women's self-ratings and ratings of other women. ACKNOWLEDGMENTS I would like to thank Elaine Donelson, Chairperson of my Dissertation Committee, for her generous contributions of knowledge, time and energy to my research. Elaine's support and her willing- ness to be avaialble for consultation enabled me to persevere past moments of discouragement. I would also like to thank Don Grummon, Al Aniskiewicz and Jeanne Gullahorn, who served as committee members. Al Hammer, who was my statistical and computer consultant, was very helpful during the analysis stage of the research. Al and other members of the Office for Research Consultation gave special attention to my dissertation, and enabled me to relate in a productive way to the computer. Finally, I would like to acknowledge the contributions of the friends who assisted with typing, co-rating, and other necessary tasks. Most of all, I would like to thank the people who continued to believe in me during my long struggle to finish graduate school. ii TABLE OF CONTENTS Page LIST OF TABLES ......................... vi LIST OF FIGURES ........................ viii INTRODUCTION .......................... l COMPONENTS OF THE PROBLEM ................... 7 Definitions ....................... 8 LITERATURE REVIEW ....................... l3 Hostile Humor ...................... 15 Physical Conflict .................... 20 Sex-linked Stereotypes .................. 24 Traditional Femininity and Masculinity ........ 24 Sex-Role Stereotypes and Mental Health ........ 3O Androgyny ...................... 35 Success and Failure ................... 36 Competition and Conflict with Men ............ 4l Summary of Previous Research ............... 45 DYNAMIC FORMULATIONS ...................... 48 A Cross-Cultural Perspective ............... 53 Roles of Women in Mythology ............. 57 Statement of Purpose ................... 6O PILOT STUDY .......................... 6l Instruments ....................... 63 Bem Sex-Role Inventory ................ 63 Projective Instrument ................ 66 Demographic Questionnaire .............. 69 Procedures ........................ 69 Introduction to Experiment .............. 69 Subjects ......................... 7O Page Results .......................... 72 Scoring ........................ 73 Discussion ........................ 80 METHOD ............................. 85 Design .......................... 87 Instruments .............. , .......... 88 Projective Instrument ................. 88 Story Character Evaluation Scale ........... 9O Inter—Rater Reliability of the SCES .......... 94 Attitudes Toward Women Scale ............. 98 Demographic Information ................ 103 Subjects ......................... 103 Procedures ........................ 104 Hypotheses ........................ 105 I. AWS x_Sex Interaction .............. 107 II. Argument Outcome x_AWS x_Sex Interaction . . . . 108 III. Secondary Hypotheses .............. 109 RESULTS ............................. 111 Description of the Subjects ................ 111 Demographic Data ................... 111 Attitudes Toward Women's Roles ............ 112 Content of Projective Stories ............... 117 Outcome of Argument .................. 117 Causes of Argument .................. 120 Story Character Evaluation Scale ............. 126 Multivariate Analysis of Variance ............. 134 Analysis of SUM Scores ................ 146 Second Analysis .................... 146 DISCUSSION ........................... 154 Summary of Primary Findings ................ 154 Theoretical Implications ................. 157 Explanations of Findings ............... 159 Summary ........................ 167 Attraction on the Basis of Similarity ......... 168 Causal Attribution for Positive and Negative Outcomes ..................... 170 Differences Between Pro-feminist and Traditional Women ....................... 175 iv Page Assessment of the Study ................. 184 Projective Instrument ................ 184 The Attitudes Toward Women Scale .......... 188 Sex of Researcher and Group Composition ....... 189 Impact of the Sample Characteristics ........ 192 Directions for Future Research ............. 195 SUMMARY ............................ 200 REFERENCES .......................... 204 APPENDICES .......................... 212 A BEM SEX ROLE INVENTORY ................. 121 B SAMPLE TEST FORM WITH SAMPLE STORY ........... 214 C PILOT STUDY EXPERIMENTER'S MANUAL ............ 215 D PARTICIPANTS' BACKGROUND DATA .............. 219 E STORY CHARACTER EVALUATION SCALE ............ 220 F ATTITUDES TOWARD WOMEN SCALE .............. 224 G BACKGROUND INFORMATION ................. 228 H EXPERIMENTER'S MANUAL .................. 229 I STORY ABOUT JEALOUSY .................. 231 J STORY BY BLACK SUBJECT ................. 232 Table 1a. lb. £0me 10. ll. 12. 13. 14. 15. 16. 17. 18. LIST OF TABLES Page BSRI Means for Stanford and Foothill Samples ....... 65 Sex—typing of Bem Subjects ................ 66 Pilot study: Main Effect ................ 75 Pilot Study: Effect of Femininity vs Non-Femininity of Subject ......................... 76 Effect of Experimental Condition ............. 78 Effect of Femininity of Subject x_Experimenta1 Condition . 79 SCES Scales ....................... 95 Description of SCES Clusters ............... 96 Outline of Hypotheses .................. 107 Distribution of AWS Scores ................ 112 Range of AWS Subtotal Scores ............... 113 Distribution of Argument Outcomes ............ 118 Distribution of Jealousy Stories ............. 121 Kuder-Richardson Reliability Coefficients for SCES Scales: First Version .................. 127 Kuder-Richardson Reliability Coefficients for SCES Scales: Final Version .................. 128 Items Deleted from NEG Scale ............... 129 Itesm Deleted from POS Scale ............... 130 SCES Scales: Final Version ............... 131 Pearson Product-Moment Correlations Among SCES Scales . . 133 vi Table 19. 20. 21. 22. 23. Page Source Table for Multivariate Analysis of Variance: Analysis I ........................ 135 Univariate Analyses of Variance for Sex Factor: Analysis I ........................ 136 Mean DIF Scores: Analysis I ............... 137 Cell Means for Rating of Lisa and Brian: Analysis I . . . 138 Source Table fOr Multivariate Analysis of Variance: Analysis II ....................... 153 vii LIST OF FIGURES Figure Page 1. Design ......................... 105 2a. Responses to AWS Item (39) by Experimental Groups . . . 114 2b. Responses to AWS Item (39) by Three AWS Groups ..... 116 3a-c. Mean Ratings on 4-Point SCES Scales for Nine Groups of Ss ......................... 139 4. Mean Rating on 18-Point SCES Scales for Nine Groups of S; ......................... 141 5. Responses to SCES Item (18) by Three AWS Groups . . . . 148 6. Responses to SCES Item (17) by Three AWS Groups . . . . 149 7. Responses to Item (17) in Three Experimental Con- ditions ........................ 150 8. Design for Second Analysis ............... 152 viii INTRODUCTION The formation of solidarity groups by people with common goals and class identifications is an essential step towards the attainment of social power. In American society, political powerful groups have typically been composed of white males, while women and minorities have not until recently formed potentially enduring alliances among themselves. One reason why white men have maintained a monopoly on such political alliances is that they have reserved for themselves access to resources and employment situations which facilitate the develop- ment of effective groups. A second reason is that white male groups have discouraged and even prohibited the formation of self-interested organizations among women and non-whites. These prohibitions have been enforced by techniques ranging from public ridicule to legal har- assment and physical coercion. These external factors would not, however, be sufficient to prevent a rebellion by oppressed classes, were it not for an addi- tional disrupting force arising from the Oppressed themselves. Peo- ple who have been victims of predudice often seem to internalize the prejudicial beliefs of their oppressors and act them out at each others' expense (Allport, 1954, p. 151). During the past two decades, this mechanism has been readily observable in both the black power and feminist movements. At times, conservative women and blacks have been even more strident than white males in criticizing their re- spective radical companions, and in enforcing conformity to the status quo. The focus of this dissertation is the operation of this divi- sive dynamic in women's groups, and the resulting polarization of women into those who are willing to attack male privilege in order to advance women's rights and those who support male privilege at the expense of women. The fact that a significant number of women and men are now willing to promote women's rights indicates that radical changes have already occurred. Previously, the movement for women's rights was not considered to be a serious political issue, and soli- darity among women was viewed by many people as unimportant, unnec- essary and even unattainable. Popular myth asserts that women do not like each other, that they are competitive and critical of each other, and that they are unable to work cooperatively towards common goals. The opposite behaviors are expected of men. Men are encouraged to develop "team spirit," through which they support each other and overcome external opposition. The history of women's political movements partially supports the myth, in that they have been few in number and have seldom en- dured beyond the attainment of specific, circumscribed goals. Many important female political groups have championed the cause of other oppressed people, as in the case of the Abolitionists, or the cause of moral reform, as in the case of the Christian Temperance Union. The Suffrage movement was a notable exception in that women were working for the benefit of women, but this thrust stopped with the realization of universal suffrage. In contrast, men's groups have traditionally included not only those organizations officially de- signated as such, but also the entire established government and most high status professional groups. The ruling bodies of our society have functioned as exclusive men's clubs both socially and politi- cally,and have served primarily the needs of their white male consti- tuents. Women have not only failed to form self-interested political groups, they have been in the forefront of opposition to women's rights organizations and legislation. Examples of recent organiza- tions which oppose an increase in women's legal rights are the Right- to-Life movement and the various groups working against ratification of the Equal Rights Amendment. A recent article (Wohl, 1976) re- ported that women in New York who voted against the ERA often cited opposition to feminism as a motive for rejecting the ERA. The women expressed a desire to dissociate themselves from the negative ster- eotypes of feminists popularized by the mass media. The attitude that women's liberation is a ridiculous and shameful endeavor sup- ported by neurotic, frustrated, man-hating women discourages many con- ventional women from supporting women's rights. Some female opponents of feminism have a more sophisticated grasp of ideological differences between liberationists and tradition- alists, and fear the social changes which feminists actually advocate. Anti-feminists derive their sense of importance and self-esteem from the home, and from their ability to serve the needs of their husbands and children. They fear that a broadening of women's roles will under- rnine the family structure upon which they depend. Anti-feminists 4 therefbre insist upon a determined loyalty to their mates and chil- dren,and to the male supremacist status quo. Because women have been compelled under threat of social dis- approval to support men and male dominance, the most insulting label which has been applied to feminists is "man-hater." Not only do women fear becoming feminists lest they grow to hate men or be seen as hating men, but even avowed feminists sometimes moderate their rhetoric to avoid being so labelled. This repressive attitude is in marked contrast to the typical response to anti-female sentiment, which is either overlooked because it is so normative, or is actively rein- forced and perpetuated. By the mid-1970's, the issue of whether to include men in fem- inist organizations, or whether to relate to them at all, became the most divisive and universally-debated tapic in feminist meetings. Moderates proposed that feminism should be for all people instead of being for women exclusively, that women have a responsibility to educate men, or simply that men might be hurt by being left out. Radicals argued for a separate women's space, the desirability of a more selfish stand, and claimed that the presence of men was sti- fling. The fact that moderate groups sometimes invited men to critique women's panels and participate in woman's meetings permitted informal observations of conflicts among women who responded either hostilely or protectively towards the men. Frequently when the men were crit- icized, other women in the group leaped to their defense even before the men themselves could respond. The female critics were in effect ostracized and their remarks ignored by many of the other feminists present. After an especially heated meeting, a man's defenders could sometimes be observed clustering around him, offering support and re- pudiating his critics. Chesler (1972, p. 275) seems to be describing the same phenom- enon when she speaks of women "policing" women. According to Chesler, women are socialized to idealize and support male heroes, while female heroes are starved for protection and nurturance. Women further help men by enforcing upon other women standards for appropriate feminine behavior, which includes self-sacrifice and compulsory compassion for men. Chesler (1972, p. 267) believes that the cultural ideal of the warm, self-sacrificing wife and mother is hypocritical, since only men benefit from women's compassion. She attributes the discrepancy between women's nurturant response to men and harsh response to women to the difference in men's and women's social status. Traditionally, women as well as men expect or demand another woman's help or sacrifice more quickly and easily than they demand a man's sacrifice or even his co-operation . Psychologically it represents our cultures higher evaluation of men, as well as the assigned female role of ”policing" other women in the service of male supremacy. Women mistrust and men destroy those women who are not interested in sacrificing at least something for someone for some reason. Rather than achieve at least half or all of Caesar's power, many women, including some feminists, would prefer to leave it in Caesar's hands altogether and, . sacrifice their individual advancement for the sake of less fortunate women, Third World people, one' s biological child- ren, one's weary husband, etc. (p. 277) The purpose of this dissertation is to explore more systemat- ically the ways in which sexist prejudices may influence women's views of arguments between women and men. Stereotypic perceptions of the sexes will be examined, and the role of conflict between men and women in exacerbating biases will be considered. Finally, two research studies will be presented in which women's reactions to an argument between a woman and a man were examined in a laboratory setting. COMPONENTS OF THE PROBLEM An argument between a woman and a man involves a number of different interpersonal behaviors and attitudes; the addition of a third party as a judge of the argument adds yet another dimension. First and most obviously, an argument is one type of interpersonal conflict. The particular arguments being examined here are verbal rather than physical conflicts. They are accompanied by anger on the part of one or both persons, and usually involve assertive or aggressive behavior. Arguments have some common features with other aggressive acts, and are so judged by outside observers. A person's fighting style or inhibitions against fighting in an argument are re- lated to that person's general attitude about self-assertion, aggres- sion, and self-defense. Among the major instigating factors behind most arguments are disagreements of opinion, or frustration of the felt needs of one or both parties. "Winning" an argument generally consists of having one's point recognized as valid, or of influencing some course of action, or both. "Losing" consists of admitting that one is in the wrong, of submitting to the opponents' behavioral prescriptions, or of accepting frustration. A successful verbal fighter must have a propensity for competitive behavior, together with the cluster of personality traits, values and attitudes which facilitate a compet- ‘itive stance. In this respect, arguments have features in common 8 with other situations in which one may succeed or fail in a competitive struggle for self-gratification. Successful competition requires those personality tratis de- scribed as "agentic" (Bakan, 1966), such as assertiveness, dominance and egoism. "Communal" traits, like sensitivity to people's feelings and altruism, are associated with a spirit of co-operation and com- promise rather than competition. Girls and women in this culture are socialized into communalroles and communal personality traits to the almost complete exclusion of agentic roles and traits (Donelson, 1977b). The reverse is true for boys and men. The arguments observed in women's groups occurred between a woman and a man, with secondary conflicts erupting among women. The sex of the interactants seemed to be an important factor in trigger- ing the conflicts. Attributions of who deserved to win and who was to blame may therefore have been affected by sexism, which is one type of prejudice. Knowledge of how prejudice operates in other situa- tions may help in understanding how sexism influences the judgements made about these arguments. Definitions The term prejudice has several accepted meanings; Allport (1954) includes at least six different definitions in his discussion of the concept. The most literal definition is "a judgment formed before due examination of the facts." In an interpersonal context, prejudice is a favorable or unfavorable attitude formed towards a person "prior to, or not based on, actual experience" (Allport, 1954, p. 6). Allport explains that, although prejudice may be favorable or unfavorable, the term more commonly refers to "an aversive or hostile attitude towards a person who belongs to a group, simply be- cause he belongs to that group, and is therefore presumed to have the objectionable qualities ascribed to the group" (Allport, 1954, p. 7). For the purposes of this dissertation, however, prejudice will be de- fined as either a favorable or unfavorable attitude formed towards a person because of an extraneous characteristic defining that person as a member of a particular group. ngj§m_is prejudice in which sex is the relevant variable de- fining group classification. The term sexism came into vogue in the early 1970's, and is therefore not defined in dictionaries or other literature written before the current decade. Allport (1954) uses the term "anti-feminism" to refer to prejudice based on sex; this term has acquired a more specific meaning, however, because of the renewed visibility of the feminist movement in recent years. Actually, the term "sexism" was intended to correspond to "racism," a concept with which people were already familiar by the late 1960's. The Random House Dictionary defines racism as: A belief that human races have distinctive characteris- tics that determine their respective cultures, usually involv- ing the idea that one's own race is superior and has the right to rule others. By substituting the word "sex" or "gender" for race this definition becomes an accurate description of sexism, the key element of which is the assumption that men should dominate the family, the society, and women. The belief that men have the right to rule family and society because of the superiority of the male sex is the basis of the social 10 system called patriarchy. Anthropologists generally agree that all societies are, and probably always have been, patriarchies rather than matriarchies or even egalitarian systems. That is, in every known society women are excluded from certain privileged economic or pol- itical activities, and husbands have more publicly recognized powers and prerogatives than do their wives (Rosaldo & Lamphere, 1974). The fact that patriarchy is probably a cultural universal means that the term "ruler" implies maleness both linguistically and in practice. This system implies some devaluation of women relative to men, in that women as a group are judged as being unfit to rule. Most societies, however, go beyond this level of devaluation in their attributions of an inferior and undesirable nature to women. The power to defile or destroy is often attributed to women, who are then restricted by taboos designed to counteract "bad magic" (Rosal- do, 1974; Hays, 1964). Women are also regarded as subhuman in socie- ties where maleness and humanity are equated. In our own society there are a large number of animal terms applied to women, such as "bitch," "chick," and "filly." The attribution of superior qualities to men and inferior qual- ities to women is an example of stereotypic thinking,which often functions to support prejudice. Allport (1954, p. 191) defines stereotype as "an exaggerated belief associated with a category," the function of which is to "justify conduct in relation to that category." The process of overcategorization, which is essential to stereotypic thinking, consists of generalizing certain attributions to all members of a category, like sex or race, rather than evaluat- ing people as individuals (Allport, 1954, p. 8). Stereotypes may be 11 based upon completely false assumptions, or they may be generaliza- tions based upon a kernel of truth. Groups who are victims of pre- judice may unintentionally validate stereotypes by acting out char- acteristic "victim behaviors," or by internalizing and acting out the stereotypes themselves (Allport, 1954, p. 142). Sex-role stereotyping is one aspect of sexism in which cert- ain roles and traits are selectively attributed to men and other roles and traits are attributed to women. Masculinity usually con- notes dominance and aggression both in terms of the prescribed mas- culine personalityand in terms of men's actual position in patri- archal societies. Femininityconnotesrelative passivity and helpless- ness. Interactions involving women and men may be perceived and judged in terms of these stereotypes and cultural role prescriptions. If individuals in arguments conform to cultural expectations, men may be likely to dominate women because of men's socialization into dominant roles. If individuals do not confbrm to the usual sex-role stereotypes, they may be judged negatively by other people. Observ- ers' judgments of the antagonists are likely to be related to the degree of consistency between the actual behavior of the people in the argument and the observers‘ stereotypes about the nature of women and men, as well as the extent to which the observer idealizes or rejects these stereotypes. In the arguments initially observed in women's groups, some women supported members of their own sex while other women supported men. Loyalty to a group other than one's own is often seen in vic- tims of prejudice. Identification with the dominant group is one way in which self-hate is expressed by the subordinated group 12 (Allport, 1954, p. 150). Not all people behave in this manner, however. Members of oppressed groups may adopt a militant posture towards the dominant class, and may also sympathize with other victims of prejudice (A11- port, 1954, p. 154). Even people who have not been victims them- selves may champion underdogs. Thus, there have always been indi- vidual men who have struggled to improve the lot of women. In re- cent years the feminist movement has increased public awareness of sexism to the point where women as a group are now seen by many peo- ple as being victims of prejudice. Support for women as underdogs may therefore become more prevalent. In summary, arguments between women and men involve several components of interpersonal behavior, including assertiveness, compe— tition, and agentic or communal attitudes. Judgments made about these arguments may also be influenced by a number of observer var- iables. Some of the more salient observer variables which may be relevant are prejudice for or against women, sex-role stereotypes held by the observer, attitudes toward assertive and aggressive be- havior, attitudes toward success and failure by women and men, and the observer's tendency to consistently identify with either social victims or their oppressors. LITERATURE REVIEW To date there have been no published studies directly con- cerned with peoples' reactions to ordinary arguments between women and men. 'There are an increasing number of studies, however, which have reported data which may be helpful in understanding the argu- ment situation. The most directly relevant research concerns reactions to sexist humor, in which a member of one sex is made the butt of a joke by a member of the other sex. Studying reactions to jokes about conflict between women and men is a desirable way to study sexism for several reasons. Humor is a socially acceptable way to express hostility and therefore reveals hostile attitudes without triggering inhibitions against socially undesirable forms of aggression. This is particularly true of anti-female humor, since sexist jokes are a popular means for expressing anti-female sentiment, as in mother-in- law stories and lockerroom humor. Sexist jokes provide researchers with naturally disguised material for exciting emotional reactions without simultaneously eliciting defensiveness and socially desirable responses. Studies of serious, real life conflicts between women and men have focussed primarily on physical aggression. Feminist groups have recently been compiling statistics on crimes in which women are the usual victims and men the attackers, as in rape and wife-beating. l3 14 One additional study reviewed below (Cookie, 1974), examines "jurors'" judgments of conjugal murder in a simulated trial, in which either a woman or a man is presented as the alleged assailant. A few studies using "shock machines" have also considered sex differences in phys- ical aggression directed at the other sex. There is other recent research which does not directly examine judgments about people in conflict, but which does examine attitudes about women and men in other competitive contexts. The practical need to improve women's social status and economic lot has led to an intense interest among feminist psychologists in women's achievement motivation. Included in this general area of research are studies of peoples' reactions to success and failure, studies of womens' performance in competitive achievement situations, and studies of discrimination in ratings of female and male workers and their pro- ducts. Since an argument is a competitive encounter in which out- comes may be classified as successes or failures, achievement-related data may help in understanding women's and men's behavior during arguments. Yet another area of research which focuses on evaluations of women and men is the study of sex-role stereotypes. Data in this area reveal attitudes towards men and masculinity versus women and femininity, and therefore suggest what judgments would be made when women and men are compared with one another. An additional benefit (3f research on sex-role stereotypes is that it clarifies the role of assertiveness and fighting behaviors in culturally defined mas- culinity and feminity. Studies of sex-linked stereotypes therefore icontribute to an understanding of attributions made about women and 15 men both within the argument situation and in abstract comparisons occurring outside of any situational context. The primary findings of all of the above research areas will be reviewed in the next section. Following the literature review, some theoretical fOrmulations will be presented in an attempt to offer a dynamic explanation of the initial observations in women's groups and of the relevant research findings. Hostile Humor Researchers interested in hostile humor have generally reported two kinds of findings, either an anti-female bias shown by both sexes, or prejudice in favor of the subjects' own sex. The difference in the findings may be attributable in part to the content of the jokes used; if the male and female interactants in a humorous story are in- volved in a dominance-submission relationship, sex prejudice seems to become more evident in subjects' evaluations of the joke. Priest and Wilhelm (1974) presented college students with forty jokes, twenty of which used either a man or a woman as the butt of the joke and twenty of which were not hostile in content. Male subjects preferred jokes in which women were the victims over other jokes, while female subjects preferred jokes in which men were the victims. When subjects were divided according to their scores on a self-actualization scale, however, women and men with high scores were more pro-female than were women and men with low scores. Priest and Wilhelm conclude that this finding is consistent with other data reporting an anti-authoritarian bias among self-actualizers. The cultural association between maleness and authority suggested to the 16 researchers that self-actualizers might make a point of championing women as social underdogs. Losco and Epstein (1975) designed four cartoons depicting a hostile act by a member of one sex against a victim of the other sex. These four cartoons were duplicated with the sexes of the victim and attacker reversed. Separate groups of subjects rated each set of cartoons for degree of funniness. Subjects were also asked to rate a list of adjectives for how well they applied to most men, most women, the ideal man and the ideal wman. Cartoons with male victims were rated as less funny than neutral cartoons, and cartoons with female victims were rated as more funny than neutral cartoons by subjects of both sexes. Anti-female prejudice was more marked among male subjects, especially in re- sponse to a cartoon showing one person attempting to dominate another, only to become a victim of the other's retaliation. When the man was shown first dominating and then being the butt of the joke, male sub- jects rated this cartoon as the least funny in the set, whereas women rated it as about average. When the woman was dominant, men rated the cartoon as funnier than average, while women again rated it as average. Losco and Epstein (1975) then divided the subjects into three groups, according to whether they preferred jokes in which men were the butt, jokes in which women were the butt, or neither, and examined their responses on the adjective list. Subjects of both sexes who (Jreferred female targets attributed more competence and related qual- 'ities to "most men," and more nurturance and related qualities to 'hnost women." Women who preferred male targets exaggerated this 17 pattern, which the authors attribute to a hypersensitivity to sexist injustice. Men who preferred male targets did not assign high levels of positive qualities to either sex, suggesting a general alienation from people. Cantor (1976) designed two versions of jokes in which a member of one sex "one-ups" a member of the other sex. Both men and women preferred jokes in which the man had the last word over those in which the woman had the last word, but male subjects were the more pre- judiced of the two sexes. Cantor (1976) conducted a second experi- ment which included same-sexed dyads along with male-female dyads in the stimuli, and also asked subjects to rate the dominant character on several personality dimensions. Subjects again found it funnier to have the woman be the butt of the joke than to have the man be the butt, but female subjects were more prejudiced than men against the woman in this study. A female character who insulted a male charac- ter was also rated as more critical, cruel, and domineering than was a man who insulted a woman. People who dominated members of the other sex were, however, rated as more intelligent than those who dominated members of their own sex. Chapman and Gadfield (1976) presented subjects with cartoons involving sexual innuendos, some of which were also sexist. Male subjects generally appreciated sexual humor more than did female sub- jects, but their greater appreciation was attributable to men's greater positive response to anti-female sexist humor. Women did not appreciate either anti-female or anti-male sexist humor, although they enjoyed sexual humor which was not sexist. Women and men who reported sympathizing with feminism found female target items unfunny 18 and male target items funnier than average. Zillman and Stocking (1976) played an audiotaped humor routine in which a man put down either himself, a friend of his, or an enemy. Subjects then evaluated the disparager's personality. Both women and men rated the self-disparager as less intelligent and secure than the person who disparaged others. Women f0und it funnier, however, for the man to disparage himself than to disparage others, whereas men f6und it funnier for him to disparage others. In a second experiment, Zillman and Stocking (1976) added mat- erials in which women were disparagers and materials in which women were disparaged. Men enjoyed hearing disparagement of a woman by others, but they disliked self-disparaging women even more than they disliked self—disparaging men. Women enjoyed hearing people of either sex disparage themselves and rated self-disparagers as "appeal- ing." Although degree of appreciation of hostile humor has generally been interpreted as a measure of hostility towards the person who is the butt of the joke, one study indicates that researchers should be cautious in equating these variables. Grote and Cvetkovich (1972) found that women who listened to an anti-female humor monologue rated it as funnier than did women who listened to a neutral humor monologue by the same comedian. After listening to the taped monologues, sub- jects were given an opportunity to present any number of arguments they could think of in favor of a woman's right to choose a non- traditional vocation. Women who had listened to the anti-female routine presented more arguments than any of the other groups of subjects, including women who had listened to excerpts from 19 Sisterhood is Powerful. The researchers concluded that women in the anti-female group were actually angered by the monologue, despite the high funniness rating they assigned to it. Women may inhibit angry, aggressive responses to insults and attempt to pass them off as jokes. This interpretation is consistent with observation of female psychotherapy clients, who often laugh when discussing incidents which angered them. The adaptive value of this behavior is obvious, in that a victim who laughs at herself is more likely to diffuse an attackers' hostility than is a victim who counterattacks. Allport (1954, p. 147) includes self-directed humor as one of the behaviors frequently practiced by minorities and social victims. Women's adoption of the submissive,self—ridiculing posture is strongly reinforced by the frequently expressed criticism that "women's libbers have no sense of humor." In contrast, men may more likely to react to ridicule of themselves with anger. Men's hostile humor towards outside groups may be an expression of dominance, which helps to explain why men favor humor which disparages others and women favor humor which disparages themselves. Dominance over others is viewed as appropriate behavior for men, but inappropriate behavior for women. Future research in this area should take into account the fact that, while women and men both rate a sexist joke as funny, they may do so for different reasons. In summary, findings of studies on humorous conflicts between the sexes are mixed with respect to the direction Lisa's POS score 2. For trad gs: Lisa's NEG score > Brian's NEG score 3. For lib Ss: Lisa's POS score > Brian's POS score, but the difference will be less than in 1. 4. For lib Ss: Brian's NEG score > Lisa's NEG score, but the difference will be less than in 2. The pattern described above is predicted to occur in Conditions 108 A and 8, thus yielding the significant two-way interaction. A three- way interaction effect is also predicted, because of the varying magnitude of differences between ratings made by liberal and tradi- tional subjects within Conditions A and B, and because Condition C is expected to eliminate differences altogether for liberal subjects. II. Argument Outcome x AWS x Sex Interaction A. Condition A (outcome not specified) The pattern for the two-way interaction will be found, but lib- eral subjects will be less influenced than traditional subjects by Sex. Thus: 1. For trad Ss: Brian's POS score > Lisa's POS score 2. For trad S5: Lisa's NEG score > Brian's NEG score 3. For lib S5: Lisa's POS score > Brian's POS score, but the difference will be less than in l. 4. For lib gs: Brian's NEG score > Lisa's NEG score, but the difference will be smaller than in 2. 8. Condition 8 (negative outcome) Information that the outcome is negative is predicted to in- crease the sex differences found with the outcome unspecified. Thus: 1.-4. The same pattern will appear as in A. l.-4. above. The DIF scores will be greater in 8 than in A. C. Condition C (positive outcome) Information that the outcome is positive is expected to amel- iorate the effect of Sex of person rated relative to Conditions A and B, and in fact to eliminate sex differences in amount of criticism of the story characters. Thus: D. dicted to 1. 109 For trad S5: Brian's POS score > Lisa's POS score. The DIF will be smallest in C relative to trads in A and B. For trad §s: Neither character will be evaluated negatively. There will be no difference between Lisa and Brian. For lib Ss: Lisa's POS score > Brian's POS score. The DIF will be smallest in C, relative to libs in A and B. For lib S5: Neither character will be evaluated negatively. No difference between Lisa's and Brian's scores is pre- dicted. As the above predictions imply, the groups of subjects pre- show the most favoritism toward one story character are: Liberals in Condition A will register the largest DIF score favoring the woman. Traditionals in Condition 8 will register the largest DIF score favoring the man. 111. Secondary Hypotheses For theoretical reasons discussed above, certain of the SCES subscales are expected to reflect sex—linked prejudice more strongly than others. Predictions are made only for the subscales which seem theoretically most salient. 1. 2. DOMP a. For trad §_s: BDOMP > LDOMP b. For lib §_S: LDOMP > BDOMP DOMN a. For trad Ss: LDOMN > BDOMN b. For lib SS: BDOMN > LDOMN 110 3. LOVP a. For trad Ss: BLOVP > LLOVP b. For lib Ss: LLOVP > BLOVP 4. LLOVN a. For trad Ss: LLOVN > BLOVN b. For lib Ss: LLOVN = BLOVN, and neither character will be rated negatively on this scale. 5. DEPP a. For trad Ss: BDEPP > LDEPP b. For lib Ss: LDEPP > BDEPP 6. DEPN a. For trad Ss: LDEPN > BDEPN b. For lib Ss: BDEPN > LDEPN The above set of predictions concerning the SCES subscales bas- ically mean that traditional women will perceive the man as a good person who is vulnerable to the hostile attacks of a destructively powerful women. Liberal subjects are expected to criticize the man for displays of weakness, but not to see him as a genuine victim. Liberal subjects are predicted to perceive the woman in a favorable light and the man less favorably. Liberal women are expected to be more permissive than traditionals with respect to arguments, so lib- eral subjects are also predicted to be less critical than traditional subjects toward either story character for expressing hostility. RESULTS Description of the Subjects Demographic Data Two hundred and one undergraduate women participated as subjects in the study. The majority of the women (77 percent) were first term students at MSU. There was a smaller number of sophomores (12 percent), juniors (5 percent) and seniors (3 percent). Most of the subjects were seventeen to nineteen years old, although the age range extended up to twenty-six years. The women in the sample described their race as either "black" or "white" on the demographic questionnaire. The black subsample included twenty women, who were distributed fairly evenly by chance across experimental conditions (N = 5 in Condition A, N = 8 in Con- dition B, and N = 7 in Condition C). Most subjects (75 percent) were from a Christian religious background, with approximately equal numbers of women identifying themselfes as Protestants or Roman Catholics. Eighteen subjects were from a Jewish background, and three subjects were from various other religions. Nineteen women described themselves as agnostics, atheists, or as having no religious beliefs. The sample was in summary fairly homogeneous demographically. The women were mostly recent high school graduates who had been attending MSU for one or two months. They were hence expected to be 111 112 fairly naive about feminism and other liberal values commonly pro- fessed on the university campus. Attitudes Toward Women's Roles The distribution of AWS scores for the present sample is very similar to the distribution described by Dunbar (1975) for MSU women tested several years ago. The statistics for the two distri- butions are shown in Table 9. Although attitudes toward women may have changed within the university community over the past several years, the beliefs of incoming students have remained remarkably stable. The women in both MSU samples registered AWS scores which were more liberal than those reported by Spence and Helmreich (1972) for students at the University of Texas, but which were more conser- vative than those reported by Lunneborg (1974) for students at the University of Washington. The fact that the differences in regional norms are in the expected directions provides further evidence for the constuct validity of the AWS. Table 9. Distributions of AWS Scores Present Study Dunbar (1975) Range 62-159 pts 62-160 pts. Mean 115.71 115.63 50 18.08 20.50 Reliability .90 .92 113 Subject classification by AWS. In previous research using the AWS, subjects were rank-ordered using the AWS total score, then were divided into pro-feminist liberals, moderates and traditionals. In the present study a scoring variation was introduced because the con- tent of one of the AWS items (39) is too closely related to the de- pendent variables for the item to be used in partitioning the sample. Item (39) states that "A wife should make every effort to mini- mize irritation and inconvenience to the male head of the family." Because the AWS was administered after the SCES, subjects' responses to item (39) may have been influenced by their previous experiences in the experiment. Figure 2a shows the distribution of responses to item (39) for subjects in Conditions A, B, and C. Although agreement with the item is elevated slightly in Condition C, the three distri- butions are very similar. AWS subtotal scores were calculated by subtracting subjects' scores for item (39) from their AWS total scores. The subtotal scores were then rank-ordered, and the sample was partitioned into thirds. Table 10 shows the range of subtotal scores for the three resulting groups, who were designated liberals, moderates and tra- ditionals. Table 10. Range of AWS Subtotal Scores AWS Class N Scores Range Liberal 66 124-155 31 pts. Moderate 66 108-123 15 pts. Traditional 65 62-107 45 pts. 114 23 2: av 83 e mazoea Pepcmswemaxm an Ammv scum m3< op mmmcoqmmm m N F 23 2: av 83 a m N P o eeab_eeeo m eewe_eeeu .mN acumen A