Love and anger in romantic relationships: a discrete systems model.

In a study of 124 dating couples, we tested a discrete systems model of the functions of two emotion systems in romantic relationships: love and anger/upset. This model posits that the operation of these systems reflects adaptations shaped by natural selection to solve different adaptive problems. Accordingly, we hypothesized that the love and anger/upset emotion systems would be largely independent in the classes of information they track in romantic relationships, in the psychological mechanisms that process that information, and in the resultant behavior generated. Consistent with the discrete systems model, and in contrast to a competing "crossover" model, differences across relationships in feelings of love covaried with differences in strategic facilitation but not in strategic interference by partners. Similarly, differences in feelings of anger/upset during conflict covaried with differences in strategic interference but not strategic facilitation. In turn, feelings of love predicted commitment-promoting behavior but not partner-directed aggression, whereas levels of anger/upset predicted aggression but not commitment. As also predicted by our model, the love and anger/upset emotion systems converged to predict relationship satisfaction.

Pornography and sexual aggression: are there reliable effects and can we understand them?

In response to some recent critiques, we (a) analyze the arguments and data presented in those commentaries, (b) integrate the findings of several metaanalytic summaries of experimental and naturalistic research, and (c) conduct statistical analyses on a large representative sample. All three steps support the existence of reliable associations between frequent pornography use and sexually aggressive behaviors, particularly for violent pornography and/or for men at high risk for sexual aggression. We suggest that the way relatively aggressive men interpret and react to the same pornography may differ from that of nonaggressive men, a perspective that helps integrate the current analyses with studies comparing rapists and nonrapists as well as with cross-cultural research.

Characteristics of men who aggress sexually and of men who imagine aggressing: risk and moderating variables.

The authors showed that the extent to which men's personalities were self-centered rather than sensitive to others' needs moderated the connection between risk factors and sexually aggressive behavior. Men who were at risk for committing aggression but who were also sensitive to others' feelings aggressed less than the corresponding group, who had relatively self-centered personalities. However, both groups showed high levels of imagined sexual aggression. The authors suggest that imagined sexual aggression may reveal information about the presence of underlying risk factors even when actual aggression is inhibited by personality characteristics such as those studied here. The implications for therapeutic interventions of the finding of aggression attenuation are also discussed.

The longitudinal impact of demand and withdrawal during marital conflict.

Forty-eight couples completed a measure of relationship satisfaction and participated in 2 video-taped problem-solving interactions, 1 focused on an issue identified by the woman and 1 focused on an issue identified by the man. Thirty-six men and 36 women completed the satisfaction measure again 2.5 years later. Demandingness, and to a lesser extent withdrawal, during the interactions showed many significant associations with both Time 1 and Time 2 satisfaction. The relationship of demandingness and withdrawal to change in satisfaction was also examined using both change scores and partial correlations. Withdrawal by men and woman demand-man withdraw during discussions of issues identified by the women reliably predicted change (decline) in wives' relationship satisfaction.

Using the confluence model of sexual aggression to predict men's conflict with women: a 10-year follow-up study.

We tested a model describing the characteristics of sexually aggressive men that may also be useful for understanding the causes of other antisocial acts against women. This model hypothesizes that sexual aggressors can be identified by two sets of characteristics, labeled hostile masculinity and impersonal sex. To test this model, we followed up a sample of men 10 years after first studying them when they were young adults. We sought to predict which men would be in distressed relationships with women, be aggressive sexually, be nonsexually aggressive, or some combination of these. These behaviors were measured not only by questioning the men themselves but also by questioning many of the men's female partners. Some couples' videotaped conversations were also analyzed. The data supported the ability of the model to predict behavior 10 years later. We also developed the model further and identified the common and unique characteristics contributing to sexual aggression as compared with the other conflictual behaviors studied. The data supported the usefulness of hierarchical modeling incorporating both general factors that contribute to various interpersonal conflicts as well as specific factors uniquely pertaining to dominance of women.

Sexually aggressive men's perceptions of women's communications: testing three explanations.

The authors tested three explanations of findings that sexually aggressive men perceive women's communications differently than less aggressive men. The first suggests that aggressors are incompetent in decoding women's negative emotions. The second posits that they fail to make subtle distinctions between women's friendliness and seductiveness and between assertiveness and hostility. The third explanation contends that sexual aggressors use a suspicious schema and therefore discount the veridicality of women's communications. These explanations were tested using videotaped scenarios in which a woman's responses to a man's advances were systematically varied. The data were most supportive of the suspicious schema explanation. These findings are integrated with other research on the characteristics of sexual aggressors and on the perceptions of aggressive children and of maritally violent men. Implications for interventions are also discussed.

Characteristics of aggressors against women: testing a model using a national sample of college students.

Structural equation modeling was used to study the characteristics of college men (N = 2,652) who aggressed against women either sexually, nonsexually, or both. According to the model, hostile childhood experiences affect involvement in delinquency, leading to aggression through two paths: (a) hostile attitudes and personality, which result in coerciveness both in sexual and nonsexual interactions, and (b) sexual promiscuity, which, especially in interaction with hostility, produces sexual aggression. In addition, sexual and nonsexual coercion were hypothesized to share a common underlying factor. Although its development was guided by integrating previous theory and research, the initial model was refined in half of the sample and later replicated in the second half. Overall, it fitted the data very well in both halves and in a separate replication with a sample for whom data were available about sexual but not about nonsexual aggression.

Predictors of naturalistic sexual aggression.

This research integrated within a theoretical and empirical framework varied predictor factors pertaining to males' sexual aggression against women. The selection of predictors was guided by theorizing that sexual aggression is caused by the interaction among multiple factors, including those creating the motivation for the act, those reducing internal and external inhibitions, and those providing the opportunity for the act to occur. The predictor factors assessed were sexual arousal in response to aggression, dominance as a motive for sexual acts, hostility toward women, attitudes accepting of violence against women, psychoticism, and sexual experience. A measure assessing self-reported sexual aggression (primarily among acquaintances) in naturalistic settings served as the dependent measure. The subjects were 155 males. As expected, nearly all the predictor factors significantly related to sexual aggression. In addition, much better prediction of such aggression was achieved by a combination of these factors than by any one individually. It was also found that including interactions among these predictors yielded a regression equation that was more successful in relating to sexual aggression than an equation using an additive combination only. The relevance of these data to the causes and prediction of violence against women is discussed.

Sexual arousal in response to aggression: ideological, aggressive, and sexual correlates.

In two experiments we assessed the impact of aggression on nonrapists' sexual arousal. In the first, both male subjects (n = 37) and female subjects (n = 43) reported more sexual arousal in response to nonaggressive than to aggressive depictions when the portrayals were sexually explicit, but the opposite occurred when the portrayals were nonsexual. Only male subjects (N = 359) participated in the second experiment. On the basis of their self-reported sexual arousal to the use of force, they were classified into the no arousal, moderate arousal or high arousal from force groups. To evaluate the veridicality of this classification, we assessed some subjects' (n = 118) penile tumescence in response to various depictions. The findings generally replicated those of the first experiment and confirmed the accuracy of the arousal-from-force classification. The no arousal and the moderate arousal from force subjects were less sexually aroused by aggressive than by nonaggressive portrayals, but the opposite was found for the high arousal from force group. Using the entire sample (N = 359), we also assessed differences on various factors among these three groups. We found strong differences on ideological factors, including acceptance of violence against and dominance over women. We also found differences in acceptance of nonsexual aggression and in subjects' beliefs that they might actually use force against women. In contrast, differences were not found on sexuality factors. The data's implications for theories on the causes of rape are discussed.

Factors associated with rape as predictors of laboratory aggression against women.

This study examined the relation between factors associated with "real"-world aggression against women and laboratory aggression. In the first phase of the research, assessment was made of the attitudes about aggression against women and of the sexual responsiveness to rape of 42 male subjects. In the second phase of the research, which subjects believed was a totally independent experiment, aggression was assessed within a Buss paradigm. It was found that the factors assessed in the first phase successfully predicted men's laboratory aggression in the second research phase. The data are interpreted as supporting the construct validity of (a) theory that suggests that common factors underlie varied acts of aggression against women, (b) the measures designed to predict aggressive tendencies, and (c) the methodology of assessing aggression within a laboratory context.

Rape fantasies as a function of exposure to violent sexual stimuli.

Twenty-nine male students, classified on the basis of questionnaire responses as sexually force oriented or non-force oriented, were randomly assigned to exposure to rape or mutually-consenting versions of a slide-audio show. All subjects were then exposed to the same audio description of a rape read by a female. They were later asked to create their own fantasies. Penile tumescence and self-reports of arousal indicated that relatively high levels of sexual arousal were generated by all of the experimental stimuli. No differences in arousal during the exposure phase were found as a function of the manipulation in the content of the slide-audio show. Sexual arousal during the fantasy period, assessed by means of self-reports, indicated that those who had been classified as force oriented created more arousing fantasies after having been exposed to the rape version of the show, whereas those classified as non-force oriented created more arousing fantasies following the mutually-consenting version. Most significantly, those exposed to the rape version, irrespective of their sexual classification, created more violent sexual fantasies than those exposed to the mutually-consenting version. The possible role of such media-elicited fantasies in the development of antisocial attitudes and behavior is considered. Also discussed are ethical issues arising from implementing or failing to implement research in this area.

Sexual responsiveness of college students to rape depictions: inhibitory and disinhibitory effects.

Two experiments were conducted to identify the specific dimensions in portrayals of sexual violence that inhibit or disinhibit the sexual responsiveness of male and female college students. The first experiment replicated earlier findings that normals are less sexually aroused by portrayals of sexual assault than by depictions of mutually consenting sex. In the second experiment, it was shown that portraying the rape victim as experiencing an involuntary orgasm disinhibited subjects' sexual responsiveness and resulted in levels of arousal comparable to those elicited by depictions of mutually consenting sex. Surprisingly, however, it was found that although female subjects were most aroused when the rape victim was portrayed as experiencing an orgasm and no pain, males were most aroused when the victim experienced an orgasm and pain. The relevance of these data to pornography and to the common belief among rapists that their victims derive pleasure from being assaulted is discussed. Misattribution, identification, and power explanations of the findings are also discussed. Finally, it is suggested that arousing stimuli that fuse sexuality and violence may have antisocial effects.

Ethical issues and exposure to rape stimuli: a reply to Sherif.

The three ethical issues raised by Sherif are discussed. In response to the first issue, we recognize the need to measure the effectiveness of our attempts to counteract possible undesirable effects of the research materials. Problems that arise in implementing such an assessment are considered. Also discussed is the possibility that exposure to such materials combined with an effective discrediting communication may have beneficial effects. In addressing the second issue, it is argued that Sherif's criticisms of our alternative data interpretations are unfounded. Finally, Sherif's suggestion to formalize the process of having investigators consider the social policy implications of their research is whole-heartedly supported.