Predicting responses to sexually aggressive stories: the role of consent, interest in sexual aggression, and overall sexual interest.

Explanatory models of sexual aggression have been proposed from a variety of different disciplines within the social sciences and from diverse theoretical perspectives within psychology (Hudson & Ward, 1997); however, a view discussed by many theorists is that sexual aggression is better understood as an act of violence and aggression rather than a sexual act (Brownmiller, 1975; Malamuth, Heavey, & Linz, 1993; Palmer, 1988). Although initially important, both scientifically and politically, the emphasis on aggression and violence has led many researchers to neglect the sexual response that is an integral component of sexually aggressive behaviors. Recent models predicting sexually aggressive behaviors have often focused on the usefulness of variables measuring hostility and aggression toward women, but have only occasionally included variables measuring men's interests and attitudes about sexual behavior in general (e.g., Malamuth et al., 1993). In fact, no study has ever directly compared and contrasted these two types of predictor variables. It is our hypothesis that prediction models of sexually aggressive behaviors could be improved by including general sexual interest variables with the more typical measures of hostile and aggressive attitudes toward women and male-female relationships.

The recent literature examining the relationship between sexually aggressive behaviors and men's interests and attitudes about sexual behavior in general (e.g., Byers & Eno, 1991) suggests that the following variables warrant further consideration: erotophobia, sexual inhibition, and sexual excitation. For example, Byers and Eno (1991) examined the association between the tendency to perceive oneself as easily sexually aroused (sexual excitation) and sexual aggression. Their data indicated that men who characterized themselves as high on sexual excitation were more likely to report having used physical force to obtain sex.

There is also some evidence that the tendency to respond to sexual cues with negative affect (erotophobia) has a relationship to sexual aggression (Byers & Eno, 1991; Caron & Carter, 1997; Hurlbert & Apt, 1991). These studies show small but statistically significant effects suggesting that men who tend to be erotophobic are more tolerant of rape (Caron & Carter, 1997), are more likely to physically abuse their wives (Hurlbert & Apt, 1991), and are more likely to use physical force to obtain sex (Byers & Eno, 1991). These findings appear discordant with the findings described above concerning sexual excitation. It is possible, however, that these findings in the literature reflect relationships with sexual aggression in different subsets of men. Specifically, it is possible that being easily sexually aroused may be related to sexually aggressive behaviors in some men, whereas in other men, being erotophobic and therefore more sexually conflicted may lead to the same behaviors. Given the state of the literature, this is a highly speculative interpretation of these findings, especially given that the effects associated with erotophobia tend to be small.

Other research has suggested that high levels of sexual inhibition or conflict are associated with rape-supportive attitudes but not with actual sexual aggression (Brier & Malamuth, 1983; Briere, Malamuth, & Check, 1983; Malamuth, Check, & Briere, 1986). In these studies, however, operational definitions of "sexual inhibition or conflict" were not provided. In contrast, a recent study using a clearer definition of sexual inhibition (i.e., concerns about the potential negative consequences of sexual activity) found a relationship between it and responses to sexually coercive stimuli (Janssen, Vorst, Finn, & Bancroft, 2002b). In this study, participants scoring low on a measure of sexual inhibition showed significantly more physiological sexual arousal to films depicting coercive sexual interactions than those scoring high on the sexual inhibition measure. Although sexual arousal to coercive sexual stimuli is not synonymous with engaging in sexually aggressive behavior, at least one empirical study found it to be one of the strongest single predictors of self-reported history of engaging in sexually coercive behavior (Malamuth, 1986).

In addition to the overall sexual interest variables discussed above, research has examined variables associated with the hypothesis that sexual aggression is best conceptualized as an act of aggression, the expression of which happens to be sexual (e.g., Burt, 1980; Check, 1985; Malamuth, Heavey, & Linz, 1993). One such variable is attraction to sexual aggression which refers "to the belief that aggressing sexually is likely to be a sexually arousing experience, both to aggressors and victims, so that the respondent believes that he might aggress if were it not for fear of punishment or other inhibitory factors" (Malamuth, 1989a, p. 30). Thus, this construct, unlike the overall sexual interest variables discussed above, refers to a specific sexual interest in coercive sexual behaviors. Research indicates attraction to sexual aggression is associated with attitudes supporting violence against women, hostility toward women, physiological and self-report indices of sexual arousal to sexually aggressive stimuli, and a self-reported history of sexual aggression (Malamuth, 1989a; Calhoun, Bernat, Clum, & Frame, 1997). In addition, men attracted to sexual aggression tended to perceive a rape victim's experience as more positive than those not attracted to sexual aggression (Malamuth, 1989b).

The preceding discussion supports the hypothesis that both self-reported interest in sexual aggression and overall sexual interest variables are associated with sexually aggressive behavior and sexual arousal in response to sexually coercive stimuli. Unfortunately, it is not clear from the research the extent to which each of these types of variables would predict responses to sexually aggressive stimuli if both were to be examined in the same prediction model. In other words, it isn't known whether these two types of predictors are unique and would complement each other if used together or if they share predictive variance and would be redundant if used as predictors in the same model.

A second issue that arises when reviewing this literature involves the manner in which sexual aggression is typically conceptualized and subsequently operationally defined. Operational definitions of sexually aggressive behavior almost always include aggressive and violent behavior by a man directed toward a woman to force sexual activity despite her resistance and/or lack of consent. Study stimuli reflecting this conceptualization of sexually aggressive behavior are then typically compared to stimuli emphasizing the consensual nature of sexual activity that does not include any forceful, rough, or aggressive elements. The use of such definitions, though common, appears to confound the issue of consent with presence of forceful or aggressive behaviors. Thus, studies using such definitions may confuse the prediction of an interest in rough or forceful but consensual activities (e.g., sadomasochistic sex) with that of an interest in forcing an unwilling person to engage in sexual activity (e.g., rape). This is an important issue because the intent of the studies in this area is to predict a proclivity to rape, not a proclivity to engage in sadomasochistic or "rough" sex.

In our study, sexual aggression was conceptualized as any nonconsensual

sexual behavior. In this conceptualization the degree of force or violence associated with the behavior is irrelevant. Therefore, consistent with our conceptualization and to remedy what we saw as a confound in previous studies, we had participants respond to two stories that both contained equal amounts of forceful sexual contact (e.g., shoving, ripping clothes, holding a woman down). These stories differed only with respect to whether the woman expressed interest, arousal, and a willingness to participate in the sexual activity or expressed her aversion by repeatedly resisting, refusing, and trying to escape from the activity (the specifics of the scenarios can be found in the methods section). Moreover, in an effort to emphasize equally both the sexual and forceful components of sexual aggression, and because the participants were male, the stories were highly explicit sexually and tended to be more stereotypically male-oriented with regard to erotic content (i.e., in addition to nonconsensual and forceful elements, emphasis was placed on detailed and explicit descriptions of sexual behaviors). Last, our study addressed the issue of whether overall sexual interest variables and interest in sexual aggression might combine in a complimentary fashion to improve our prediction of a man's proclivity toward sexually aggressive behaviors.

We hypothesized that overall sexual interest variables (sexual excitation, sexual inhibition, and erotophobia) would predict self-report responses to both the consensual and nonconsensual stories (though more to the consensual story), and interest in sexual aggression would predict responses only to the nonconsensual story. Our specific hypotheses were as follows:

1. Overall sexual interest variables (high levels of erotophobia and sexual excitation and low levels of sexual inhibition) would be associated with a greater interest in (i.e., greater self-reported sexual arousal and desire) and a more "positive" response to (i.e., greater likelihood of acting like the man in the story and greater willingness to read similar stories) both the consensual and nonconsensual stories. It was expected, however, that these relationships would be stronger for the consensual story.

2. Overall sexual interest variables (high levels of erotophobia and sexual excitation and low levels of sexual inhibition) would predict more positive perceptions of the woman's experience in both stories (i.e., higher ratings of the woman's level of enjoyment and sexual desire), though again this effect would be stronger for the consensual story.

3. Interest in sexual aggression (high levels of attraction to sexual aggression) would be associated with a greater interest in (i.e., greater self-reported sexual arousal and desire) and a more "positive" response to (i.e., greater likelihood of acting like the man in the story and greater willingness to read similar materials) the nonconsensual story.

4. Interest in sexual aggression (high levels of attraction to sexual aggression) would predict more positive perceptions of the woman's experience in the nonconsensual story (i.e., higher ratings of the woman's level of enjoyment and sexual desire).

METHOD

Participants

One-hundred eight college males participated in the study. The average age of the participants was 19.95 years (SD = 3.07), and most had never been married and were not living with a partner (98%). Ninety-seven percent of participants were non-Hispanic White with the remaining 3% falling evenly into each of the following categories: Native American, African American, and Asian American. Furthermore, most of the participants were first- (50.9%) and second-year (34.3%) college undergraduates. The remaining were evenly divided between third- (7.4%) and fourth-year (7.4%) undergraduates.

Materials

Attraction to Sexual Aggression Scale. The Attraction to Sexual Aggression Scale is a 14-item questionnaire that evaluates a respondent's reactions to two types of behavior: forcing a female to do something sexual she did not want to do and rape. The respondent evaluates the extent to which he finds these behaviors attractive, whether or not he has thought about such behaviors, his likelihood of engaging in each of these behaviors, the degree to which he would be sexually aroused by these behaviors, and the extent to which he believes other men and women would be aroused by these behaviors (Malamuth, 1989a). Responses to the items are made on an l 1-point scale, ranging from 0 to 10 with higher scores indicating a greater level of the construct measured in the question. The questions are embedded within an assessment of several other sexual behaviors such as bondage, pedophilia, transvestism, and heterosexual intercourse. An overall score is calculated by summing across individual items with higher scores indicating a greater level of attraction to sexual aggression. The scale has been shown to have adequate test-retest reliability at a one-week interval (r = .75). In the present sample, Cronbach's alpha was .86. Research by Malamuth (1989a, 1989b) supports the validity of the instrument.

The Sexual Inhibition and Sexual Excitation Scales. The Sexual Inhibition and Sexual Excitation Scales consist of 45 items that measure men's tendencies to respond with sexual inhibition and excitation to a variety of stimuli (Janssen, Vorst, Finn, & Bancroft, 2002a). The instrument contains one Excitation Scale and two inhibition scales: Threat of Performance Failure and Threat of Performance Consequences. The Excitation Scale contains items such as "When a sexually attractive stranger looks me straight in the eye, I become aroused," and "When I look at erotic pictures, I easily become sexually aroused." The Threat of Performance Failure Scale contains items such as "I need my penis to be touched to maintain an erection," and "If I am concerned about pleasing my partner sexually, I easily lose my erection." Finally, the Threat of Performance Consequences Scale contains items such as "If I can be heard by others while having sex, I am unlikely to stay sexually aroused," and "If my new sexual partner does not want to use a condom, I am unlikely to stay aroused."

Responses to items are made on a Likert-type scale (ranging from 1 = strongly agree to 4 = strongly disagree). An overall score for each scale is calculated by summing across items for that scale. Higher scores on each scale indicate a greater degree of the construct measured. The test-retest reliability has been shown to be adequate for the Excitation Scale (r = .76), the Threat of Performance Failure Scale (r = .67) and the Threat of Performance Consequences Scale (r = .74). In the current study, Cronbach's alpha for each scale was as follows: .89 (Excitation), .76 (Threat of Performance Failure), and .69 (Threat of Performance Consequences). Research by Janssen et al. (2002a, 2002b) supports the validity of the instrument.

The Sexual Opinion Survey. The Sexual Opinion Survey is a 21-item questionnaire that assesses the personality dimension of erotophobia-erotophilia or the "the disposition to respond to sexual cues along a negative-positive dimension of affect and evaluation" (Fisher, Byrne, White, & Kelley, 1988, p. 123). Each item describes a sexual situation and provides a negative or positive affective response to the situation. Example items include "Masturbation can be an exciting experience," and "Seeing a pornographic movie would be sexually arousing to me." Participants then rate the extent to which they agree or disagree with the affective response on a 7-point scale (ranging from 1 = strongly agree to 7 = strongly disagree). For this study, we obtained a total score by following the instructions outlined by previous research. The test-retest reliability of the instrument over 2 months ranges from .80 to .85 (Fisher et al., 1988). In the present study, Cronbach's alpha was .81. Research has also supported the validity of the instrument (Fisher et al., 1988).

Personal Reaction Inventory. The Personal Reaction Inventory, also know as the Marlowe-Crowne Social Desirability Scale, is a 33-item questionnaire assessing an individual's need to obtain approval by responding in a socially appropriate and acceptable manner (Crowne & Marlowe, 1960); it was included in the present study to control for social desirability. The instrument contains items such as "No matter who I'm talking to, I'm always a good listener," and "I have never intensely disliked anyone." Responses to items are made in a true/false format, and an overall score is computed by summing across all items, with higher scores indicating greater social desirability. The scale has been shown to have adequate test-retest reliability (r = .89), and it demonstrated acceptable internal consistency in the present study (Cronbach's alpha = .77). Finally, research has also supported the validity of the instrument (Crowne & Marlow, 1960).

Audio scripts. The two senior authors of this study constructed three stories using a face valid approach with the goal of portraying, as realistically as possible, acquaintance rape, consensual but forceful (i.e., rough) sexual activity, and consensual, nonforceful sexual activity. The scripts were repeatedly revised until a consensus was reached among the authors that this goal had been achieved. All three scripts were 4-minute narratives recorded on audiotape by the same male voice.

The consensual story containing no forceful behaviors was constructed to serve only as a filler story. Its purpose was simply to mask the similarities between the other two audio scripts and participants' responses to it were not included in data analyses.

The two scripts of interest in this study (nonconsensual and consensual but forceful sexual activity) both involved the ending of a date between a male and female undergraduate who had dated a few times and had sex once before.

The consensual but forceful (i.e., rough sex) script described a sexual episode in which the man behaved in a rough, forceful, and even aggressive manner by shoving and restraining the woman, slapping her, pulling her hair, and ripping her clothes. Although forceful and even aggressive, this script went to great pains to clearly depict a sexual encounter in which the woman not only consented to the activity but clearly enjoyed it and was aroused by it.

The nonconsensual, forceful (i.e., acquaintance rape) audio script included the same forceful and even aggressive behaviors, but was clearly nonconsensual. In other words, the woman did not enjoy nor was she sexually aroused by the activity, and, most importantly, she did not consent to it (repeatedly saying no, telling the man to stop, and trying to resist and escape).

Finally, the occurrence of the forceful and aggressive behaviors (e.g., slapping, restraining) occurred with the same frequency, in the same order, and approximately at the same time in each of the two experimental audio scripts. The occurrence of each forceful or aggressive behavior was then yoked to an appropriate response by the woman. In the consensual but forceful audio script the woman's responses communicated her desire for the behavior and willingness to engage in it, while in the nonconsensual audio script the same behavior by the man was followed by verbal and physical resistance, displeasure, and disapproval of his behavior.

Sexual arousal questionnaire. Participants were asked to provide several ratings in response to each of the independent-variable audio scripts. Items were originally constructed by the two senior authors using a face validity approach and were repeatedly revised until consensus was reached among the four authors. Responses were made on a 7-point Likert-type scale. Specifically, participants were asked how sexually aroused they felt after each story (ranging from 1 = not at all sexually aroused to 7 = very sexually aroused), the extent to which the story made them feel like having sex (ranging from 1 = not at all to 7 = very much), their likelihood of acting as the man in the story did if they could be assured that no one would know (ranging from 1 = not at all likely to 7 = very likely), and the extent to which they would like to read a similar story if they could be assured that no one would know (ranging from 1 = not at all likely to 7 = very likely). Participants also responded to items assessing their perceptions of the extent to which the woman enjoyed what happened (ranging from 1 = not at all to 7 = very much) and wanted to have sex (ranging from 1 = not at all to 7 = very much). Eight questions (four per story) regarding the content of each story were asked as manipulation checks to insure that participants listened to the audio scripts.

Procedure

Six hundred male undergraduates completed the Attraction to Sexual Aggression Scale in groups of 25 to 30 as part of a larger study. Manila folders with large flaps on the sides were provided to participants so that they could shield their responses from others sitting nearby. Participants provided their phone numbers on consent forms to allow the researchers to contact them if they wished to participate in the present study and other ongoing studies. Names appeared only on the consent forms, which participants turned in to the researchers before completing the questionnaires. The questionnaires and consent forms were then stored separately. An assigned identification number appeared on both the questionnaires and consent forms as a way to match participants' responses to the data collected in the present study.

One-hundred eight participants were later randomly selected from the original participant pool of 600 and contacted by phone to participate in the present study. They were informed that they would be listening to some erotic stories and filling out some questionnaires. If a participant agreed to participate, he was scheduled for an appointment.

Upon arriving at the laboratory, the participant was given a consent form explaining the nature of the study and its confidentiality. The consent form explained that the purpose of the story was to determine people's reactions to sexually explicit stories. After they signed the consent form, they were asked to listen to three erotic stories on audiotape. Specifically, they were instructed to listen, from beginning to end, to stories about some couples who had been out on a few dates, and to answer some questions about the stories after hearing each one.

All participants heard the filler story in the middle of the two experimental stories; however, the position each experimental story occupied was counterbalanced across subjects so one half heard one story in the first position and one half heard the other story in the first position. A researcher played the audiotapes from a control room and the participants listened to them through headphones in a private laboratory room. Following each audiotape the participants completed the sexual arousal questionnaire.

After listening to all three stories, participants filled out the Sexual Opinion Survey, the Personal Reaction Inventory, and the Sexual Inhibition and Sexual Excitation Scales. In addition, other questionnaires not relevant to the study and not assessing sexual aggression or sexuality were included in the packet in an effort to obscure the potential relationship of the experimental measures to the stimuli.

The final part of the study involved debriefing participants. All participants received a debriefing statement similar to that of Malamuth and Check (1984) along with extra credit in undergraduate psychology classes for their participation. The debriefing statement presented several points intended to dispel rape myths and to emphasize the serious and deleterious consequences of rape.

RESULTS

Manipulation Checks

Descriptive data from the eight questions designed to ensure that participants listened attentively to both the nonconsensual and consensual stories suggested that participants did in fact pay attention to both stories. Over 90% of the participants answered seven of the eight questions correctly. The remaining question was answered correctly by 62% of participants.

Two questions ("To what extent do you do you think the woman really wanted to have sex?" and "To what extent do you think the woman enjoyed what was happening to her?") on the sexual arousal questionnaire were analyzed to determine whether participants differentiated between stories. Specifically, the goal of these two analyses was to determine if participants accurately perceived that the woman in the nonconsensual story did not want to have sex and did not enjoy herself relative to the woman in the consensual but forceful story. Two one-way repeated-measure ANOVAs [F(1,107) = 2946.36, p < .001; F(1,107) = 1887.839, p < .001] were conducted, revealing that the woman in the consensual but forceful story was perceived as enjoying herself more (M = 6.29, SD = .95; M = 1.13, SD = .36) and as desiring sex more (M = 6.03, SD = 1.08; M = 1.19, SD = .50) than the woman in the nonconsensual story.

Data Analysis

Preliminary analyses indicated that many of the items on the sexual arousal questionnaire were moderately to highly intercorrelated. Thus, an exploratory factor analysis of the questions on the sexual arousal questionnaire was conducted for each of the stories. A principal component analysis, a varimax rotation, and a .512 factor-loading cut-off (Stevens, 1996) produced the same two-factor structure for responses to both stories. The first factor was composed of the four questions assessing sexual arousal, sexual desire, likelihood of reading similar stories, and likelihood of acting like the man in the story. This factor was labeled Participant's Sexual Activation. The second factor consisted of the two questions assessing the extent to which participants thought the woman enjoyed what had happened and the extent to which they thought she wanted to have sex. This factor was labeled Perceptions of Woman's Sexual Responsiveness.

Cronbach's alphas for the two factors ranged from .67 to .84 for responses to both stories, and the factor loadings for individual items ranged from .55 to .93 for both stories. In addition, the two factors together accounted for 73% of the variance in responses for both stories. Factor scores were computed for data analyses by summing across individual items within each factor. Higher scores on Perceptions of Woman's Sexual Responsiveness reflected a more positive appraisal of the woman's sexual experience, and higher scores on Participant's Sexual Activation reflected a greater level of overall sexual responsiveness to a story. Means and standard deviations of the factor scores for the nonconsensual story are as follows: Participant's Sexual Activation (M = 7.7, SD = 4.1) and Perceptions of Woman's Sexual Responsiveness (M = 2.3, SD = .76). Means and standard deviations of the factor scores for the consensual story are as follows: Participant's Sexual Activation (M = 15.4, SD = 5.2) and Perceptions of Woman's Sexual Responsiveness (M = 12.3, SD = 1.9)

Hierarchical multiple regression analyses were performed on the two factor scores for both the consensual and nonconsensual sexually aggressive story. The order in which a story was presented and social desirability were entered as predictors in the first block of each hierarchical regression. Story order was included to control for the possibility that the order in which the stories were presented would affect participants' responses to both stories, despite counterbalancing and the insertion of the filler or masking story. Erotophilia-erotophobia, attraction to sexual aggression, sexual excitation, sexual inhibition-threat of performance failure, and sexual inhibition-threat of performance consequences were then entered as predictors in the second block of each hierarchical regression. Bivariate correlations between the predictor variables were all less than .80 (see Table 1), and the tolerances for each of the predictors were well above .10, suggesting that multi-collinearity was not a concern in these analyses.

Responses to the Consensual Story

In the first step of the first analysis, social desirability and story order were entered and they accounted for a statistically significant [F(2,105) = 4.15, p < .05] amount of variance ([R.sup.2] - .073) in first factor, Participant's Sexual Activation, with order being a statistically significant predictor. The subsequent entry of interest in sexual aggression and overall sexual interest predictors revealed that there was a statistically significant increment [F change (5, 100) = 12.15, p < .001] in the amount of variance explained ([R.sup.2] change = .35). With the addition of these predictors, the overall model, including social desirability and story order, accounted for a statistically significant [F (7, 100) = 10.49, p < .001] amount of variance ([R.sup.2] = .42). Sexual excitation was the only statistically significant predictor, indicating that higher levels of sexual excitation were associated with a greater level of sexual responsiveness to the consensual story. The standardized weights (betas) for the predictors can be found in Table 2.

After controlling for social desirability and order, the interest in sexual aggression and overall sexual interest predictors did not account for a statistically significant amount of variance in scores on the second factor, Perceptions of Woman's Sexual Responsiveness, for the consensual story.

Responses to the Nonconsensual Story

In the first step of the analysis, social desirability and story order were entered but did not account for a significant amount of variance ([R.sup.2] = .024) in the first factor, Participant's Sexual Activation. When the interest in sexual aggression and overall sexual interest predictors were entered in the second step, there was a statistically significant increment [F change (5, 100) = 4.26, p < .001] in the amount of variance explained ([R.sup.2] change =. 17). With the addition of these predictors, the overall model, including social desirability and story order, accounted for a statistically significant [F (7, 100) = 3.45, p < .001] amount of variance ([R.sup.2] = .195). Sexual excitation and attraction to sexual aggression were statistically significant predictors, indicating that higher levels of both constructs were associated with a greater level of sexual responsiveness to the nonconsensual story. The standardized weights (betas) for the predictors can be found in Table 3.

In the first step of the analysis for the second factor, Perceptions of Woman's Sexual Responsiveness, social desirability and order accounted for a statistically significant [F(2,105) = 3.81, p < .05] amount of variance ([R.sup.2] = .068), with order being a statistically significant predictor. Interest in sexual aggression and overall sexual interest predictors were entered in the second step, and there was a statistically significant increment [F change (5, 100) = 2.53, p < .01] in the amount of variance explained ([R.sup.2] change =. 105). With the addition of these predictors, the overall model, including story order and social desirability, accounted for a statistically significant [F (7, 100) = 2.98, p < .01] amount of variance ([R.sup.2] = .17). The Sexual Inhibition-Threat of Performance Failure subscale of the Sexual Inhibition and Sexual Excitation Scales was a statistically significant predictor in the equation, indicating that higher levels of sexual inhibition were associated with perceiving the woman's experience as more positive. Attraction to sexual aggression was also a statistically significant predictor, suggesting an association between higher levels of this construct and an increase in positive perceptions of the woman's experience. The standardized weights (betas) for the predictors can be found in Table 3.

DISCUSSION

The responses to the sexual arousal questionnaire were factor analyzed because of statistically significant intercorrelations among items. This factor analysis revealed a two-factor structure. The first factor (Participant's Sexual Activation) was a composite of the items assessing sexual arousal, sexual desire, likelihood of reading similar stories, and likelihood of acting like the man in the story. The second factor (Perceptions of Woman's Sexual Responsiveness) was a composite of the items assessing the extent to which participants thought the woman enjoyed what had happened and the extent to which they thought she wanted to have sex. Participant's Sexual Activation reflected a participant's sexual responsiveness to a story while the Perceptions of the Woman's Sexual Responsiveness reflected the extent to which participants perceived the woman's sexual experience as positive.

In general, this study's results suggest the woman's consent, or lack thereof, influenced the predictive utility of both sexual excitation and attraction to sexual aggression. When the woman's consent was clearly present, sexual excitation was the primary predictor of participants' self-report responses, whereas when the woman did not consent, attraction to sexual aggression became important.

In response to the consensual story, sexual excitation was the strongest predictor of scores on the first factor, Participant's Sexual Activation, indicating that men who reported being easily sexually aroused had a higher level of sexual responsiveness to this story. This finding was expected because men who report greater sexual excitation by definition have a greater affinity for sexually explicit materials. The data for the second factor, Perceptions of Woman's Sexual Responsiveness, indicated that only the two control variables, social desirability and order, were significantly associated with perceptions of the woman's experience. This finding suggests that men who are not as concerned about presenting themselves in a socially desirable manner may be similarly less concerned with reporting that a woman's experience in an aggressive or forceful but consensual sexual encounter was positive.

Responses to the nonconsensual story yielded a different pattern of results. Although higher levels of sexual excitation were associated with a greater level of sexual responsiveness for the nonconsensual story, as predicted, this relationship was weaker than it was for the consensual story. This finding is consistent with the results found by Byers and Eno (1991) which suggest that the tendency to perceive oneself as easily sexually aroused is associated with sexual arousal to coercive sexual stimuli and a willingness to behave in a sexually coercive manner. In addition, as predicted, attraction to sexual aggression emerged as a predictor of this factor for the nonconsensual story. This finding is consistent with previous research on attraction to sexual aggression (Malamuth, 1989a, 1989b) which indicates that an interest in sexual aggression tends to be linked to a greater level of sexual responsiveness to coercive sexual stimuli, as well as a greater willingness to behave in a sexually coercive fashion.

The data for the nonconsensual story for the second factor, Perceptions of Woman's Responsiveness, revealed a surprising pattern of results. Sexual inhibition emerged as a significant predictor of this factor. Specifically, high levels of sexual inhibition involving concerns about sexual performance failure were associated with more positive perceptions of a rape victim's experience. We had predicted, based on previous research, that low levels of sexual inhibition would be associated with this outcome. This previous research, however, found a relationship between sexual arousal to sexually coercive stimuli and low levels of sexual inhibition concerning the negative consequences of sexuality and not fears of sexual performance failure (Janssen, Vorst, Finn, & Bancroft, 2002b). The current finding may be similar to other research that has found a relationship between high levels of sexual inhibitions or conflict and the endorsement of rape-supportive attitudes (Briere, Malamuth, & Check, 1983). That is, sexually conflicted men may endorse rape-supportive attitudes, and in turn may give a more positive appraisal of a rape victim's sexual experience. This, however, is a highly speculative conclusion given the current data and previous literature.

Attraction to sexual aggression was also a significant predictor of participants' ratings of the woman's sexual responsiveness in the nonconsensual story, suggesting that there is a relationship between a greater interest in sexual aggression and more positive perceptions of a rape victim's sexual experience. This finding is consistent with previous research from Malamuth and Brown (1994) which found that self-reported history of sexual aggression and rape-supportive attitudes were associated with a tendency to discredit a woman's sexual communications. Similarly, the present study demonstrated that interest in sexual aggression was associated with a greater likelihood of viewing a rape victim's experience as pleasant, despite the obvious pain and discomfort she experienced. In general, it appears that college men who have a propensity for or interest in sexually coercive behavior discount the veracity of a rape victim's sexual communications.

There are a few noteworthy limitations of the present study. The first involves the ecological validity of the stimuli. The extent to which the stories accurately depict rape and sexual aggression is unknown. Moreover, the stories have a strong male-oriented erotic quality to them and this could have augmented the association between sexual excitation and some of the dependent measures; however, this association is consistent with previous studies that have found a similar effect (e.g., Byers & Eno, 1991). Unfortunately, the design of the present study does not lend itself to an examination of this issue. Thus, the findings of this investigation should be generalized with care beyond the laboratory and the context of this study.

A third limitation involves the presence of order effects. When participants listened to the consensual, forceful story after the nonconsensual story, they reported a more positive impression of the woman's experience and a higher level of sexual responsiveness in response to the consensual story; these effects were in the opposite direction when the consensual story was presented first. Similarly, when participants listened to the nonconsensual story second, they reported a less positive impression of the woman's experience in response to the nonconsensual story. Again, this effect was in the opposite direction when the nonconsensual story was presented first.

It appears that these order effects reflect an effect of context in which the two stories were contrasted with each other and the first affected the ratings of the other. Again, this suggests that cautiousness is warranted when attempting to generalize these findings beyond the context of this study. The order effects also suggest that the context in which analogue stimuli are presented is important and should be carefully considered by those conducting research in this area. It is important to note that order effects were entered into the multiple regressions as controls, and therefore significant effects were found even after controlling and accounting for the order in which the stories were presented (i.e., the context). Nevertheless, researchers interested in conducting investigations using similar methods may want to consider a between-subjects design as a way of eliminating order effects and potentially clarifying interpretations.

The present investigation has two important implications for sexual aggression research. First, the presence or absence of the woman's consent in sexually coercive analogue stimuli has often been obscured in the literature. Our data demonstrate the importance of making the woman's nonconsent clear in such stimuli because this can affect the outcome of a study. For example, results from studies (e.g., Norris & Kerr, 1993) that use sexually aggressive but consensual stimuli (i.e., fantasy rape) may not have external validity (i.e., not generalize to assault or rape in the real world). Thus, it may be important to convey clearly in analogue stimuli a woman's nonconsent, regardless of the extent of forceful sexual behaviors, if the findings of such studies are to be generalized to real-world sexual violence.

Our data also suggest that attempts to study men with proclivities for sexually aggressive behavior could benefit from the inclusion of overall sexual interest measures. It appears that sexual excitation and attraction to sexual aggression complement one another in the prediction of responses to sexually coercive stimuli. Specifically, men who report high levels of sexual excitation may be more likely to have an affinity for sexually coercive behaviors; however, this conclusion is tenuous considering the analogue nature of our study. In addition, the relationship between sexual inhibition and perception of a rape victim's experience warrants further attention as well. Future studies that include more naturalistic measures of sexual aggression than does the present study should be conducted to examine the relationship of sexual aggression to sexual excitation and sexual inhibition. It is possible that these constructs would add predictive strength to comprehensive models of sexual aggression like that of Malamuth and his colleagues (Malamuth, Linz, Heavey, Barnes, & Acker, 1995).

Table 1. Intercorrelations Between Predictor Variables

         Predictor                    1    2     3      4       5

Sexual excitation                     --  .069  .009  -.117   .611 *
Attraction to sexual aggression            --   .092  -.123   .049
Sexual inhibition-
  threat of performance failure                  --    .126   .064
Sexual inhibition-
  threat of performance consequences                    --   -.099
Erotophobia-erotophilia

* p < .01.

Table 2. Standardized (Beta) Weights From the Multiple Regression
Analyses for the Consensual Story

                                  Participant's      Perception of
                                     sexual         woman's sexual
                                  activation (a)   responsiveness (b)

Block one
   Order                             .194 *              .321 **
   Social desirability              -.185               -.24 **

Block two
   Sexual excitation                 .525 ***             .076
   Sexual inhibition-threat of
     performance failure             .161                -.043
   Sexual inhibition-threat of
     performance consequences        .023                 .078
Erotophobia-erotophilia              .086                 .171
Attraction to sexual aggression      .056                 .049

Note. Block one df = (2, 105). Block two df = (6, 99).

(a) Composite of items assessing self-reported sexual arousal,
sexual desire, likelihood of reading similar stories, and
likelihood of acting like the man in the story.

(b) Composite of items assessing perceptions of the woman's level
of enjoyment and sexual desire.

* p < .05.

** p < .01.

*** p < .001.

Table 3. Standardized (Beta) Weights From the Multiple Regression
Analyses for the Nonconsensual Story

                                  Participant's      Perception of
                                      sexual        woman's sexual
                                  activation (a)   responsiveness (b)

Block one
   Order                              -.132              -.256 **
   Social desirability                -.079              -.04

Block two
   Sexual excitation                   .261 *             .144
   Sexual inhibition-threat of         .111               .200 *
     performance failure
   Sexual inhibition-threat of        -.128              -.031
     performance consequences
Erotophobia-erotophilia               -.053              -.063
Attraction to sexual aggression        .276 **            .210 *

Note. Block one df = (2, 105). Block two df = (6, 99).

(a) Composite of items assessing self-reported sexual arousal,
sexual desire, likelihood of reading similar stories, and likelihood
of acting like the man in the story.

(b) Composite of items assessing perceptions of the woman's level of
enjoyment and sexual desire.

* p < .05.

** p < .01.

REFERENCES

Briere, J., & Malamuth, N. M. (1983). Self-reported likelihood of sexually aggressive behavior: Attitudinal versus sexual explanations. Journal of Research in Personality, 17, 315-323.

Briere, J., Malamuth, N. M., & Check, J. V. P. (1983). Sexuality and rape supportive beliefs. In P. Caplan (Ed.), Feminist psychology. Toronto: Eden Press.

Brownmiller, S. (1975). Against our will: Men, women, and rape. New York: Simon and Schuster.

Burt, M. (1980). Cultural myths and supports for rape. Journal of Personality and Social Psychology, 38, 217-230.

Byers, S. E., & Eno, R. J. (1991). Predicting men's sexual coercion and aggression from attitudes, dating history, and sexual response. Journal of Psychology and Human Sexuality, 4, 55-70.

Calhoun, K. S., Bernat, J. A., Clum, G. A., & Frame, C. L. (I 997). Sexual coercion and attraction to sexual aggression in a community sample of young men. Journal of a Interpersonal Violence, 12, 392-406

Caron S. L., & Carter, B. D. (1997). The relationships among sex role orientation, egalitarianism, attitudes toward sexuality, and attitudes toward violence against women. Journal of Social Psychology, 137, 568-587.

Check, J. V. P. (1985). The hostility towards women scale. Unpublished doctoral dissertation, University of Manitoba, Winnipeg, Manitoba, Canada.

Crowne, D. P., & Marlowe, D. (1960). A new scale of social desirability independent of psychopathology. Journal of Consulting Psychology, 24, 349-354.

Fisher, W. A., Byrne, D., White, L. A., & Kelley, K. (1988). Erotophobia-Erotophilia as a dimension of personality. The Journal of Sex Research, 25, 123-151.

Hudson, S. M., & Ward, T. (1997). Rape: Psychopathology and theory. In D. R. Laws & W. O'Donohue (Eds.), Sexual deviance: Theory, assessment, and treatment (pp. 332-355). New York: Guilford Press.

Hurlbert, D. F., & Apt, C. (1991). Sexual narcissism and the abusive male. Journal of Sex and Marital Therapy, 17, 279-292.

Janssen, E., Vorst, H., Finn, P., & Bancroft, J. (2002a). The Sexual Inhibition (SIS) and Sexual Excitation (SES) Scales: I. Measuring sexual inhibition and excitation proneness in men. The Journal of Sex Research, 39, 114-126.

Janssen, E., Vorst, H., Finn, P., & Bancroft, J. (2002b). The Sexual Inhibition (SIS) and Sexual Excitation (SES) Scales: II. Predicting psychophysiological response patterns. The Journal of Sex Research, 39, 127-132.

Malamuth, N. M. (1986). Predictors of naturalistic sexual aggression. Journal of Personality and Social Psychology, 50, 953-962.

Malamuth, N. M. (1989a). The Attraction to Sexual Aggression Scale: Part one. The Journal of Sex Research, 26, 2649.

Malamuth, N. M. (1989b). The Attraction to Sexual Aggression Scale: Part two. The Journal of Sex Research, 26, 324-354.

Malamuth, N. M., & Brown, L. M. (1994). Sexually aggressive men's perceptions of women's communications: Testing three explanations. Journal of Personality and Social Psychology, 67, 677-712.

Malamuth, N. M., & Check, J. V. P. (1984). Debriefing effectiveness following exposure to rape depictions. The Journal of Sex Research, 20, 1-13.

Malamuth, N. M., Check. J. V. P., & Briere, J. (1986). Sexual arousal in response to aggression: Ideological, aggressive, and sexual correlates. Journal of Personality and Social Psychology, 50, 330-340.

Malamuth, N. M., Heavey, C. L., & Linz, D. (1993). Predicting men's antisocial behavior against women: The interaction model of sexual aggression. In G. C. N. Hall, R. Hirschman, J. R. Graham, & M. S. Zaragoza (Eds.), Sexual aggression: Issues in etiology, assessment and treatment (pp. 63-97). Washington, DC: Taylor and Francis.

Malamuth, N. M., Linz, D., Heavey, C. L., Barnes, G., & Acker, M. (1995). Using the confluence model of sexual aggression to predict men's conflict with women: A 10-year follow-up study. Journal of Personality and Social Psychology, 69, 353-369.

Norris, J., & Kerr, K. (1993). Alcohol and violent pornography: Responses to permissive and nonpermissive cues. Journal of Studies on Alcohol, 54(Suppl. 11), 118-127.

Palmer, C. T. (1988). Twelve reasons why rape is not sexually motivated: A skeptical examination. The Journal of Sex Research, 25, 512-530.

Stevens, J. (1996). Applied multivariate statistics for social sciences (3rd ed.). Mahwah, NJ: Lawrence Erlbaum Associates.

Manuscript accepted April 23, 2002

Brian A. Wilson, Jeffrey E. Holm, Kelly L. Bishop, and Dana M. Borowiak
University of North Dakota

Address correspondence to Jeffrey Holm, Department of Psychology, University of North Dakota, P.O. Box 8380, Grand Forks, ND 58202; e-mail: [e-mail].

-0.5 cially unacce