Men sometimes attempt to sexually coerce their intimate partners. In the context of an intimate relationship, sexual coercion can include forcible rape but often takes the form of more subtle tactics, such as withholding financial resources if a woman does not consent to sex (Carr and VanDeusen 2004; Johnson and Sigler 2000; Marshall and Holtzworth-Munroe 2002; Shackelford and Goetz 2004). Regardless of whether sexual coercion is physical or non-physical, women who have been sexually coerced by an intimate partner experience negative physical and psychological consequences (e.g., Campbell 1989; Livingston et al. 2004; Zweig et al. 1999). If we can identify the predictors and correlates of sexual coercion, we will be better positioned to reduce or eliminate this costly behavior.

Previous research has identified several predictors of sexual coercion of women by men. Examples of these predictors include male acceptance of traditional gender roles and male sexual dominance (Muehlenhard and Falcon 1990), fraternity affiliation among male college students (Lackie and de Man 1997), men’s lack of empathy for women (Simons and Wurtele 2002), and male narcissism in conjunction with a woman’s refusal of a man’s sexual advances (Baumeister et al. 2002). Several predictors of sexual coercion in the context of an intimate relationship also have been identified. These include male low self-esteem (Burke et al. 1988), male physical and psychological partner-directed aggression (Marshall and Holtzworth-Munroe 2002), male alcohol and pornography consumption (Carr and VanDeusen 2004), female infidelity (Goetz and Shackelford 2006), and male sexual jealousy (Frieze 1983). The current research investigated another potential statistical predictor of sexual coercion in intimate relationships: men’s partner-directed insults. Previous research has documented a positive relationship between men’s partner-directed insults and men’s use of partner-directed non-physical coercive behavior as well as physical violence (Goetz et al. 2006). Given the established link between men’s partner-directed insults and their general coercive behavior in a relationship, it is reasonable to hypothesize that insults are related to specific forms of coercive behavior, in this instance sexual coercion.

The current studies investigate men’s sexual coercion using the Sexual Coercion in Intimate Relationships Scale (SCIRS; Shackelford and Goetz 2004). The SCIRS differs from other measures of sexual coercion, such as the Aggressive Sexual Behavior Inventory (Mosher and Anderson 1986), the Sexual Situation Questionnaire (O’Sullivan and Byers1993), and the Coercive Sexuality Scale (Rapaport and Burkhart 1984), in that it specifically assesses coercion in the context of an intimate relationship, rather than between casually dating partners. Assessing sexual coercion in the context of a casual dating relationship is important. However, sexual coercion in a committed intimate relationship may be different from sexual coercion in a casual dating relationship. Because a couple in a committed relationship may likely be more compatible and be more considerate and caring toward one another than a couple in a casual dating relationship (see Buss 2004), sexual coercion tactics may be more innocuous, subtle, and discrete in a committed relationship. The SCIRS is unique in that it accounts for these variations by including assessments of tactics that vary in subtlety (e.g., withholding benefits and hinting about withholding benefits). The assessment of tactics that vary in subtlety may be important because sexual coercion in intimate relationships can be both conspicuous and discreet (Shackelford and Goetz 2004).

The current studies investigated the content of men’s insults using the Partner-Directed Insults Scale (PDIS; Goetz et al. 2006). Prior to the development of the PDIS, no measure was available to assess the specific content of the insults men use to derogate their partners. Previously established measures that broadly assess verbal abuse in an intimate relationship, such as the Conflict Tactics Scale (Straus 1979), the Psychological Maltreatment of Women Inventory (Tolman 1989), the Index of Psychological Abuse (Sullivan et al. 1991), and the Measurement of Wife Abuse (Rodenburg and Fantuzzo 1993), typically assess only the frequency with which an individual yells at or insults their partner—they do not assess the specific content of the insults directed at their partner. For example, although the Conflict Tactics Scale (Straus 1979) includes a subscale on verbal aggression, it does not assess the content of the partner-directed insults. Some measures of psychological abuse include subscales of verbal abuse that assess the content of insults, but these assessments are typically restricted to a few items. The Psychological Maltreatment of Women Inventory (Tolman 1989), for example, includes items such as “My partner told me my feelings were irrational or crazy” and “My partner blamed me for his problems,” but of the 50 total items, only five mention the specific content of an insult.

The PDIS is the first scale designed to evaluate explicitly both the content of the specific insults as well as the frequency with which a man uses these insults against his partner. Goetz et al. (2006) documented positive relationships between scores on the PDIS and measures of both physical and non-physical partner-directed violence. Given the established link between partner-directed insults and general coercive behavior, it is reasonable to hypothesize a relationship between partner-directed insults and sexually coercive behavior. Specifically, we hypothesized that men’s use of partner-directed insults is related positively to their sexual coercion in the context of an intimate relationship. We tested this hypothesis in two studies. Study 1 secured men’s self-reports of their own sexual coercion and partner-directed insults in their current intimate relationship. Study 2 secured women’s reports of their partners’ sexual coercion and insults in their current relationship.

Study 1: Men’s Self-Reports of Partner-Directed Insults and Sexual Coercion

Methods

Participants

Two hundred forty-seven men, each of whom was in a committed, sexual relationship with a woman, participated in this study. The mean age of the participants was 25.8 years (SD = 10.0), the mean age of the participants’ partners was 24.7 years (SD = 8.9), and the mean relationship duration was 43.2 months (SD = 63.6). All participants were drawn from universities and surrounding metropolitan communities. About half of the participants were university students approached at the beginning of several psychology, sociology, and biology class meetings. The remaining participants were community members who were known and recruited by students of the university. Unfortunately, we did not record whether a participant was a current student and so cannot include this as a variable in the analyses. No additional demographic information is available on these participants.

Materials

Participants completed a survey that included several sections. The first section solicited demographic information, including the participant’s age, his partner’s age, and the duration of his current relationship. The second section of the survey assessed men’s partner-directed insults using the Partner-Directed Insults Scale (PDIS; Goetz et al. 2006). The PDIS evaluates both the content of the specific insults as well as the frequency with which the participant uses these insults against his partner. Each of the 47 insults is categorized into one of four components: Derogating Physical Attractiveness (e.g., “I told my partner that her breasts are ugly”), Derogating Value as a Partner/Mental Capacity (e.g., “I told my partner that she makes my life miserable”; “I called my partner an idiot”), Derogating Value as a Person (e.g., “I told my partner that nothing she does is important”), and Accusations of Sexual Infidelity (e.g., “I accused my partner of having sex with many other men”). Instructions for the PDIS are as follows: “Men sometimes try to hurt their female partner’s feelings by saying insulting things to them. The following list includes insulting things that a man might say to his partner. In the column labeled ‘How often (Use scale),’ write the number from the scale below to indicate HOW OFTEN you have said each insulting thing to your partner.”

Responses are recorded using a six-point ordered-category scale with the following values: 0 = I have never said this insulting thing to my partner, 1 = I have said this insulting thing to my partner 1 time, 2 = I have said this insulting thing to my partner 5 times, 3 = I have said this insulting thing to my partner 6 to 10 times, 4 = I have said this insulting thing to my partner 11 to 24 times, 5 = I have said this insulting thing to my partner 25 or more times. Scores for each component were calculated by summing the response values for each item in that component. Full scale scores were calculated by summing response values for each item in the entire scale. Previous research has established the reliability, validity, and utility of the PDIS as an assessment of the content and frequency of the insults that men direct at their intimate partners (Goetz et al. 2006).

To assess men’s sexual coercion in the current relationship, the last section of the survey included the Sexual Coercion in Intimate Relationships Scale (SCIRS; Shackelford and Goetz 2004). The SCIRS secures information about how often the participant performed 34 sexually coercive acts. Items in the SCIRS vary in subtlety, ranging from hinting and subtle manipulations to outright physical force. These 34 items cluster into three components: Resource Manipulation/Violence (e.g., “I hinted that I would withhold benefits that my partner depends on if she did not have sex with me;” “I physically forced my partner to have sex with me”), Commitment Manipulation (e.g., “I told my partner that if she loved me she would have sex with me”), and Defection Threat (e.g. “I hinted that I would have sex with another woman if my partner did not have sex with me”). Instructions for the SCIRS are as follows: “Sexuality is an important part of romantic relationships and can sometimes be a source of conflict. Your honest responses to the following questions will contribute profoundly to what is known about sexuality in romantic relationships and may help couples improve the sexual aspects of their relationships. We appreciate that some of the questions may be uncomfortable for you to answer, but keep in mind that your responses will remain confidential. Below is a list of acts that can occur in a romantic relationship. Please use the following scale to indicate HOW OFTEN in the past ONE month these acts have occurred in your current romantic relationship. Write the number that best represents your response in the blank space to the left of each act.”

Responses were recorded using a six-point ordered-category scale with the following values: 0 = Act did not occur in the past month, 1 = Act occurred 1 time in the past month, 2 = Act occurred 2 times in the past month, 3 = Act occurred 3 to 5 times in the past month, 4 = Act occurred 6 to 10 times in the past month, and 5 = Act occurred 11 or more times in the past month. Scores for each component were calculated by summing the response values for each item in that component. Full scale scores were calculated by summing response values for each item in the entire scale. Previous research has established the reliability, validity, and utility of the SCIRS as an assessment of sexual coercion in intimate relationships (Goetz and Shackelford 2006; Shackelford and Goetz 2004).

Procedure

Three criteria had to be met to qualify for participation. The prospective participant had to be (1) male, (2) at least 18 years of age, and (3) currently involved in a committed, sexual relationship with a woman. If the criteria were met, the researcher handed the participant a consent form, the survey, and a security envelope. The participant was instructed to read and sign the consent form, complete the survey, place the completed survey in the envelope, and then seal the envelope. The participant was instructed not to seal the consent form inside the envelope to maintain anonymity.

Results and Discussion

The alpha reliabilities for the full-scale PDIS and four components of the PDIS (Derogating Physical Attractiveness, Derogating Value as a Partner/Mental Capacity, Derogating Value as a Person, and Accusations of Sexual Infidelity), were α = 0.91, 0.91, 0.83, 0.74, and 0.83, respectively. The alpha reliabilities for the full-scale SCIRS and the three components of the SCIRS (Resource Manipulation/Violence, Commitment Manipulation, and Defection Threat) were α = 0.90, 0.77, 0.83, and 0.87, respectively.

The hypothesis that men’s use of partner-directed insults is related positively to sexual coercion in the context of an intimate relationship was supported: men’s full-scale scores on the PDIS correlated positively with their full-scale scores on the SCIRS: r = 0.37, p <0.001. Men who reported greater use of insults against their partner also reported greater sexual coercion against their partner. This relationship remained positive and statistically significant even after controlling for participant age, partner age, and relationship duration: partial r = 0.37, p < 0.001. A test of the difference between the zero-order and partial correlations, using Fisher’s r-to-z transformation, did not reach significance (z < 1.0, p > 0.05). For reportorial completeness, we correlated each of the four components of the PDIS with the three components of the SCIRS (see Table 1). All correlations were positive and significant except for one. We instituted a Bonferroni correction for α inflation that produced a per-prediction corrected α level of (0.05/20) = 0.0025 (see Cohen and Cohen 1983). Using this corrected α, only two of the 19 originally significant correlations became non-significant (see Table 1). These relationships remained positive and statistically significant after controlling statistically for participant age, partner age, and relationship duration. In addition, none of the tests of the differences between zero-order and partial correlations reached significance (all zs < 1.0, all ps > 0.05; analyses available on request).

Table 1 Study 1: Correlations between the PDIS (total and four components) and the SCIRS (total and three components) according to men’s self-reports
Full size table

Unique Predictive Utility of the Insult Components

To identify whether any of the PDIS components uniquely predicted men’s sexual coercion against their partners, we conducted a multiple regression using scores on the four PDIS components to predict full-scale SCIRS scores. The overall model was significant (F = 12.58, R 2 = 0.17, p < 0.001). Investigation of the individual standardized regression coefficients indicated that three of the four PDIS components uniquely predicted men’s total SCIRS scores: Derogating Value as a Partner/Mental Competency, Derogating Value as a Person, and Accusations of Sexual Infidelity (see Table 2). These results did not change after controlling for participant age, partner age, and relationship duration (analyses available on request).

Table 2 Study 1: Multiple regression analyses (reported in standardized beta weights), using men’s self-reports
Full size table

For reportorial completeness, we conducted three additional multiple regressions, using all four of the PDIS components to predict each of the three SCIRS components. The results are displayed in Table 2 and indicate that Resource Manipulation was significantly and uniquely predicted by Derogating Value as a Person and Accusations of Sexual Infidelity and that Commitment Manipulation was significantly and uniquely predicted by Derogating Value as a Partner/Mental Competency. Defection Threat was not significantly predicted by any of the PDIS components. This pattern of results persisted after controlling statistically for participant age, partner age, and relationship duration (analyses available on request).

Study 2: Women’s Reports of Men’s Partner-Directed Insults and Sexual Coercion

Men’s self-reports of their partner-directed insults and sexual coercion may not provide accurate assessments of these behaviors (e.g., Dobash et al. 1998; Magdol et al. 1997). Men may be reluctant to report their partner-directed insults and sexual coercion or, if they do, they may underreport the most egregious insults or the most severe forms of sexual coercion (e.g., Dobash et al. 1998). Women’s reports of their partner’s sexual coercion and partner-directed insults may reflect more accurately the incidence of such behaviors. Using an independent sample of women in committed, sexual relationships, Study 2 secured women’s reports of their partner’s sexual coercion and insults. These independent reports offered an additional test of the hypothesis tested in Study 1.

Methods

Participants

Three hundred seventy-eight women, each of whom was in a committed, sexual relationship with a man, participated in this study. The mean age of the participants was 25.5 years (SD = 7.9), the mean age of the participants’ partners was 28.4 years (SD = 9.3), and the mean relationship duration was 51.4 months (SD = 62.8). Participants were obtained in the same manner as in Study 1. No additional demographic information is available on these participants. None of the women in Study 2 were partners of the men in Study 1.

Materials

The survey for Study 2 paralleled the survey used in Study 1. Participants in Study 2 reported their partner’s use of insults and sexual coercion in the current relationship using versions of the SCIRS and PDIS in which the wording was changed to accommodate reporting of a partner’s behavior.

Procedures

As in Study 1, three criteria must have been met to qualify for participation. The prospective participant had to be (1) female, (2) at least 18 years of age, and (3) currently involved in a committed, sexual relationship with a man. The same procedure was followed as in Study 1.

Results and Discussion

The alpha reliabilities for the full-scale PDIS and four components of the PDIS (Derogating Physical Attractiveness, Derogating Value as a Partner/Mental Capacity, Derogating Value as a Person, and Accusations of Sexual Infidelity), were α = 0.92, 0.85, 0.87, 0.82, and 0.87, respectively. The alpha reliabilities for the full-scale SCIRS and the three components of the SCIRS (Resource Manipulation/ Violence, Commitment Manipulation, and Defection Threat) were α = 0.90, 0.85, 0.86, and 0.90, respectively.

The hypothesis tested in Study 2 paralleled the hypothesis tested in Study 1: Men’s use of partner-directed insults is related positively to their sexual coercion in the context of an intimate relationship. Consistent with this hypothesis, and with the results reported in Study 1, women’s reports of men’s use of partner-directed insults correlated positively with their sexual coercion: r = 0.64, p < 0.001. According to women’s partner-reports, men who more frequently directed insults at their partners also were more sexually coercive against their partner. As in Study 1, this correlation remained positive and statistically significant even after controlling for participant age, partner’s age, and relationship duration: partial r = 0.63, p < 0.001. A test of the difference between the zero-order and partial correlations, using Fisher’s r-to-z transformation, did not reach significance (z < 1.0, p > 0.05). As in Study 1, we correlated each of the four components of the PDIS with the three components of the SCIRS. All correlations were positive and significant (see Table 3). Again, we instituted a Bonferroni correction for α inflation that produced a per-prediction corrected α level of (0.05/20) = 0.0025 (see Cohen and Cohen 1983). All of the correlations remained significant using the corrected α. These relationships remained positive and statistically significant after controlling statistically for participant age, partner age, and relationship duration (analyses available on request). In addition, none of the tests of the differences between zero-order and partial correlations reached significance (all zs < 1.0, all ps > 0.05; analyses available on request).

Table 3 Study 2: Correlations between the PDIS (total and four components) and the SCIRS (total and three components) according to women’s partner-reports
Full size table

Unique Predictive Utility of the Insult Components

Paralleling Study 1, we conducted a multiple regression using scores on the four PDIS components to predict SCIRS scores to identify whether any of the PDIS components uniquely predicted men’s use of sexual coercion against their partners. Again, the overall model was significant (F = 84.35, R 2 = 0.48, p < 0.001). Investigation of the individual standardized regression coefficients indicated that three of the four PDIS components uniquely predicted men’s total SCIRS scores: Derogating Physical Attractiveness, Derogating Value as a Person, and Accusations of Sexual Infidelity (see Table 4). These results did not change after controlling for participant age, partner age, and relationship duration (analyses available on request).

Table 4 Study 2: Multiple regression analyses (reported in standardized beta weights), using women’s partner-reports
Full size table

For reportorial completeness, we conducted three additional multiple regressions, using all four of the PDIS components to predict each of the three individual SCIRS components. The results are displayed in Table 4 and indicate that Resource Manipulation was significantly and uniquely predicted by all four of the PDIS components and that Commitment Manipulation was significantly and uniquely predicted by each of the PDIS components except Derogating Value as a Person. Defection Threat was significantly and uniquely predicted by each of the PDIS components except Derogating Value as a Partner/Mental Competency. This pattern of results persisted after controlling statistically for participant age, partner age, and relationship duration (analyses available on request).

Comparison of Men’s Self-Reports (Study 1) and Women’s Partner-Reports (Study 2)

We performed Fisher’s r-to-z transformations to compare the magnitudes of the parallel correlations generated by men’s self-reports and women’s partner-reports. Eleven of the 20 correlations obtained from men’s self-reports were significantly smaller than the parallel correlations obtained from women’s partner-reports, and none were significantly larger (analyses available on request). Thus, the magnitudes of these relationships between men’s use of partner-directed insults and their use of sexual coercion against their partner were significantly greater for women’s partner-reports than for men’s self-reports. In addition, all four of the multiple regression models produced from women’s partner-reports accounted for greater variance than did the parallel models produced from men’s self-reports.

For reportorial completeness, we performed tests of the difference between performance frequencies reported by men and women (see Table 5). The results indicated just two sex differences. Women reported a significantly higher performance frequency on one of the SCIRS components, Commitment Manipulation. Men reported significantly higher performance frequency on one of the PDIS components, Derogating Physical Attractiveness.

Table 5 Descriptive statistics of performance frequencies and test of sex differences difference for Study 1 (men’s self-reports) and Study 2 (women’s partner-reports)
Full size table

General Discussion

We hypothesized that men’s sexually coercive behaviors would be related positively to their use of partner-directed insults in the context of an intimate relationship. The results from men’s self-reports (Study 1) and from women’s partner-reports (Study 2) provided two independent lines of support for this hypothesis. Men’s use of sexually coercive behaviors can be statistically predicted from the frequency and content of the insults that they direct at their intimate partners.

A comparison of the results of Studies 1 and 2 indicates that the relationships between men’s use of partner-directed insults and sexual coercion are stronger for women’s partner-reports than for men’s self-reports. These sex-differentiated relationships do not appear to be attributable to sex differences in reported performance frequencies of men’s partner-directed insults and sexual coercion. Future research might investigate the possibility that, relative to men, women may be more attuned to the relationship between men’s use of partner-directed insults and sexual coercion. We speculate that women may be more attuned to the relationship between men’s use of insults and sexual coercion because of the potentially dangerous consequences of being a victim of these behaviors and that this relationship may be less salient for men because men do not have as much to lose from engaging in these behaviors as women have from being victims of these behaviors. However, because the men in Study 1 were not partnered to the women in Study 2, we cannot assess the possibility, for example, that these sex-differentiated relationships might be attributable to differences in the veridicality of men’s reports and women’s reports (see Dobash et al. 1998; Magdol et al. 1997).

The overall pattern of positive links between partner-directed insults and sexually coercive behavior was replicated across both studies. In addition, two of the PDIS components, Derogating Value as a Person and Accusations of Sexual Infidelity, uniquely predicted men’s overall sexual coercion across both studies. Derogating Value as a Person includes insults such as “I called my partner a nobody” and “I told my partner that she is worthless.” Relative to items in the other components, items in this component appear to be the most broad (i.e., not insulting a specific feature or attribute of the partner, but instead derogating the value of the person as a whole). It is possible that the insults are used in a hierarchical fashion, such that the most specific insults are used first and most often, with the most broad and general insults—such as those included in the component Derogating Value as a Person—used as a last effort in a poor relationship nearing termination, a situation which also may foster sexual coercion (e.g., Goetz and Shackelford 2006; Shackelford and Goetz 2004).

The component Accusations of Sexual Infidelity includes insults such as “I accused my partner of having sex with many other men,” and “I called my partner a ‘whore’ or a ‘slut’.” Thus, men who accuse their partners of being sexually unfaithful also are more likely to sexually coerce them. This finding is consistent with research indicating that men sometimes behave as if their partners are their exclusive sexual property, attempting to re-assert “ownership” if these exclusive “rights” are threatened (Frieze 1983; Goetz and Shackelford 2006; Russell 1990; Wilson and Daly 1992).

The strength of the relationships between the use of insults and sexual coercion also may depend on the nature of the relationship. For instance, because the use of insults is related to the use of sexual coercion and partner-directed violence (Goetz et al. 2006), we might expect that men who are physically abusive use a greater number of more emotionally hurtful partner-directed insults than do men who are not physically abusive. The current research indicates that 92% of men and women reported at least one occurrence of insulting behavior in their current relationship (see Table 5). This suggests that almost all men insult their intimate partners at some point in the relationship. However, it may be that some of these insults more strongly predict sexual coercion in abusive relationships than in non-abusive relationships.

One limitation of the current research is the lack of paired partner reports. Because the men and women surveyed were not paired, we cannot assess the possibility that apparent sex differences in the strength of these associations are attributable to differences in the veracity of men’s self-reports and women’s partner-reports. Future research would benefit from obtaining cross-spouse reports to address such concerns. A second limitation is that the data presented here reflect a single assessment. Further research using a methodology that includes repeated assessments over time may provide greater insight into the nature of the links between insults and sexual coercion as well as how these links may change over time. Although we have used insults to predict statistically men’s sexual coercion, it is possible that the actual causal direction is reversed, and that men’s sexual coercion leads to their use of partner-directed insults. Regardless of the causal direction, it may be more practical to use verbal insults, an overt class of behaviors, to predict sexual coercion, behaviors that are less generally apparent and often rather subtle and covert (with the exception of direct physical force; see review by Goetz et al. 2006). Another potential limitation is that social desirability concerns by men and by women might have affected the results. Inclusion of a measure of social desirability in future research in this area will help to address and clarify whether and to what extent social desirability concerns might have affected the current results.

Conclusion

Men’s use of sexual coercion in the context of an intimate relationship can be statistically predicted by the frequency and content of the insults that men direct at their partners. Given the negative consequences of sexual coercion, the ability to predict sexual coercion in intimate relationships could have practical applications, particularly in the identification of and treatment for those involved in abusive intimate relationships. For example, a preventative measure might include informing women that men who engage in verbally abusive behaviors, such as those assessed by the PDIS, may be more likely to engage in sexually coercive behaviors. Another preventative measure might be to inform men who are attempting to limit or reduce their sexual coercion of their partners that they must not only consider changing what they do to their partners, but also what they say to them. Arming health professionals and at-risk men and women with the knowledge of the dangers of sexual coercion as well as other behavioral markers of men who may engage in such coercion could prove useful in preventing abuse in intimate relationships.