Sexual coercion perpetration and victimisation in females: The influence of borderline and histrionic personality traits, rejection sensitivity, and love styles

ABSTRACT

This study investigated sexual coercion (perpetration and victimisation) in women. Women (N = 151) aged 18–63 years (M = 23.34, SD = 8.80) completed standardised questionnaires measuring sexual coercion (nonverbal sexual arousal, emotional manipulation and deception, and exploitation of the intoxicated), personality disorder traits (Borderline and Histrionic), love styles (Eros, Ludus, Storge, Pragma, Mania, and Agape), and rejection sensitivity. Data analyses revealed that together, personality disorder traits, love styles, and rejection sensitivity predicted coercion perpetration involving emotional manipulation and deception. These variables also predicted victimisation involving nonverbal sexual arousal and emotional manipulation and deception. Of these predictors, borderline traits predicted coercion involving emotional manipulation and deception (as both a perpetrator and victim) and victimisation from nonverbal sexual arousal-based coercion. Furthermore, Ludus predicted victimisation involving emotional manipulation and deception, while rejection sensitivity predicted the use of emotional manipulation and deception to coerce a partner.

KEYWORDS:

• Coercion
• female perpetration
• personality disorder
• sexual aggression
• victimisation

Introduction

Sexual coercion, defined as “the act of using pressure, alcohol or drugs, or force to have sexual contact with someone against his or her will” (Struckman-Johnson et al., 2003, p. 76), is associated with increased psychological distress (de Visser et al., 2014) and poor health (Williams et al., 2013). Coercion may involve a range of behaviours that can be separated into four categories. Specifically, sexual arousal (e.g. persistent kissing and touching), emotional manipulation (e.g. blackmail, questioning, or using authority), alcohol and drug intoxication (e.g. purposefully getting a person drunk or taking advantage whilst intoxicated) and physical force (e.g. using physical harm) (Struckman-Johnson et al., 2003).

Research to date has largely focused on male perpetrators and female victims, and consistently reports that men are more likely than women to perpetrate sexual coercion (Struckman-Johnson et al., 2020). For example, in a multi-level analysis with 3480 participants across 10 European counties, nearly 50% of males reported engaging in a least one act of perpetration, compared with 15% of female – rates were higher for men than for women in all countries (see Krahé et al., 2015). Yet evidence from both research studies and large-scale federal agency incident data shows that a proportion of women also engage in each sexually coercive behaviour (e.g. Fernández-Fuertes et al., 2018; Krahé et al., 2003; Stemple & Meyer, 2014; Stemple et al., 2017; Struckman-Johnson et al., 2003; Tomaszewska & Krahé, 2018). There is some suggestion that men and women engage in specific types of coercion – manipulation, intoxication and force tactics – to a roughly similar degree (Schatzel-Murphy et al., 2009), and that there are common predictors of sexual coercion in both sexes (e.g. adversarial sexual beliefs: Hines, 2007). Yet, the importance of predictors of coercion can vary by sex (Hoffmann & Verona, 2019; Schatzel-Murphy et al., 2009). For example, Schatzel-Murphy et al. (2009) found that sexual compulsivity, sexual dominance and sociosexuality were important predictors of sexual coercion in both men and women, but the former most strongly predicted sexual coercion in women while the latter two were stronger in men. As research in the area is scant, it is important to consider factors that elevate and explain females’ engagement in sexually coercive behaviour, as both perpetrators and victims.

The present study tested both the predictive and explanatory power of models of sexual coercive behaviour in women. The models included theoretically pertinent correlates, including personality disorder traits, rejection sensitivity, and love styles. As a secondary objective, we compared these same predictors to models of sexual coercion victimisation. We discuss the theoretical and empirical basis of each predictor below, but our overall aim was to test models that include both traits and attitudinal factors, as both have been identified as predictors of sexual coercion.

Borderline and histrionic personality disorder traits

Personality traits may influence the likelihood of women's use of sexually coercive behaviour (Krahé et al., 2003). Characteristics of the dramatic, emotional, and erratic Cluster B personality may be particularly influential on sexually exploitative tactics as Borderline Personality Disorder (BPD) and Histrionic Personality Disorders (HPD), for example, are characterised by poor impulse control, problematic emotional response, and dysfunctional personal relationships (Bender et al., 2001; Hill et al., 2011). Linehan's biosocial theory of BPD (1993) can be used to understand sexual coercion. That is, sexual coercion – like sexual promiscuity – may be an impulsive and maladaptive behaviour that is either a response to extreme biologically-based emotion dysregulation, or an attempt to modulate one's emotions. Some relationships in which one partner has a borderline personality diagnosis feature low relationship satisfaction, high attachment insecurity, and poor communication (Bouchard, Sabourin, et al., 2009). Indeed, female-only studies have found an association between BPD diagnosis and self-reported psychological dating abuse perpetration (Clift & Dutton, 2011), and that women court mandated to attend violence intervention programmes may be more likely to have BPD diagnosis, in contrast to women in the US general population (Stuart et al., 2006).

It is notable that due to a compulsive fear of rejection, BPD is especially associated with behaviour intended to reduce the threat of real or imagined abandonment (APA, 2013; Staebler et al., 2011). Yet, a feature common to the intimate relationships of individuals with borderline disorder traits is swinging between “hot” and “cold” treatment of partners – that is, idealisation to devaluation – leading to fractured intimacy and elevating the risk of this abandonment or rejection (Schmahl et al., 2004). It is unsurprising, therefore, that some individuals with BPD have been found to ease this implicit fear by using sexually coercive behaviours in their efforts to achieve emotional stability, to counteract their disjointed interpersonal style (Agrawal et al., 2004). It is also suggested that for some women with higher levels of BPD traits, this fear of abandonment may lead to engagement in unwanted sex (Willis & Nelson-Gray, 2017). Women with BPD may also be prone to have sexual difficulties including dysfunctional attitudes toward sex and experiencing sexual pressure from partners (Bouchard, Godbout, et al., 2009).

These findings indicate that BPD may be associated with both increased incidence of sexual victimisation (Sansone et al., 2011; Zanarini et al., 2005) and sexual coercion perpetration (Khan et al., 2017). Based on previous literature, we predict a positive association between borderline personality traits and nonverbal sexual coercion (nonverbal sexual arousal and emotional manipulation and deception) as both a perpetrator and as a victim.

While borderline personality traits may heighten the risk of sexual coercion, the explanatory power of BPD may be less than that of Histrionic Personality Disorders (HPD), which has clear conceptual overlap with sexual coercion perpetration. HPD is characterised by attention seeking, excessive emotionality, and using provocative behaviour to manipulate others (APA, 2013; Dorfman, 2010). For example, women with HPD sometimes demand attention from relationship partners and display increased sexual preoccupation, which may lead to sexual coercion if the partner does not reciprocate (AlaviHejazi et al., 2016; Apt & Hurlbert, 1994). This sexual preoccupation was demonstrated in a study of “sexting” behaviours (sending or receiving erotic or nude images to/from others), which was more common in women with HPD (Ferguson, 2011). HPD traits were also found to be a significant predictor of sexual coercion perpetration by females who exploited intoxicated partners (Hughes et al., 2020). However, the relationship between HPD and sexual aggression remains unclear. For example, women with HPD may display low levels of sexual assertiveness and sexual desire (Apt & Hurlbert, 1994) which are inconsistent with both the perpetration of and being victim of sexually aggressive behaviour. Additional research is therefore required to establish the relationship between HPD or histrionic personality traits and sexual aggression. Based on this, we predict that women who display HPD traits will be more likely to engage in emotionally manipulative sexual coercion and be at greater risk of sexual coercion victimisation.

Rejection sensitivity

Rejection sensitivity is a trait-based construct that characterises individuals with BPD (e.g. Gardner et al., 2010; Sato et al., 2020) and HPD (Lyddon & Sherry, 2001; Meyer & Pilkonis, 2005) but which also presents in individuals without these traits (Norona et al., 2016). Rejection sensitivity refers to “the disposition to anxiously expect, readily perceive and intensely react to rejection by significant others” (Downey et al., 2000, p. 45). Those concerned with sexual abandonment may be more sexually compliant in an attempt to foster relationship closeness and reduce insecurities that arises from rejection sensitivity (Schachner & Shaver, 2004). Women with greater rejection sensitivity, therefore, may consent to unwanted sexual activity in order to fulfil perceived relationship obligations (Impett & Peplau, 2002); indeed, rejection sensitivity is associated with increased sexual aggression victimisation (Young & Furman, 2008). Rejection sensitivity may also influence perpetration of sexual coercion. For example, rejection sensitive women are more likely to report hostility and relationship conflict in response to a perceived rejection (Ayduk et al., 1999; Ayduk et al., 2008). Similarly, rejection is associated with the use of aggression and violence (Romero-Canyas et al., 2010; Volz & Kerig, 2010). Hence, a substantial proportion of violent acts are influenced by the extent to which a person feels loved and accepted (Kahya, 2021) and it is possible that sexually coercive behaviours are used to avoid feelings of rejection. At present, there is a paucity of research investigating the association between rejection sensitivity and sexual coercion (perpetration and victimisation) in women. We predict those with greater sensitivity to rejection will be more likely to both perpetrate sexual coercion and experience it as a victim.

Love styles

While individual differences in personality disorder traits or features thereof may afford some predictive and explanatory power in a model of sexual coercion, other non-trait based individual differences are important, for example, attitudes towards love. One taxonomy conceptualises attitudes towards love in terms of “love styles”, that is, the way in which romantic love is expressed and experienced (Lee, 1973). From a psychosocial perspective, romantic love instils people's expectations and notions of how partners behave in intimate relationships, compared with platonic relationships (Chung, 2005). Yet, the complex interplay between romantic love and abusive conduct in intimate relationships has been almost entirely overlooked in aggression research. This is despite evidence that love and abuse co-exists in many abusive relationships (Pocock et al., 2020), and that different love styles are associated with sexually coercive behaviours in both men (Kalichman et al., 1994; Sarwer et al., 1993) and women (Russell & Oswald, 2001).

According to Lee (1973), primary love styles include Eros (passionate love characterised by strong physical and emotional attraction), Ludus (game playing love based on conquest and “winning” partners), and Storge (love that starts as a friendship and develops into romance). Secondary love styles are a combination of the primary love styles and include Pragma, Mania, and Agape. Pragma (Ludus and Storge) is a practical realistic love where common goals are shared. Mania (Eros and Ludus) is an obsessive, possessive, controlling love. Agape (Eros and Storge) is an altruistic, selfless, and unconditional love based on commitment (Lee, 1973). Previous research has revealed that sexual coercion in men is underpinned by a Ludus love style (Kalichman et al., 1994; Sarwer et al., 1993). Furthermore, women engaging in sexually coercive behaviour are also more likely to adopt a ludic (game playing, manipulative) love style and are less likely to pursue a pragmatic (logical) love style (Russell & Oswald, 2001) whereas men who display Storge and Pragma love styles are more likely to report sexual coercion victimisation (Russell & Oswald, 2002). In this study, we predicted that women who score higher on the Ludus love style and potentially those secondary love styles that include Ludus (i.e. Pragma and Mania) will be more likely to engage in coercive behaviour and less vulnerable to coercion from a partner.

In sum, this study is novel in its aim to test models of sexual coercion that include an interplay of the specific traits and attitudinal factors described. We predicted that high levels of BPD traits, HPD traits, and rejection sensitivity along with a Ludus love style predict and explain engagement in sexually coercive behaviour in women. As a secondary objective, we tested whether these variables also predict sexual coercion victimisation. It is important to note that while there is both theoretical and empirical overlap between BPD and HPD (Reise & Wright, 1996), personality disorder traits and rejection sensitivity (Gardner et al., 2010; Lyddon & Sherry, 2001; Meyer & Pilkonis, 2005; Sato et al., 2020), and love styles and “disordered” personality traits (Jonason et al., 2020), each construct is distinct and may uniquely contribute to sexual coercion. Studies that utilise broader traits such as The Big Five (Davies, 1996) have already established a relationship between love styles and personality, including specific disordered personality traits. Yet the latter are conceptually relevant to sexual coercion. While both BPD and HPD include traits such as sexual impulsivity/promiscuity, the former also involves traits such as rapid shifts between idealisation and devaluation and the latter, excessive attention seeking. A unique contribution of rejection sensitivity can occur when conceptualised and operationalised as a separate construct i.e. rejection sensitivity is comprehensively modelled, rather than being assessed directly as related aspects of personality disorder.

Methods

Participants

A prior power analysis for multiple regression to detect a medium effect size with power set at .80 and p < .05 produced a total sample size of 113. This was a questionnaire-based study with participants completing the study either online or in paper form. Women (N = 151) aged 18–63 years (M = 23.34, SD = 8.80) were recruited via opportunity sampling at a British University and through social media (e.g. Facebook). Participants were typically students (75.5%), or in full time (14.6%) or part time (8.6%) employment. Relatively few participants were unemployed, homemakers, or retired (1.4%). Participants were single (31.8%), in the early stages of a relationship (10.6%), in a moderate-term relationship (33.1%), in a long-term relationship (17.2%), or married (7.35%). Participants were not typically cohabiting (72.8%) at the time of the study.

Measures

The Postrefusal Sexual Persistence Scale (Struckman-Johnson et al., 2003) is 19 item measure of post-refusal sexual persistence, defined as pursuing sexual contact with a partner after they initially refused (Struckman-Johnson et al., 2003). The original scale is separated into four sections, and the following three were used in this study in relation to both coercing others and their experience of being coerced: (1) nonverbal sexual arousal tactics (three items); (2) emotional manipulation and deception strategies (eight items); (3) exploitation of the intoxicated (two items). Example items include “persistent kissing and touching” and “taking off your clothes” (nonverbal sexual arousal), “telling lies” and “questioning their sexuality” (emotional manipulation and deception), and “purposely getting them drunk” and “taking advantage while they were intoxicated” (exploitation of the intoxicated). Participants report perpetration and victimisation since the age of 16 years old and items are scored 1 (yes) or 0 (no) and summed to create a total for each subscale. Each subscale demonstrated acceptable internal consistency (nonverbal sexual arousal perpetration: α = .79; nonverbal sexual arousal victimisation: α = .74; emotional manipulation perpetration: α = .82; emotional manipulation victimisation: α = .70; exploitation of the intoxicated victimisation: α = .71), except exploitation of the intoxicated perpetration (α = .42) which may reflect the short (i.e. two item) nature of the subscale.

The Borderline Personality Disorder – BPD (9 items) and Histrionic Personality Disorder – HPD (8 items) subscales were included from the Personality Diagnostic Questionnaire-4 (Hyler, 1994). The questionnaire was developed from the DSM-IV and DSM-V to screen for the presence of personality disorder traits. Participants respond to a series of questions by circling either “true” (scored 1) or “false” (scored 0) in relation to how they have tended to feel, act, and think over the past several years. Items are summed to create a total for each subscale, with a higher score indicating a higher level of traits associated with borderline and histrionic personality. Example items include “I often wonder who I really am” (BPD) and “I use my looks to get the attention that I need” (HPD). Cronbach alphas for the present study were: BPD: α = .66 and HPD: α = .59.

The Love Attitudes Scale: Short Form (Hendrick et al., 1998) is an 18-item version of the original Hendrick and Hendrick questionnaire. Six subscales each represent a different love style: Eros (passionate love); Ludus (game-playing love); Storge (friendship love); Pragma (practical love); Mania (possessive, dependent love); and Agape (altruistic love). Example items include “Our love is the best kind because it grew out of a long friendship” (Storge) and “I would rather suffer myself than let my partner suffer” (Agape). Participants responded to each item (in relation to their current or most recent relationship) on a five-point scale which ranged from 1 (strongly agree) to 5 (strongly disagree). The scores were added together for each subscale and were scored continuously. Each subscale displayed adequate internal consistency: Eros: α = .73; Ludus: α = .71; Storge: α = .89; Pragma: α = .75; Mania: α = .76; and Agape: α = .75.

The Adult Rejection Sensitivity Questionnaire (Berenson et al., 2009) contains nine hypothetical scenarios. Participants indicate their level of concern and level of expectancy in response to each scenario on a six-point scale from 1 (very unconcerned or very unlikely) to 6 (very concerned or very likely). The total concern and total expectancy scores are then multiplied together and then divided by nine to create a total rejection sensitivity score. Internal consistency was acceptable (α = .74) for the present study.

Procedure

Women, recruited from social networking sites or a British University participation point scheme, were asked to complete a series of standardised questionnaires. These were completed in the order outlined in the measures section. Participants completed the questionnaires either in hard-copy or online. The study was granted ethical approval from the University Ethics Committee and was conducted in accordance with The Code of Ethics of the World Medical Association (Declaration of Helsinki). All participants provided informed consent prior to participation, were debriefed after contributing, and did not receive financial reward for their participation.

Results

Subscale scores were calculated prior to analysis using SPSS (version 24). Analyses indicate that 42.4%, 19.2%, and 4.0% of women report some degree (i.e. indicating “yes” to at least one item) of nonverbal sexual arousal, emotional manipulation and deception, and exploitation of the intoxicated perpetration respectively. The proportion of women who had been victim to some level (i.e. indicating “yes” to at least one item) of nonverbal sexual arousal, emotional manipulation and deception, and exploitation of the intoxicated were 74.8%, 71.5%, and 39.1%. These data were, of course, recorded (and treated in subsequent analyses) continuously reflecting the degree of perpetration or victimisation reported by women. Descriptive statistics for all variables are shown in Table 1. There was no missing data. Correlation analyses revealed a number of significant relationships between predictor and criterion variables but no evidence of multicollinearity between predictors. These data are also shown in Table 1.

Table 1. Descriptive statistics and correlations between personality traits (Borderline and Histrionic), love styles (Eros, Ludus, Storge, Pragma, Mania, and Agape), rejection sensitivity, and sexual coercion (nonverbal sexual arousal, emotional manipulation and deception, and exploitation of the intoxicated).

	BP	HP	ERO	LUD	STO	PRA	MAN	AGA	REJ	NVP	EMP	EXP	NVV	EMV	EXV
BP
HP	.37**
ERO	.01	.05
LUD	−.15	−.19*	−.25**
STO	−.00	−.02	.25**	−.15
PRA	−.12	−.00	.07	.10	.01
MAN	−.47**	−.35**	.04	.20*	.02	.19*
AGA	−.39**	−.28**	.21*	.01	.03	.07	.37**
REJ	.21**	.06	−.13	−.08	−.08	.01	−.18*	−.23**
NVP	.14	.14	−.07	−.09	−.03	−.12	−.20*	−.04	.15
EMP	.26**	.09	−.07	.00	.04	−.01	−.19*	−.12	.24**	.34**
EXP	.01	.08	.04	−.26**	.02	.05	.02	.08	.05	.10	.11
NVV	.32**	.26**	.11	−.21	.05	−.19*	−.28**	−.04	.13	.48**	.25**	.11
EMV	.30**	.09	.07	−.21	.05	.01	−.11	−.19*	.08	.00	.31**	.17*	.39**
EXV	.16*	.01	.06	−.11	−10	.05	−.10	−.08	.14	.00	.12	.17*	.30**	.51**
M	3.19	2.57	6.15	11.81	9.40	9.94	10.70	7.15	16.37	.85	.36	.05	1.60	1.72	.58
SD	2.12	1.81	2.63	2.79	4.00	3.29	3.22	2.71	5.73	1.11	1.09	.24	1.17	1.68	.80

BP = Borderline Personality, HP = Histrionic Personality, ERO = Eros, LUD = Ludus, STO = Storge, PRA = Pragma, MAN = Mania, AGA = Agape, REJ = Rejection Sensitivity, NVP = Nonverbal Sexual Arousal Perpetration, EMP = Emotional Manipulation and Deception Perpetration, EXP = Exploitation of the Intoxicated Perpetration, NVV = Nonverbal Sexual Arousal Victimisation, EMV = Emotional Manipulation and Deception Victimisation, EXV = Exploitation of the Intoxicated Victimisation

*p < .05, **p < .01.

A series of multiple linear regressions were conducted using SPSS (version 24) to determine whether personality disorder (borderline and histrionic), love styles (Eros, Ludus, Storge, Pragma, Mania, and Agape), and rejection sensitivity predict sexual coercion (nonverbal sexual arousal, emotional manipulation and deception, and exploitation of the intoxicated) perpetration and victimisation. These data are shown in Table 2.

Table 2. Predictors of sexual coercion perpetration and victimisation.

	Perpetration			Victimisation
	Nonverbal sexual arousal	Emotional manipulation and deception	Experience of exploitation when intoxicated	Nonverbal sexual arousal	Emotional manipulation and deception	Experience of exploitation when intoxicated
	B	B	B	B	B	B
Borderline personality	.04	.21*	.01	.22*	.28***	.14
Histrionic personality	.10	.00	.07	.16	−.06	−.10
Rejection sensitivity	.12	.20*	.05	.09	.00	.09
Love styles
Eros	−.08	−.06	−.05	.06	.04	.80
Ludus	−.05	.06	−.27	−.09	−.18*	−.10
Storge	.00	.08	−.01	.03	.01	−.13
Pragma	−.08	.03	.06	−.14	.05	.08
Mania	−.14	−.09	.06	−.12	.07	−.03
Agape	.11	.05	.10	.15	−.13	−.04
R^2	.08	.12	.09	.20	.14	.07
F	1.36	2.17*	1.49	3.99***	2.52**	1.25

Together, personality disorder, love styles, and rejection sensitivity significantly predicted emotional manipulation and deception victimisation (F (91.41) = 2.52, p = .010) and perpetration (F (91.41) = 2.17, p = .028). Borderline personality traits was a significant positive predictor in both models, such that women high on Borderline personality traits were more likely to report being the victim (B = .28, t = 2.96, p = .00) and perpetrator (B = .21, t = 2.21, p = .03) of sexual coercion involving emotional manipulation and deception. For the perpetration model however, rejection sensitivity (B = .20, t = 2.35, p = .02) was also a significant positive predictor, whilst for victimisation the Ludus love style (B = −.18, t = −2.17, p = .03) was a significant negative predictor. Hence, women high on rejection sensitivity were more likely to use emotionally manipulative and deceptive forms of coercion against their partner whilst those high on the Ludus love style were less likely to be the victim of emotionally manipulative and deceptive coercion.

The only other significant model was nonverbal sexual arousal for victimisation (F (91.41) = 3.99, p < .001). Borderline personality traits (B = .22, t = 2.40, p = .02) represent the only significant individual predictor of nonverbal sexual arousal victimisation, such that women high on BPD traits were more likely to report being a victim of this form of coercion. Personality disorder traits, love styles, and rejection sensitivity did not significantly predict nonverbal sexual arousal perpetration or coercion involving exploitation of the intoxicated as either a victim or perpetrator.

Discussion

This study makes a novel contribution to the existing literature by investigating the influence of personality disorder traits (BPD and HPD), love styles (Eros, Ludus, Storge, Pragma, Mania, and Agape), and rejection sensitivity on women's sexual coercion – both perpetration and victimisation. With regards to the three models of perpetration, one model was significant while for victimisation, two of the three models were significant, suggesting that these constructs may increase the risk of perpetrating only specific types of sexual coercion. Findings indicate that women with higher levels of BPD traits are more likely to perpetrate sexual coercion characterised by both emotional manipulation and deception, and be the victim of emotional manipulation, deception and nonverbal sexual arousal. The latter is consistent with previous literature, suggesting that BPD increases the risk of sexual victimisation (Zanarini et al., 2005), albeit specific aspects thereof. This could reflect fears of abandonment which subsequently leads to engagement in unwanted sex. Use of emotional manipulation and deception might also be driven by specific BPD features such as abandonment fears and difficulty managing negative affect, but our study examined only global BPD and thus further research is necessary to identify the relative explanatory power of specific BPD criteria.

It is not clear why BPD failed to significantly predict all types of sexual victimisation and perpetration in our study. One possibility is that our sample did not capture a large enough proportion of individuals with specific BPD criteria (e.g. impulsivity) that may better explain/predict specific aspects of sexual coercion, consistent with the heterogeneous and polythetic nature of BPD. Similar arguments could be made for HPD. That is, our sample did not include sufficiently high levels of HPD criteria that overlap conceptually with sexual coercion e.g. sexual promiscuity. As HPD was not significant in any model, this explanation seems plausible.

BPD traits was not the only variable to explain variance in victimisation and perpetration of sexual coercion. Those with higher levels of rejection sensitivity were more likely to perpetrate sexual coercion involving emotional manipulation and deception. Results are consistent with literature suggesting that higher levels of hostility, aggression and violence are present when feelings of rejection arise (Ayduk et al., 1999, 2008; Romero-Canyas et al., 2010; Volz & Kerig, 2010). Future research should consider precursors of sexual coercion, such as relationship insecurity or instability. Women may display more emotionally manipulative tactics in an attempt to reduce the risk of sexual abandonment (Schachner & Shaver, 2004) and reduce feelings of rejection and it is important to consider the manner in which perceptions of relationship stability may change post coercion.

Finally, love styles did not improve the predictive power of all models of sexual coercion, but women displaying the Ludus love style were less likely to be the victim of emotional manipulation and deception based sexual coercion. Women with a ludic love style may be more aware of their partner's manipulative and coercive behaviour (due to their “game-playing” love style) or may be less likely to be as emotionally involved in their romantic and sexual relationships. We did not find that women adopting the Ludus love style were more likely to perpetrate sexual coercion; this was not consistent with previous research that linked Ludus love styles to sexual coercion in men and women (Kalichman et al., 1994; Russell & Oswald, 2001). Future studies may further consider the influence of Ludus on willingness to engage in sexual coercion and perceived consequences of such behaviour.

These results have both theoretical and practical implications. Our findings improve the limited understanding of factors that predict both sexual coercion and victimisation in women. Traditional theories of sexual coercion, such as feminist theory (Brownmiller, 1986), emphasise men's dominance and power over women, and while we did not assess prevalence of sexual coercion in women, our data show that sexual coercion is present in females. These findings challenge the myth that females do not engage in sexual aggressive conduct (for discussion, see Struckman-Johnson et al., 2020; Weare, 2018), and our results specifically indicate the value of interpreting sexual coercion within the context of enduring personality disorder traits and attitudinal factors.

Limitations and future directions

Future studies should determine whether the pattern of relationships reported in this study exist in clinical samples of women diagnosed with BPD and HPD who experience more extreme dysfunction. This study should also be replicated with wider and mixed community populations that reflect diversity in relation to cultural and ethnic background, as well as minority sexual orientation (Coulter et al., 2017; Ray et al., 2021). We suggest this not for blinkered comparative analysis but for encouraging inclusivity and broadening our understanding of sexually coercive conduct and experiences beyond samples of European-heritage, heterosexual women. As the women in this study were relatively young, future studies would also benefit from recruiting a better representation of older women. This may be particularly salient considering the higher rates of sexual aggression reported by younger Millennial generation women compared to older women representative of two past generations, considered to reflect a change in traditional sexual scripts (Anderson et al., 2021).

Findings are of course limited by a reliance on self-report questionnaire measures which means that common method variance may have artificially inflated relationships; moreover, self-report studies may be influenced by social desirability (e.g. reluctance to disclose the use of coercion) and poor recall (Johnson & Fendrich, 2005), and our results are not therefore generalisable to all cases. Yet, this methodology is consistent with research in the area and self-report measures of BPD and HPD are reliable and valid assessments of personality disorder symptoms, especially experiential ones (Hopwood et al., 2008). There is also scope for future studies to adopt a mixed method approach for a meaningful qualitative exploration of sexual coercion in women with BPD and HPD, providing a clearer insight into sexually coercive behaviours, lived experiences of victimisation and perceptions of their abusive conduct (O’Sullivan et al., 1998). This approach would also indicate important antecedent factors worthy of further, more holistic exploration. For example, this might include the role of trauma histories in childhood and adulthood as well as alcohol and substance use – factors that have been found to play a significant role in the use of sexual coercion in non-incarcerated, community-based females (Ecott et al., 2020) as well as the sex offences of convicted females (Fazel et al., 2010; Levenson et al., 2015; Wijkman et al., 2010).

Second, this study focused on exposure to sexual coercion (as a perpetrator or victim) rather than the frequency with which such behaviour has been perpetrated or experienced as a victim. Differences may occur between those experiencing one coercive event and those with repeated exposure (Zinzow & Thompson, 2015). Therefore, it would be prudent for future studies to consider repeated incidence of sexual coercion. In addition, we did not investigate responses to coercive behaviour or perceptions of coercive behaviour. Future research should establish factors predicting whether coercive behaviour results in sexual activity together or related consequences such as distress or relationship dissolution (Collibee & Furman, 2014). Perpetrator and victim gender have been found to influence perceptions of sexually coercive behaviour (Huitema & Vanwesenbeeck, 2016; Judson et al., 2013) and future research should consider responses to women's disclosure of sexual coercion perpetration or victimisation. Finally, previous research has established the effectiveness of interventions designed to reduce women's exposure to sexual violence (e.g. Senn et al., 2015). Future research should evaluate interventions targeted at both female victims and perpetrators to inform practice.

Conclusions

Our findings indicate that women with higher levels BPD traits are more likely to perpetrate sexual coercion characterised by emotional manipulation and deception, as well as be victim of sexual coercion involving nonverbal sexual arousal. Women with higher levels of the Ludus love style were less likely to be the victim of emotional manipulation and deception based sexual coercion. Finally, those with higher levels of rejection sensitivity were more likely to perpetrate sexual coercion involving emotional manipulation and deception. As not all models or predictors were significant, our findings are consistent with previous studies that highlight the importance of operationalising perpetration and victimisation of sexual coercion as multidimensional constructs because there are different correlates that may predict and explain each type.

Disclosure statement

No potential conflict of interest was reported by the author(s).