Evaluating perceptions of sexual coercion: the role of personality, gender, and motive in birth control sabotage

ABSTRACT Public attitudes are considered influential in terms of determining criminal justice responses to offending behaviour, however, research into sexual coercion and specifically Birth Control Sabotage (BCS) has received little attention. The aim of this study was to explore the influence of dark triad traits, gender, and motive on perceptions of BCS. Participants (N = 273) were recruited from a general population sample. All participants completed the Short Dark Triad (SD3) and read four vignettes relating to BCS, where perpetrator gender and function of sabotage (motive) were manipulated. Participants responded to these vignettes on a scale examining victim blame, criminality and victim impact. The results are discussed with reference to previous research exploring victim blame in other aspects of non-consensual sexual behaviour. As one of the first studies in this area, possible real-world implications and future directions are discussed in terms of jury decision making and victim support. PRACTICE IMPACT STATEMENT This article assists professionals in developing educational strategies and policies for how police and the legal system approach birth control sabotage and reproductive autonomy by explaining how attributing responsibility to victims and perpetrators and the need for police intervention is related to gender, intent and perceptions of affirmative consent.


Introduction
Prior research has argued that Birth Control Sabotage (BCS) transforms consensual sex to non-consensual sex (Brodsky, 2016;Clough, 2018;Rowlands & Walker, 2019), which may be considered a sexual offence. However, the law in relation to BCS is complex and BCS has also been considered within the framework of Intimate Partner Violence (IPV). Thus attitudes towards BCS may have important implications for influencing policy and the law as well as potential jury decision making. Given the dearth of studies addressing individual differences in attitudes towards BCS, the present study aimed to explore factors which may influence these. Specifically, the study focussed on the impact of gender and personality on attributions of blame in relation to BCS. In addition, the function of BCS was also explored in terms of whether this related to impregnation or heightened sexual gratification. The paper will begin by outlining the background literature pertaining to BCS and legal In this context BCS is deemed as consisting of pregnancy pressure (behaviours to coerce pregnancy), hiding oral contraceptive pills; removal of vaginal rings, contraceptive patches or intrauterine devices (IUD's) without a partners permission, removing or breaking condoms, or not withdrawing when agreed (Park et al., 2016).
Thus, Reproductive Control has recently been recognised as a form of domestic abuse (Rowlands & Walker, 2019), as it has a strong association with IPV (Katz et al., 2017;Miller et al., 2010). For example, Miller et al. (2010) found that 7% of women have experienced reproductive coercion, but this increases to 35% in the presence of IPV. In particular, women experiencing IPV are twice as likely to have a male partner who refuses to use contraception (Silverman & Raj, 2014) and are also twice as likely to report an unintended pregnancy . Thus, whilst reproductive coercion is not limited to women experiencing IPV, IPV does appear to increase the likelihood of experiencing reproductive coercion. This could be because BCS is not always sexually motivated but is often an expression of power by asserting control and dominance (Goode, 1971).
However, in order for BCS to be classified as an offence under the Serious Crimes Act (2015) there must be evidence that exposure to the behaviour has been "repeated" or "continuous". Thus, single instances of BCS would not fall under this category. Hence, in UK law BCS may also be considered within the realms of the Sexual Offences Act, 2003 under the notion of conditional consent which may apply to the offences of: rape; assault by penetration; sexual assault; and causing a person to engage in sexual activity. In relation to these offences Person A is guilty of an offence if he/she: 1) acts intentionally; 2) Person B does not consent to the act; and 3) Person A does not reasonably believe Person B consents. Section 74 defines consent as "if he agrees by choice, and has the freedom and capacity to make that choice" (p45). According to the CPS Guidance (Chapter 6: Consent), prosecutors should consider this in two stages: Stage 1) whether a complainant had the capacity (i.e. the age and understanding) to make a choice about whether or not to take part in the sexual activity at the time in question.
Stage 2) whether he or she was in a position to make that choice freely, and was not constrained in any way. Assuming that the complainant had both the freedom and capacity to consent, the crucial question is whether the complainant agrees to the activity by choice.
Thus, BCS has been regarded as relating to stage 2 in terms of whether the person consented to the activity by choice. Section 74 was subsequently considered by the High Court and the Court of Appeal in a series of cases where consent in relation to sexual offences was considered not to be true consent, either because a condition upon which consent was given was not complied with or because of a material deception (other than one which falls within section 76 of the Sexual Offences Act, 2003 [SOA]). The resultant judgments identified sets of circumstances in which consent to sexual activity might be vitiated where the condition was breached. The first consisted of Non-Consensual Condom Removal (NCCR) whereby it was concluded that "having sexual intercourse without a condom in circumstances where she had made clear she would only have sexual intercourse if he used a condom would therefore amount to an offence under the Sexual Offences Act 2003 … ." (CPS, Rape and Sexual Offences: Chapter 3: Consent). The other related to a case where the individual agreed to withdraw before ejaculation but he deliberately ejaculated inside the complainant, the result, the Court stated was: "She was deprived of choice relating to the crucial feature on which her original consent to sexual intercourse was based. Accordingly her consent was negated. .. In law, this combination of circumstances falls within the statutory definition of rape".
Therefore, sexual coercion via BCS may be considered a sexual offence (Brodsky, 2016;Clough, 2018), as stealthing blurs the boundaries between consent and violation, transforming consensual sex to non-consensual sex (Brodsky, 2016;Ebrahim, 2019). NCCR increases the risk of bodily harm, as NCCR can lead to unwanted pregnancies and sexually transmitted infections (Brodsky, 2016). However, this is further complicated in UK law because deception that vitiates consent must also demonstrate that that the nature of the deception is closely linked to the purpose of sexual intercourse. Hence, the CPS guidance notes that "A lie about wearing a condom is sufficiently closely connected because it physically changes the nature of penetration. In contrast, a lie about fertility is not, because it is not related to the performance of the sexual act" [R v Lawrance, 2020].
Thus, it would seem that aspects BCS may be considered under the framework of the Sexual Offences Act and/or the framework of Coercive and Controlling Violence. However, it is noteworthy that these two pieces of legislation discuss the application of the law within the context of females as victims and males as perpetrators. Additionally, research has focused on males as perpetrators of reproductive coercion, and females as victims (Hughes et al., 2019;Rowlands & Walker, 2019). Whilst research supports that men are more likely to perpetrate BCS (Davis et al., 2019;Krahé et al., 2015), there is growing evidence that women also engage in sexually coercive behaviours (Hines, 2007;Struckman-Johnson et al., 2003;Wegner et al., 2018). In addition, in the same way that deceptive condom removal has been regarded as an act of sexual coercion, deception in relation to contraception may also be adopted by females. For example, lying to a partner about contraception use (such as the pill) would be regarded as an act of BCS. This is sometimes referred to as "spurgling".
Few academic studies have been undertaken to reflect these trends (Brennan, 2017), especially when considering females as perpetrators. However, in a single case in the USA [P.P vs D.D (2017)], the male plaintiff consented to sex with the female defendant several times without using a condom, with the belief that the defendant was using the birth control pill. It was later found that D.D was pregnant with P.P's child, and the plaintiff alleged that his consent was vitiated. In this case, the Judge rejected the claims that P.P's consent was vitiated, concluding that whilst there was emotional harm resulting from the shock of unplanned parenthood, there was no physical harm caused to P.P from the sexual act itself. Physical harm includes pregnancy and sexually transmitted infections (Brodsky, 2016), and the Judge noted both were known risks before consenting to sex. Thus, the law in relation to BCS may vary dependent on the gender of the perpetrator and victim and the method of BCS. In addition, interpretations of the law may also vary and individual differences in attitudes could have important implications in situations such as jury decision making. However, previous research exploring attitudes towards sexual coercion has primarily investigated non-consensual sexual behaviour, typically rape (Diamond-Welch et al., 2017).

Individual differences
There is a lack of research exploring attitudes towards specific acts of sexual coercion such as BCS.
However, it has been demonstrated that many factors, such as gender, can influence attitudes held towards non-consensual sexual behaviour (Burt, 1980;Van der Bruggen & Grubb, 2014). Findings have indicated that men tend to: attribute less blame to perpetrators and more blame to victims compared to women (Diamond-Welch et al., 2017;Gerber et al., 2004;Van der Bruggen & Grubb, 2014); and hold less favourable attitudes towards rape victims than women (Jimenez & Abreu, 2003;Nagel et al., 2005;Ward, 1988). Gunby et al. (2012) suggest this could be due to gendered perceptions of what constitutes legal consent.
Gender can also be influential when explored as a victim or perpetrator characteristic, whereby studies have indicated that male victims are blamed more than female victims for their rape experience (Burczyk & Standing, 1989;Howard, 1984). It has been suggested that this is because male victims have been subject to societal stereotypes that expect them to be able to fight or escape the perpetrator (Howard, 1984). In contrast, other studies have indicated that male victims are perceived as less to blame than female victims (Kassing et al., 2005;Wakelin & Long, 2003) as they are less likely to be perceived as attracting a rapist and male rape myths are not as widely accepted (Kassing et al., 2005;Wakelin & Long, 2003). There has been less research considering gender as a perpetrator characteristic, but it has been demonstrated that both male and female perpetrators of rape receive more blame when the victim is female (Gerber et al., 2004;Rye et al., 2006). Theories used to explain gender differences in attributions of blame in non-consensual sexual behaviour relate to the sex role socialisation analysis of rape. For example, Burt (1980) suggested that men are socialised to be initiators of sexual interactions and are expected to gain sexual experience, whereas women are socialised to be passive, and are expected to remain chaste. Traditional gender roles are predictors of rape myth acceptance, and this may lead to more blame being attributed to victims of rape (for a review see Grubb & Turner, 2012). Another theory that may account for gender differences in blame attributions is ambivalent sexism which comprises of two components: hostile sexism and benevolent sexism (Glick & Fiske, 1996). Hostile sexism is where an individual has overt, negative beliefs or stereotypes about a gender. Benevolent sexism is where an individual has sexist prescriptions of stereotypical gender roles, but these are subjectively positive. An example of hostile sexism would be that women are incompetent and less intelligent than men, and an example of benevolent sexism would be that men should protect women (Glick & Fiske, 1996). Benevolent sexism has previously been associated with lower attributions of blame towards perpetrators (Chapleau et al., 2007;Viki et al., 2004;Yamawaki, 2007). It should be noted however, that studies examining sex roles and sexism are over a decade old and it is therefore uncertain whether they reflect current attitudes towards sex and gender.
Another factor that may influence an individual's attitudes and blame towards non-consensual sexual behaviour is personality (Burt, 1980;Grubb & Turner, 2012), and particularly for traits in the Dark Triad (DT). The dark triad consist of three socially undesirable personality traits; psychopathy, narcissism, and Machiavellianism (Paulhus & Williams, 2002). Psychopathy has been characterised by callousness, whereby an individual is not remorseful of their actions. Psychopathic traits can also include superficial charm, irresponsibility, fearlessness, and lack of empathy (Hare, 1991;Jonason et al., 2013). Narcissism gives individuals an inflated sense of self and is associated with grandiosity, whereby they believe they are superior to others (Jakobwitz & Egan, 2006;Paulhus & Williams, 2002). Machiavellianism is considered a manipulative personality, characterised by arrogant and immoral beliefs whereby they will do anything to get what they want (Jakobwitz & Egan, 2006). It should be noted that whilst these traits are distinct, they can overlap. All three traits have similar characteristics such as egocentricity and manipulation, those scoring high in all traits may use dishonest means to get what they want. Furthermore, individuals with DT traits have been shown to gain benefits through lying and deception (Baughman et al., 2014). Given that sexual coercion involves deception, it is not surprising that DT traits have been linked to the perpetration of sexual coercion in both men (Camilleri et al., 2009;Knight & Guay, 2006) and women (Blinkhorn et al., 2015;Ryan et al., 2008).
If those with higher DT traits are more likely to perpetrate sexual coercion, it can be reasonably assumed that people with these greater levels of these traits may also hold fewer negative attitudes towards the act of BCS. There is little research in this area, but DT traits have individually been associated with less favourable attitudes towards other non-consensual behaviour. For example, it has been found that men who score higher in narcissism have less empathy for victims of rape and endorse more rape myths (Bushman et al., 2003). More psychopathic individuals have also been found to hold more positive attitudes towards sexually coercive tactics (Marcus & Norris, 2014). Previously, these traits have been assessed using separate measures, typically the Self-Report Psychopathy Scale, the Machiavellian Scale, and the Narcissistic Personality Inventory (Furnham et al., 2013).

Aims and rationale
Thus, based on the aforementioned research the following hypotheses were generated: 1) There will be a positive correlation between all three DT traits and high scores in victim blame. 2) Regardless of the function of the behaviour, total victim blame scores will be significantly higher when the perpetrator is male, as will criminality and impact scores. 3) There will be no significant differences in total victim blame scores regardless of the function of BCS. 4) Significantly higher scores for criminality and impact on the victim will be attributed when the function is pregnancy compared to sexual gratification. 5) Female participants will attribute significantly lower blame to victims and will apportion significantly higher criminality and impact scores than male participants.

Participants
An opportunity sampling method was used, whereby participants were recruited online via community pages on social media in the UK. The sample consisted of two hundred and seventy-six participants from the general population aged over the age of 18. However, three participants, of which 1 identified as "non-binary" and 2 preferred not to say, were excluded from analysis. They were excluded as unequal sample sizes for gender identity would make it difficult to make meaningful comparisons. The final sample consisted of two-hundred and seventy-three participants, 204 (74.70%) identified as female, and 69 (25.30%) identified as male. Participants were between the ages of 18 and 63 years old (M = 22.44 years, SD = 6.28 years).

Design
The design of the study was within-participants, as participants read all four vignettes. The independent variables were: DT scores as measured by the Short Dark Triad (SD3) (Jones & Paulhus, 2014); the gender of the participant (male vs female); the gender of the perpetrator (male vs female); and the function of BCS (pregnancy vs gratification). The dependent variable measuring victim blame, criminality and victim impact was captured using the adapted version of the victim blame scale (Bothamley & Tully, 2018). Participants had to complete all questions in the study in order to participate.

Personality
Participants completed the SD3 questionnaire (Jones & Paulhus, 2014) which contained 27-items pertaining statements such as "It's not wise to tell your secrets" and "I'll say anything to get what I want". Participants had to respond to these statements on a 5-point Likert scale (1 = disagree strongly, 5 = agree strongly). The reliability of a questionnaire is widely considered acceptable when Cronbach's α > .70 (Cortina, 1993), however, other studies have indicated that reliability is reasonable provided Cronbach's a > .60 (Taber, 2018;Van der Bruggen & Grubb, 2014). Narcissism had an acceptable reliability, Cronbach's α = .70, as did the machiavellianism subscale, Cronbach's α = .69. For all items in the psychopathy subscale, the reliability was low, Cronbach's α = .50. Question 6, "People who mess with me always regret it" had the lowest correlation with the psychopathy subscale r = -.55.and was therefore excluded from analysis increasing reliability (Cronbach's α = .69).

Victim blame
Participants read each of the four vignettes that were created. The scenarios remained similar, but the following aspects were manipulated: perpetrator gender and victim gender (the scenarios were heterosexual) and the function of BCS (pregnancy or sexual gratification). In all conditions the perpetrator purposely concealed the act of BCS from the victim (e.g. lied to the other about using contraception). Examples are presented below: "Alex and Sarah had been dating for some time when they decided to go out on a dinner date. They went to their favourite restaurant and had a good time together. It was romantic, and once they got home, they decided to have sexual intercourse. Sarah put a condom on Alex before they began having sex. However, during sex, Alex took the condom off without Sarah being aware. Alex did this as he wanted Sarah to get pregnant" "Emma and James had their first date at a cinema. They enjoyed their time together and decided to become a couple. After dating for a while, Emma and James decided it was the right time to have sexual intercourse. James asked Emma if she was on any contraception. Emma said yes, mstating she was on the pill. Emma lied, she was never on the contraceptive pill, but she decided not to tell James. This is because she wanted to get pregnant with his baby" Following each vignette, participants completed an adapted version of the victim blame scale (Bothamley & Tully, 2018) relating to the behaviour described. Participants responded on a 5point Likert scale (1 = none at all/very unlikely, 5 = a great deal/very likely) to a total of eight items. The first four items measured victim blame, these were "how much do you think (victim's name) is to blame for the incident?", "how likely do you think it is that (victim's name) could have avoided the incident?", "how much do you think (victim's name) had control over the situation?" and "how sorry do you feel for (victim's name)?", with the last being reverse coded. Victim blame was analysed in terms of a total score, where the minimum score is 4 and the maximum score is 20.
Higher scores indicate that more blame is being attributed to the victim. Scores for total victim blame were highly reliable (Cronbach's α = .84).
A further two items explored criminality, these were "to what extent do you consider (perpetrator's name) behaviour to be an offence?" and "do you think police intervention is necessary for the resolution of this situation?". Although these two questions both explore criminality, there were analysed separately following the design of Bothamley and Tully's (2018) study. The last two items, "do you think (perpetrator's name) behaviour will create fear or apprehension in (victim's name)?" and "do you think (perpetrator's name) behaviour will cause mental harm to (victim's name)?", explored the impact of the scenario on the victim. These were again analysed separately for the same reason given above. For all criminality and impact questions, the minimum score was 1 and the maximum score was 5. Higher scores suggest more favourable towards victims, whereby the criminality or victim impact from the behaviour is considered more likely. Scores for these were highly reliable, both for criminality, Cronbach's α = .87 and the impact on the victim (Cronbach's α = .85).

Procedure
Participants were recruited via social media, through sharing a link to the study on Qualtrics. Before taking part, participants read an information sheet outlining the aims and purpose of the study. This also informed participants that their participation was voluntary and completely anonymous. Participants were informed before the study that they could withdraw at any stage by exiting the browser, but that once they submitted their responses there was no way to withdraw, as no identifiable information was provided. They were then required to tick one of two boxes: "I give consent" or "I do not give consent". If the latter was selected, participants were automatically sent to the end of the study. Following consent, participants were first asked demographic information relating to their gender and age and were then required to answer the SD3 questionnaire. After the personality questionnaire, participants read four vignettes, the order of which was presented randomly. After reading each vignette, participants answered a victim blame scale. Upon completion, participants received a debrief sheet that detailed the nature of the study, and reminded them that their participation was completely anonymous, and could therefore not be withdrawn.

Ethical considerations
Ethical approval was gained from the University of Central Lancashire Ethics Committee. All participants provided informed consent, as although to avoid bias they were not told that the scenarios were examples of BCS they were told that they related to "risky sexual behaviour". Prior to taking part, participants were told that they could withdraw at any time during the study, up until submitting their responses. Once submitted, it was not possible to withdraw as responses were completely anonymous. No identifiable information was provided at any time, but all data was kept on a password protected computer. The nature of the study could be considered sensitive, and as such, participants were provided with the contact information for both Victim Support and the Survivors Trust should they require support.

Data screening
Prior to data analysis, data screening processes were conducted. The data screening identified that there was no missing data, and an examination of the boxplots revealed potential outliers, but none were extreme, and all values were within range. They were therefore not removed or transformed in accordance with recommended procedures (Tabachnick & Fidell, 2001). In addition, the current sample exceeded 200 participants and, therefore, normality tests were not conducted (Tabachnick & Fidell, 2001). This is because in larger samples, normality tests can be sensitive to significant results even if there is only a small deviation from normality (Field, 2005).

Personality Scale
Machiavellianism A Pearson's correlation coefficient was used to assess the relationship between machiavellianism and total victim blame. Machiavellianism was not significantly correlated with blame for all four conditions (see table 1). This indicates that those who have higher levels of Machiavellianism do not blame victims of BCS more or less than those have lower levels of Machiavellianism. No correlation was found between Machiavellianism, total victim blame, criminality or victim impact.

Psychopathy
A Pearson's correlation coefficient was used to assess the relationship between psychopathy and total victim blame. No significant correlation was found between psychopathy and victim blame in all four conditions (see table 2). This indicates that those who have higher levels psychopathy do not blame victims of BCS more or less than those have lower levels of psychopathy. No correlation was found between psychopathy, total victim blame, criminality or victim impact.

Narcissism
A Pearson's correlation coefficient was used to assess the relationship between narcissism and total victim blame. Narcissism was also not significantly correlated with blame in any of the BCS conditions (see table 3). This indicates that those who have higher levels narcissism do not blame victims of BCS more or less than those have lower levels of narcissism. No correlation was found between narcissism, total victim blame, criminality or victim impact.

Victim Blame Scale
Each victim blame subscale was subject to a 2 (Participant gender: male vs. female) x 2 (Perpetrator gender: male vs. female) x 2 (Function: sexual gratification vs. impregnation) mixed-design factorial ANOVA. The first factor, participant gender, was a between-subjects variable. The other two factors, perpetrator gender and function, were repeated measures, within-subject variables. All effects are reported as significant at p < .05.

Victim blame
There were two significant main effects found for total victim blame. Firstly, there was a significant main effect of perpetrator gender, F(1,271) = 52.95, MSE = 6.23, p < .01, η p 2 = .16. This confirms that victims are blamed more when the perpetrator of BCS was female (M = 6.95) compared to when the perpetrator was male (M = 5.69). The second significant main effect was found for the function, F (1,271) = 5.48, MSE = 3.58, p = .02, η p 2 = .02. This indicates that when the function of BCS is heightened gratification, victim blame is significantly higher (M = 6.47), than when the function is pregnancy (M = 6.16). However, there was no significant effect of participant gender, F(1,272) = 1.59, MSE = 3.33, p = .21, η p 2 = .01, indicating that males and females attribute approximately the same amount of blame to victims of BCS.

Criminality (Offence constitution)
Two main effects were found in relation to the extent to which BCS was viewed as an offence. Similarly to total victim blame, there was a significant main effect of perpetrator gender, F(1,271) = 85.84, MSE = .67, η p 2 = .24, indicating that BCS is more likely to be perceived as an offence when the perpetrator is male (M = 4.43) compared to female (M = 3.92). There was also a significant main effect of function, F(1,271) = 97.38, MSE = .56, p < .01, η p 2 = .26. This reflects that the behaviour is more likely to be viewed as an offence when the function of the behaviour is pregnancy (M = 4.44) compared to gratification (M = 3.92). The main effect of participant gender was not significant, F(1,271) = 2.08, MSE = .58, p = .15, η p 2 = .01, indicating that there are no differences in the extent to which males and female believe BCS is an offence.
There was a significant interaction effect found between perpetrator gender and function, F (1,271) = .11, MSE = .44, p = .74, η p 2 < .01. Contrasts revealed that regardless of perpetrator gender, BCS was perceived as more of an offence when the function was pregnancy (M = 4.43) rather than gratification (M = 3.92). Contrasts also revealed that BCS is viewed as more of an offence when the perpetrator was male (M = 4.44) compared to female (M = 3.91), regardless of function.

Criminality (Police intervention)
For the second criminality item, there was a significant main effect of perpetrator gender, F(1,271) = 119.96, MSE = .75, p < .01, η p 2 = .31 whereby police intervention was viewed as more necessary when the perpetrator was male (M = 3.35) than female (M = 2.69).There was also a significant main effect of function, F(1,271) = 122.78, MSE = .58, p < .01, η p 2 = .31, confirming that police intervention is perceived as more necessary when the function of BCS is pregnancy (M = 3.31) compared to gratification (M = 2.73). Again, there was no main effect of participant gender, F(1,271) = 1.13, MSE = 1.19, p = .29, η p 2 < .01. This indicates that males and females view police intervention as necessary to the same extent.
There was a significant interaction effect between perpetrator gender and function, F(1,271) = 19.41, MSE = .53, p < .01, η p 2 = .07. Contrasts reflect that when the perpetrator is male or female, police intervention is perceived more necessary when the function of BCS is pregnancy (M = 3.32) compared to gratification (M = 2.72). Contrasts further revealed the police intervention is deemed more necessary when the perpetrator is male (M = 3.35) compared to female (M = 2.69) regardless of function.

Victim impact (Fear or apprehension)
There was a significant main effect of perpetrator gender, F(1,271) = 50.85, MSE = .45, p < .01, η p 2 = .16 in whether participants believed BCS was likely to create fear or apprehension in the victim. This reflects that fear or apprehension is considered more likely when the perpetrator is male (M = 4.58) compared to female (M = 4.25). Another significant main effect was found for function, F (1,271) = 92.40, MSE = .32, p < .01, η p 2 = .25, whereby participants viewed victim fear or apprehension more likely when the function of BCS is pregnancy (M = 4.61) versus gratification (M = 4.22). There was, however, no significant main effect of participant gender, F(1,271) = 1.45, MSE = .29, p = .29, η p 2 < .01. When exploring interactions, there was a significant interaction effect between perpetrator gender and function, F(1,271) = 18.33, MSE = .30, p < .01, η p 2 = .06. Contrasts revealed that victim fear or apprehension was considered more likely when perpetrated by a male (M = 4.58) rather than female (M = 4.25), for both gratification and pregnancy. Contrast also indicated that fear or apprehension was considered more likely when the function was pregnancy (M = 4.61) compared to gratification (M = 4.22), regardless of perpetrator gender.

Victim impact (Mental harm)
The last item also found a significant main effect of perpetrator gender, F(1,271) = 77.79, MSE = .44, p < .01, η p 2 = .22, whereby mental harm was viewed as more likely when the perpetrator is male (M = 4.34) compared to female (M = 3.93). Similar to the other items, there was also a significant main effect of function, F(1,271) = 124.74, MSE = .52, p < .01, η p 2 = .32. This reflects that mental harm is perceived more likely to the victim when the function of BCS is pregnancy (M = 4.14) compared to gratification (M = 3.85). There was no main effect of participant gender, F(1,271) = 1.10, MSE = .47, p = .30, η p 2 < .01, highlighting that males and females view the likelihood of mental harm the same in BCS.
There was a significant interaction effect between perpetrator gender and function, F(1,271) = 17.14, MSE = .32, p < .01, η p 2 = .06., with contrasts revealing that mental harm is considered less likely when the function is gratification (M = 3.85) rather than pregnancy (M = 4.14), regardless of perpetrator gender. Contrasts further indicated that mental harm is considered more likely in both functions when the perpetrator is male (M = 4.34) compared to female (M = 3.93).

Discussion
The present study aimed to explore factors which may influence attitudes towards BCS. More specifically, it investigated the impact of gender and personality on attributions of blame in relation to two types of BCS and unlike other studies in the area, it explored two different functions of this BCS: impregnation or heightened sexual gratification.

Personality
Previous research has shown that DT traits are associated with unfavourable attitudes towards nonconsensual sexual behaviour (Bushman et al., 2003;Marcus & Norris, 2014.) However, the results of the present study did not support a significant association between DT traits and attitudes towards victim blame, criminality and impact on the victim for BCS. This was a surprising finding when considering the nature of DT traits and their previous association with the perpetration of sexual coercion (Blinkhorn et al., 2015;Camilleri et al., 2009;Knight & Guay, 2006;Ryan et al., 2008). Previously it has been hypothesised that the Dark Triad is linked to attitudes endorsing non-consensual sexual behaviour due to DT traits representing a lack of empathy (Jonason et al., 2013), egocentricity, a tendency to do anything to get what they want and a sense of superiority over others (Jakobwitz & Egan, 2006). DT traits have also been linked to promiscuous mating (Koladich & Atkinson, 2016) and a reproductive focus on multiple partners (Jonason et al., 2010) known as the fast life history strategy (Kavish & Anderson, 2019). Thus, it was expected that people scoring higher on DT traits would endorse more positive attitudes towards BCS, particularly when the function was sexual gratification. However, this was not supported in the current study. Whilst it could be argued that one reason for this is that previous research exploring DT traits and attitudes towards sexual coercion was outdated in the current social climate, evidence for a link between Dark Triad traits and sexual coercion has been found in a recent study (Lyons et al., 2020). Thus, changes in socio-political attitudes generally would not appear to account for the findings in the current study.
However, given the overlap between the sexual coercion and IPV literature in relation to BCS it is possible that the literature in relation to DT and IPV may be of more relevance. For example, Carton and Egan (2017) found that specific personality traits such as low agreeableness were better predictors of IPV than DT traits whereby Machiavellianism was not found to predict psychological or physical IPV abuse. The authors propose the reasons for this may be that due to the "socially manipulative" nature of Machiavellianism whereby they may be less likely to endorse expressions of IPV. Furthermore, psychopathy was associated with dominance/intimidation and denigration and Narcissism and Machiavellianism with restrictive engulfment. Carton and Egan (2017) also found that the main source of the association between psychopathy elements on the DT and IPV related to scores on agreeableness which is a more common dispositional trait. Thus, it is possible that factor such as agreeableness would be more predictive of variations in attitudes towards BCS rather that overall DT traits. Further evidence for this comes from the findings of Lyons et al. (2020) who noted that in their analysis of individuals with DT traits, sexual assertiveness (defined as strategies that individuals use to achieve sexual autonomy [Hurlbert, 1991]) was associated with lower perpetration of sexual coercion in males.
Hence, it is possible that only certain aspects of the personality structure associated with low agreeableness and low assertiveness predict attitudes towards sexual coercion and more specifically BCS. Shafer et al. (2018) suggested this may be because people who are sexually assertive have a greater understanding and appreciation of sexual consent and are less sensitive to rejection (Lyons et al., 2020). This would fit with theories of fragile narcissism whereby it is noted that confluence models of sexual violence suggest hostile masculinity is associated with hypersensitivity to rejection and criticism (Malamuth et al., 1995) known as the narcissistic reactance theory of sexual assault (Baumeister et al., 2002). Thus, narcissism and DT traits may be considered as multi-dimensional constructs measuring different aspects of personality whereby only certain aspects of these constructs act as better predictors for attitudes towards sexual coercion (and potentially BCS) rather than overall scores.
Thus, several reasons may explain why DT traits were not associated with attitudes towards BCS. Firstly, DT traits may not adequately assess the range of multi-dimensional factors linked with personality and attitudes towards sexual coercion. Secondly, it is possible that aspects of the DT traits associated with coercion (such as dominance/intimidation and denigration and restrictive engulfment) are simply not present within the context of BCS and this may explain the lack of association between DT and BCS. In addition, it is possible that attitudes towards BCS are related more to threats of being able to participate in the "fast life" in terms of potential unplanned pregnancy posing as threat as something which may inhibit this "fast life". Thus, if as Jonason et al. (2010) posit that DT traits are designed to facilitate social behaviour that matches a fast life, then the risk of becoming a parent as a result of BCS may seem unappealing. Thus, it is possible that the attitudes captured in this study are more a reflection of egocentric "self-views" of BCS posing a potential threat to the fast life and an inconvenience to those scoring high on DT traits rather than attitudes being a reflection of empathy for the psychological effects of BCS. Further evidence for this is linked to the notion that although DT traits are associated with impulsivity and sensation seeking (especially narcissism) this is moderated by negative feedback (Crysel et al., 2013) whereby the narcissism component of the dark triad is positively associated with temporal discounting (e.g. less now, more later). In addition, this is greater for people with higher scores of ego-threat narcissism. Thus, similar to the findings for sexual violence, people scoring high on DT traits are able to inhibit impulsivity and immediate rewards if this poses a threat to their reward system in the long term. Hence, it is possible that BCS and the threat of unplanned pregnancy imposing on a fast life may trigger temporal discounting for people with DT traits thus resulting in the current findings. Furthermore, it is also possible that participants who scored highly on Machiavellianism are also more prone to engage in socially desirable responding and self-monitoring and thus their self-report does not reflect true internal values (Kowalski et al., 2018). Thus, future research exploring individual differences in relation to attitudes towards BCS would benefit from exploring more subtle aspects of personality such as agreeableness, sexual assertiveness and could also include measures of self-deceptive enhancement and socially desirable responding.

Perpetrator gender
It was predicted that total victim blame scores would be significantly higher when the perpetrator of BCS was male and the victim female. This was because female victims are more likely to be victims of non-consensual sexual behaviour (Breiding et al., 2014;Smith et al., 2018), and are perceived as more likely to attract a rapist than male victims (Kassing et al., 2005). The results did not support this hypothesis and indeed there were no significant differences between how male and female participants perceive BCS as a whole. The lack of impact found regarding participant gender does not support an array of previous research that has typically found that men attribute more blame to the victims of non-consensual sexual behaviour (Diamond-Welch et al., 2017;Gerber et al., 2004;Van der Bruggen & Grubb, 2014). Thus, it is possible that attitudes towards BCS are less influenced by participant gender.
However, the gender of perpetrators of BCS was important in this study whereby participants were noted to place significantly more blame on male victims of BCS when the perpetrator was female compared to when the perpetrator was male and the victim was female. This may potentially provide support for sex role theories whereby it has been suggested that traditional gender roles may explain gender differences in victim blame (Burt, 1980;Grubb & Turner, 2012). Male victims may be blamed more than female victims because males have been subject to societal stereotypes that expect them to be able to fight or escape the perpetrator (Howard, 1984). For example, Ayala et al. (2018) found that more blame and less sympathy was assigned to male than female rape victims which Reitz-Krueger et al. (2017) apportion to rape myths for female and male victims being subjected to gendered nuances. Javaid (2015) described this in terms of female rape myths consisting of females being deserving of blame (e.g. 'asking for it) whereas male rape myths consisted of denial (e.g. real men cannot be raped). Mulder and Bohner (2020) propose that this is due to the need for people to re-establish familiar gender and victim scripts which perpetuate ambivalent sexism. Thus, it is possible that less blame was placed on female victims and more blame attributed to the male perpetrators because men have historically been perceived as protectors of women (Glick & Fiske, 1996) and male victimisation challenges cultural perceptions that "real men" cannot be victims (Cohen, 2014). Mulder and Bohner (2020) suggest this is due to what is known as binding values taken from Moral Foundations Theory (Graham et al., 2009) which have been associated with positively endorsing rape myths. Thus, the findings of this study in relation to BCS mirror those of Walfield (2018) who found that one third participants in their study stated they struggled to believe that a male could be raped by a female and that participants endorsed myths which downplayed the harm of rape, particularly when the offender was female. However, future research is needed to explore this given the dearth of studies in relation to attitudes towards female perpetrators of BCS and male victims. It is noteworthy the National Intimate Partner and Sexual Violence Survey (NISVS) in 2010 found an estimated 10.4% of the US male population have been victims of BCS, reporting that they had a female partner who either attempted to get pregnant when the male partner did not want to or the partner tried to prevent them using a condom. Thus, the issue is clearly prevalent but under-researched to date.
It was also hypothesised that participants would perceive BCS as a more criminal act when the perpetrator was male. Results provide support for this hypothesis, as participants were significantly more likely to perceive BCS as an offence when the perpetrator was male compared to female. In addition, participants were significantly more likely to perceive police intervention as a necessary when the perpetrator was male, regardless of the function of the behaviour. This is an interesting finding as it appears to reflect the criminal nature of societal views that women should have autonomy over their body (James-Hawkins et al., 2019) and that absence of this constitutes a criminal offence because of an absence of affirmative consent (Alani, 2017). Hence, attitudes towards victim blame, criminality and BCS appear to be influenced by two separate constructs, contraceptive responsibility and bodily autonomy. This was further supported by the finding that participants in this study also viewed the creation of fear or apprehension in the victim, as well as the likelihood of mental harm as significantly more likely when the perpetrator was male rather than female. Thus, it would seem that in terms of attitudes towards BCS and victim harm this is also placed within the framework of female bodily autonomy. These findings are consistent with research which shows that unintended pregnancy is a risk factor for maternal depression, stress, and lower levels of psychological well-being (Bahk et al., 2015).
It is also possible that more criminal responsibility on male perpetrators of BCS may also be explained by sexual script theory which posits that sexual scripts are socially constructed cognitive schema that define normative sexual behaviours and inform individual actions in sexual situations (Simon & Gagnon, 2003). At a cultural level, traditional sexual scripts are gendered prescriptions for appropriate sexual conduct (Masters et al., 2013). Thus, it is proposed that sexual behaviours are modified by societal gender inequality (Martin & Ruble, 2010) and that stereotypical views of male and female sexual behaviour still exist (Petersen & Hyde, 2010). Thus, it is possible that the more blame apportioned to male perpetrators in BCS is consistent with other studies which have found that female sexual offenders are ordinarily treated more leniently (Shields & Cochran, 2020) and their offences are viewed as less threatening and harmful (Mellor & Deering, 2010). Thus it is possible that the same views apply to females engaging in BCS as those engaging in sexual offending.

Function
It was predicted that there would be no significant difference in victim blame between BCS for gratification and BCS for pregnancy. In this study participants attributed higher scores for blame to victims of BCS when the behaviour was perpetrated for sexual gratification rather than pregnancy. Thus, lower scores for victims when the function was pregnancy could be indicative of a more sympathetic attitude towards victims of reproductive coercion where the intent was to reduce reproductive autonomy. This sympathetic view towards victims of reproductive coercion resulting in pregnancy mirrors the findings of Taylor and Shrive (2021).
It was also predicted that BCS would be viewed as a more criminal act when the function was pregnancy rather than gratification. Results from this study upheld this prediction and . participants also deemed police intervention as significantly more necessary when the function of BCS was pregnancy. Thus, it would seem that participants viewed the specific planned intent of the perpetrator to induce pregnancy as more criminal and worthy of punishment. This may be explained by attitudes in relation to impulsive versus instrumental acts, whereby the condition in which the motive of the offender was to cause pregnancy whilst purposefully deceiving their partner was deemed more punishable. In criminal law the "planning of an offence" is classified as an aggravating factor in the general guidelines of the Sentencing Council (2019). In terms of culpability of the offender, this is classified as "with reference to the offender's role, level of intention and/or premeditation and the extent and sophistication of planning". The degrees of culpability lie from the highest level (defined as deliberateintentional act or omission) moving downwards to a medium level (defined as recklessacted or failed to act regardless of the foreseeable risk). Thus, whilst both scenarios in the study could have led to pregnancy, participants appeared to place greater culpability on the perpetrator who deliberately intended pregnancy as the outcome which is in line with UK law.
Another hypothesis that was supported was that BCS function would influence participants perceptions of the impact that the behaviour had on the victim. Results revealed that the creation of victim fear or apprehension were considered significantly more likely consequences when the function of BCS was pregnancy rather than gratification. Similarly, participants viewed mental harm to the victim as significantly more likely when the function was pregnancy. One reason for the influence of function may be due to the perceived seriousness of the outcome. For example, in their study Taylor and Shrive (2021) found that public perceptions of attitudes towards women who became pregnant or had children conceived in sexual violence related to themes of sympathy, that they may be expected to want an abortion and if they did not then people may see this as a sign of the person not being honest about being raped. Thus, it may be considered that pregnancy is a more serious and potentially less acceptable consequence than those associated with increased pleasure because of the additional ethical dilemmas which would accompany the complexities of pregnancy through BCS. In addition, the notion of culpability and the purposefully planned breach of trust associated with the planned conditions of imposed pregnancy may have also contributed to the findings in this study whereby participants perceived planned reproductive coercion as more serious because it impacts on the autonomous decision-making with regards to reproduction (Moore et al., 2010). Thus, it would seem that participants considered the consequence of a planned deceit leading to unintended pregnancy as having a dual effect of both a breach of trust on the victim as well as a lack of autonomy. Again, it would seem this links to the sentence guidelines in terms of culpability and criminal responsibility as well as the "harm caused by the offending" in terms of whether the harm "was intended to cause or might foreseeably have caused". Thus, it would seem both the intent of inducing an unplanned pregnancy plus the role of active deception with the intentional misrepresentation with the intent to deceive the partner (Knapp & Comaden, 1979) led participants to conclude that BCS was more serious in the actively planned pregnancy condition rather than sexual gratification. Thus, future research in relation to BCS would benefit from exploring attitudes towards trust as well as intent and culpability.

Overview and Future Directions
The present findings contradict previous research that has found benevolent sexism to be associated with lower blame attributed to perpetrators (Chapleau et al., 2007;Viki et al., 2004;Yamawaki, 2007), as regardless of perpetrator gender, more blame was placed on perpetrators rather than victims.
The negative outcomes of BCS, in terms of fear or apprehension and mental harm, were considered more likely when perpetrated by males. It can be inferred that the public perceive stealthing as more serious than lying about being on other contraceptives such as the pill. However, it should be noted that whilst significant differences were found for function and perpetrator gender, levels of victim blame across all conditions remained low, and close to the minimum score. Favourable attitudes towards victims of BCS were also demonstrated in the victim impact items when ignoring perpetrator gender and function. Participants viewed fear or apprehension and mental harm as likely victim outcomes across all BCS conditions. It can therefore be argued that the general population understand the negative consequences that non-consensual sexual behaviour can have on individuals, and that this understanding goes beyond that of unwanted pregnancy and sexually transmitted infections (Brodsky, 2016).
Participants tended to believe that BCS was likely to constitute an offence. This is encouraging as it indicates participants are aware that the behaviour depicted in the vignettes can indeed constitute an offence. This is because the behaviour described vitiates consent (Brodsky, 2016;Ebrahim, 2019), the capacity of which is a defining characteristic of sexual violence (Bagwell-Gray et al., 2015). However, despite BCS being viewed as a likely offence, participants tended not to view police intervention as somewhat or very likely necessary. This runs parallel with sexual coercion findings in realworld contexts. For example, it has been demonstrated that only around 13% of women and 8% of men who have experienced sexual coercion reported the event to the police (Macdowall et al., 2013). Whilst this study supports that the prevalence of sexual coercion is under-reported in the UK (Macdowall et al., 2013), it does not support attitudes found in previous research, whereby victims do not see the experience as a crime (Fisher, 2009;Macdowall et al., 2013). Future research should investigate why there is this discrepancy, as it may have implications for policing and crime reporting and the influence of social media on public perceptions of crime.
One strength of the current study is that participants were gained from a general population sample. Any of the participants could be called for jury duty, it is therefore useful to gain their views and attitudes regarding sexually coercive and specifically BCS. This is because it may potentially help to understand their decision-making processes in such cases. Future research could gain more insight into this by investigating BCS using mock trial transcripts rather than the vignettes used in this study. There may be different findings when participants are presented with a more ecologically valid, real-world examples.

Limitations
There are some limitations to the design of the study. Firstly, individuals who did not identify as male or female, or who did not want to disclose their gender identity had to be excluded from analysis due to unequal sample sizes. This questions whether the findings of this study are as representative to the general population as intended. In addition, participants were recruited online via community pages on social media in the UK. However, participants were not asked to identify demographical information about their cultural or national identity therefore societal and cultural norms which may have influenced participant attitudes could also not be explored. Another limitation is that both of the questionnaires used (the SD3 and the victim blame scale) required self-report techniques. Self-report techniques have been critiqued as they can often be subject to social desirability bias. This could be particularly problematic in the SD3 as it is attempting to measure socially aversive traits, but also in victim blame measures as participants may respond less honestly to avoid appearing in a negative light. Furthermore, it would be useful for future research to focus on the wider types of harm caused by BCS in terms of the sabotage to sexual health more generally such as the other negative impacts of this type of BCS in the form of sexually transmitted infections.

Conclusion
In conclusion, the results of the present study suggest that perpetrator gender and the function of BCS have an impact on victim blaming and perceptions of both criminality and victim impact. Whilst previous studies investigating attitudes towards non-consensual sexual behaviour have focused on rape, the present study has highlighted the need for more research into perceptions of sexually coercive acts without the presence of force. To date, the current study appears to be the only one investigating victim blame in BCS that also considers the motivations behind the behaviour. The results of this study are important as they may have implications for jury decision making and policing. In addition, understanding attitudes towards non-consensual sexual behaviour may help practitioners in supporting victims and may also help in improving treatment interventions for offenders. It is recommended that future studies should investigate this topic further to develop stronger, more consistent findings in this area.

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