The most widely accepted theory to explain MPR is the Multi-Factorial Theory of Multiple-Perpetrator Sexual Offending, which proposes that MPR occurs due to an interaction of individual, sociocultural, and situational factors (Harkins & Dixon, 2010, 2013). Individual factors consist of personality traits, developmental factors, and sexual preferences, with leadership and deviant sexual interests theorized as being the most impactful. Sociocultural factors consist of cultural norms, myths, values, and beliefs, with rape culture, rape myths, and patriarchy being the most important in increasing the likelihood of MPR. Lastly, situational factors consist of the setting in which MPR may occur, such as pedophile organizations or war. The current study focused on individual and sociocultural factors.

In order to further understand MPR, Alleyne and colleagues (2014) developed a novel, vignette-based, proclivity measure called the Multiple-Perpetrator Rape Interest Scale (M-PRIS), which assesses sexual arousal to MPR, behavioral propensity towards MPR, and enjoyment of MPR. The M-PRIS has demonstrated strong psychometric properties, specifically in its high internal reliability and strong test-retest reliability. Prior research with the M-PRIS has demonstrated that certain constructs hold positive associations with a proclivity for multiple-perpetrator rape, such as deviant sexual interests (i.e., high-risk sexual fantasies), rape-supportive cognitions, and antisocial attitudes (Alleyne et al., 2014; Palermo et al., 2019) while others do not, such as a dominant personality (Petrella & Harkins, 2021). These results, particularly related to deviant sexual interests and rape-supportive cognitions lend support to the Multi-Factorial Theory of Multiple-Perpetrator Sexual Offending (Harkins & Dixon, 2010, 2013), which posits that deviant sexual interests and rape-supportive cognitions are key factors underpinning MPR offences. Thus, subscales of deviant sexual interests and rape-supportive cognitions should be further investigated in their relationship with MPR.

Few studies have examined aggression levels within MPR. Woodhams and Cooke (2013) examined the relationship between suspect aggression and victim resistance in MPR. They found that victim resistance was significant and the type of rape (simultaneously versus sequential) was positively associated with suspect aggression. Additionally, larger group size was significantly associated with less victim resistance. Moreover, Hauffe and Porter (2009) compared lone and group sexual assault characteristics and found that group sexual assaults were more likely than lone sexual assaults to be hostile and violent in nature (i.e., use of violence to restrain the victim and hold the victim down, use of multiple weapons), involve multiple rapes by the same offender, and vaginal penetration. Consistent with prior research on group size and MPR, the findings suggest that MPRs consist of a higher degree of aggression (Hauffe & Porter, 2009; Woodhams & Cooke, 2013; Woodhams et al., 2020). Thus, aggression should be further investigated in their relationship with MPR.

Based on prior literature on sexual violence among lone offenders, we decided to investigate additional potential correlates that may potentially be related to MPR. The correlates map onto dynamic risk factors, which are factors that are potentially changeable (Ward & Fortune, 2016). First, loneliness and lack of intimacy have been extensively researched as potential dynamic risk factors clinically relevant in sexual offenders (Seidman et al., 1994; Wielinga et al., 2021). Marshall (2010) proposed that an interaction between poor childhood attachment style, lack of intimacy, and loneliness (both social and emotional) leads to vulnerability, which eventually leads one to sexually offend. MPR is inherently a group and social action (Harkins & Dixon, 2010), thus, perpetrators may feel a sense of camaraderie and belonging during and after committing the sexual offense. Moreover, the immense longing for social connections may lead individuals to sexually offend in a group versus alone, so they may form connections they did not have in childhood. Therefore, loneliness was investigated to further understand how it relates to MPR proclivity.

Another potential correlate of interest is anger rumination, particularly given that no research known to the authors has examined anger rumination and MPR specifically. Anger rumination consists of focusing attention on angry moods, recalling past anger experiences, and thinking about the causes and consequences of anger episodes (Sukhodolsky et al., 2001). In the sexual offense literature, anger rumination plays a key role in grievance thinking, which is a type of cognitive distortion wherein an individual feeling wronged and experiences a combination of hostility, rumination, and vengefulness (Barnett, 2011). Grievance thinking has been identified as a motivational factor for sexual murders (Wakeling & Barnett, 2011). Deficits in controlling one’s grievance thinking may lead to sexual offending as a way to obtain revenge on women who the individuals feel have wronged them. Thus, anger rumination was chosen to better understand how negative ruminative thoughts may lead to MPR. Through further understanding of how anger rumination plays a role in sexual offending, particularly MPR, theoretical and clinical implications may be drawn to further the grievance thinking model.

Lastly, psychopathy was of interest due to its mixed relationship with sexual offending, and the possibility of having a different relationship with MPR than lone sexual offenses (Darjee, 2019; Higgs et al., 2019; Moretti et al., 2024; Mouilso & Calhoun, 2013). Drawing from Hare’s (2003) four factor model (i.e., Interpersonal Manipulation, Callous Affect, Erratic Lifestyle, and Antisocial Behavior), MPR could theoretically be linked to psychopathy and its factors in multiple ways. Examples of Interpersonal Manipulation traits include pathological lying, consistently glib, manipulative, and untrustful. MPR offenders may lie and manipulate victims as a group to coerce victims into situations, where they cannot obtain help. Callous Affect traits include a lack of remorse and empathy, low affect, and tough-mindedness. MPR offenders tend to be more violent in nature when compared to other sexual violence, so having a lack of remorse and empathy would allow MPR offenders to commit the rape without internal conflicts. Erratic Lifestyle traits include impulsivity, recklessness, and thrill-seeking behaviors. Because MPR offers a sense of security in numbers, offenders may act more impulsively and recklessly, which may lead them to engage in MPR. Lastly, Antisocial Behaviors traits consist of criminal behaviors, aggressive acts, and hostility. Through the four-facet model of psychopathy, we hope to further understand how MPR may relate to psychopathic traits.

Sociocultural factors of MPR are factors ingrained in our culture or community, such as rape-myth cognitions or negative stereotypical attitudes and beliefs about women. Examples of rape-myth cognitions include victim blaming, denial of harm, misunderstanding consent, and rationalizing the perpetrator's actions (Ryan, 2019). They perpetuate sexual violence by reducing the severity of rape and its impact on victims. In this study, we focus on rape-myth cognitions, as prior research shows a positive association between rape-myth cognitions with multiple-perpetrator rape proclivity (Adolfsson et al., 2020; Alleyne et al., 2014; Palermo et al., 2019). Moreover, higher levels of rape-myth cognitions have demonstrated to predict sexual violence perpetration (Bohner et al., 1998; Chiroro et al., 2004; O’Connor, 2021; Süssenbach & Euteneuer, 2024). This underlies the importance to understand rape-myth cognitions in the context of MPR, as individuals may reinforce each other’s beliefs, which may normalize and enable sexual violence, furthering a culture where sexual violence is more likely to occur and be excused.

Research has found that men rather than women are more likely to believe rape myths, as well as individuals who did not know the victim (Crall & Goodfriend, 2016). The researchers also discovered no significant difference in rape myth acceptance based on sexual orientation or level of post-secondary education.

More specifically, Johnson and Beech (2017) observed that there haven’t been many studies to date (n = 8) that compared the general public with sexual offenders and their rape myth acceptance rates. However, in their meta-analysis, the authors discovered that sex offenders in contrast to non-offenders had lower sex role satisfaction, high trait anxiety, and high trait anger. Consequently, there was no difference in rape myth acceptance by rapists and sexual murderers. In a group of exclusively rapists, intimacy and loneliness deficits were correlated with rape myth acceptance and hostility towards women.

Work by Fisher et al. (1999) used the UCLA Emotional Loneliness Scale to evaluate child molesters and prison officers on their emotional loneliness to see if there were differences by sample. The authors discovered that loneliness was indeed higher in the child molesters. In the same vein, offenders (murder, trafficking, robbery or theft) were higher in psychopathy on the PCL-R than non-offending male youths (Castellana et al., 2014).

These results indicate that there is empirical support for the speculation that offending samples are inherently different on beliefs and personality traits (e.g., rape myth acceptance, trait anger, emotional loneliness, and psychopathy) than nonoffenders. Thus, the current study looked to expand the literature within a non-offending sample.

The current study aims to build upon the MPR research by investigating potential correlates associated with a proclivity for MPR and reconfirming prior findings in the literature. Specifically, the correlates that will be examined are loneliness, aggressive behavior tendencies, anger rumination, deviant sexual interests, rape-supportive cognitions, and psychopathy. Through an understanding of the various correlates, potential dynamic risk factors and clinical targets may be identified to reduce prevalence and recidivism.

The current study was reviewed and approved by King’s University College Research Ethics Review Committee. As part of a larger study, data was collected for individuals of all genders in a single Qualtrics survey (a high-level summary of the results for both men and women was presented in the newsletter for the Criminal Justice Section of the Canadian Psychological Association, Crime Scene; see Iqbal et al., 2024). Given that the main measure in this study (i.e., the M-PRIS) has only been validated in men, the decision was made to analyse the men and women in the sample separately (see Schwier et al., 2025, for the results from the women). This decision was made as there are distinct research questions and purposes for the two subsamples—the subsample with the men involves a replication and extension of previous research in the area and the subsample with the women involves a preliminary validation of the M-PRIS in this population and an exploration of correlates that have been previously tested with men. Unfortunately, there were not adequate numbers of non-binary individuals to allow for their inclusion in either output. The entire study was completed online via Qualtrics After giving consent, participants completed a demographics questionnaire, the Multiple-Perpetrator Rape Interest Scale, the UCLA Loneliness Scale: Short Form, the Buss-Perry Aggression Questionnaire: Short Form, the Sexual Fantasy Questionnaire, the Anger Rumination Scale, the Updated Measure for Assessing Subtle Rape Myths, and the Self-Report Psychopathy-III: Short Form with no time limit. Measures were in the aforementioned order, and titles of the measures did not appear in the survey. After completing the measures, participants were debriefed, thanked for their time, and participants could enter themselves into a raffle for the opportunity to win one of two $50 CAD gift cards.

Participants were recruited through advertisements on social media (e.g., Twitter, Facebook, Reddit, Mastodon, and message boards dedicated to research), which were posted by the researchers. Participants were men from Canada and the United States (N = 115, M_Age = 30.09, SD_Age = 7.55). They were primarily White/Caucasian (60%), and the rest consisted of African-American/Black (6.96%), East Asian (3.48%), South Asian (4.35%), South East Asian (4.35%), Middle Eastern (4.35%), Native/Aboriginal/Indigenous (4.35%), Hispanic/Latino (4.35%), West Indian (1.74%), multi-racial - 5 (4.35%), and Other (1.74%). We obtained 451 responses and removed responses (N = 336) that had a very low completion rate, responded as “Trans,” “Non-Binary,” or “Female,” and/or completed the survey in less than ten minutes (see Schwier et al., 2025, for an analysis of the data from women [n = 182]).

Means and standard deviations were derived for all scales and their subscales (see Table 1). The means for the M-PRIS Total, UCLA Loneliness Total, BPAQ Total, BPAQ Physical Aggression, BPAQ Verbal Aggression, BPAQ Anger, BPAQ Hostility, ARS Total, ARS Angry Afterthought, ARS Thoughts of Revenge, ARS Angry Memories, ARS Understanding of Causes, SRMA Total, SRMA She Asked For It, SRMA He Didn’t Mean To, SRMA Wasn’t Really Rape, SRMA She Lied, SRP Total, SRP Interpersonal Manipulation, SRP Callous Affect, SRP Erratic Lifestyle, SRP Antisocial Behavior were below the midpoint of the scale. The means for the SFQ Total, SFQ Exploratory, SFQ Intimate, SFQ Impersonal, and SFQ Sado-Masochistic were above the midpoint of the scale.

There were positive, significant associations between all exploratory variables and the outcome variable, M-PRIS Total, that ranged from weak to strong (see Table 1 for Pearson’s r). Exploratory variables are the total scores and subscales for each measure used: UCLA: SF, BPAQ: SF, SFQ, ARS, SRMA, and SRP-III: SF. The strongest relationship for M-PRIS was SFQ Sado-Masochistic (r = .79, p < .001).

We conducted a multiple linear regression to gauge which dynamic risk factors were the strongest correlates for M-PRIS Total. As seen in Table 2, all variables in the model were centered, which consisted of the UCLA Loneliness Total, the BPAQ, SFQ Total, SRMA Total, ARS Total, and SRP Total. Results showed a significant model, F(6, 108) = 28.6, p < .05, which explained 61.4% of the variance. Specifically, BPAQ Total, SFQ Total, and ARS Total were significant (sr² = .0543, 0.0499, 0.0488, respectively).

Several studies have found loneliness to be strongly related to lone sexual offending (Hanson & Morton-Bourgon, 2005; Marshall, 2010; Seidman et al., 1994, Wielinga et al., 2021), but no study investigated the relationship between loneliness and MPR. Given the inherently social nature of group offending (Harkins & Dixon, 2010), loneliness and a need for belonging could theoretically be an important influence for willingness to engage in MPR. In our study, loneliness was found to have a weak positive relationship with MPR proclivity, and the sample’s mean was below the midpoint of the scale. The positive correlation provides evidence that, despite loneliness not being a significant correlate of MPR proclivity in the multiple linear regression, an individual’s feelings of loneliness may still have an impact on proclivity for MPR and should be further investigated both by researchers and clinicians. Specifically, understanding feelings of loneliness and how to build meaningful social bonds with others may serve as a protective factor against sexual recidivism. A possible reason for the weak relationship and low mean is that the UCLA Loneliness Scale: SF is not a forensic measure or created to be identified as a risk construct, or the sample has a stronger level of social support and intimacy (Wielinga et al., 2021). Therefore, the feelings of loneliness captured in this assessment may not have been appropriate to identify the true relationship between loneliness and a proclivity for MPR–further strengthening the call for more research into the area.

Aggressive behavior tendencies have been linked with MPR, with MPRs tending to be more aggressive in nature compared to lone sexual offenders (Hauffe & Porter, 2009; Porter & Alison, 2006; Woodhams & Cooke, 2013). In our study, aggressive behavior tendencies (i.e., physical aggression, verbal aggression, hostility, and anger) were found to have a moderate positive relationship with a proclivity for MPR, which supports prior findings. One potential explanation for the heightened aggressiveness may be rooted in the cognitive appraisal that there is safety and power in numbers, and the victim physically cannot stop the offenders. Through this line of thinking, they may become unbound by any societal or personal restrictions. The findings also suggest that there was no single dimension of aggression that exclusively overpowers another; however, BPAQ Physical Aggression did have the strongest relation among the subscales. To enhance prevention efforts, forensic clinicians should strive to understand how MPR offenders' aggressive behavior tendencies begin and how they cope (or fail to cope) when feelings of hostility or anger arise.

The current study found the strongest overall relationships between deviant sexual interests (i.e., SFQ Total, SFQ Exploratory, SFQ Intimate, SFQ Impersonal, SFQ Sado-Masochistic) and a proclivity for MPR. Additionally, the means were above the midpoints for the total and subscales indicating fairly moderately high levels of these fantasies. Findings further confirm Harkins and Dixon's (2013) Multi-Factorial Theory of Multiple Perpetrator Sexual Offending, which posited that deviant sexual interest is a leading individual factor for MPR. Previous research (e.g., Alleyne et al., 2014) did not include sadomasochism in their studies, which interestingly was found to be the greatest correlate for MPR proclivity out of all the correlates that were tested. MPR can be seen as one of the most violent sexual offending types, which may explain why sadomasochism was found to hold a strong linkage to MPR. A group’s sadomasochism drive may explicitly come out to hurt and feel immense pleasure from the pain they are inflicting.

The Multi-Factorial Theory of Multiple Perpetrator Sexual Offending (Harkins & Dixon, 2010) posits the sociocultural factor that rapes myth cognitions, such as justification for sexual violence against women, may influence individuals to potentially participate in MPR. The current study supports the notion as the SRMA Total has a significant correlation with a proclivity in MPR. Alleyne and colleagues (2014) reported similar findings, where rape myth cognitions were a major correlate of MPR proclivity. The lack of significance of the SRMA in the multiple linear regression may be due to collecting data from a community sample. To support this assumption, previous work by Johnson and Beech (2017) found that rapists were significantly more accepting of rape myths than non-sexual offenders and nonoffenders.

No prior research has investigated the relationship between anger rumination and MPR, but Johnson and Beech (2017) discovered higher trait anger in sex offenders than nonoffenders. Anger rumination was selected as a prospective correlate due to its established significance in prior research as a robust predictor of various manifestations of aggressive and violent behaviors (Anestis et al., 2009; Denson, 2013; Vasquez et al., 2012). In the current study, anger rumination (i.e., ARS Total, ARS Angry Afterthoughts, ARS Thoughts of Revenge, ARS Angry Memories, ARS Understanding of Causes) had an overall moderate relationship with proclivity for MPR. As anger rumination was significant in the multiple linear regression, the relationship with MPR proclivity demonstrates the potential need for prevention programs and clinicians to target anger rumination among those who may be at risk for MPR offending. Obtaining a fuller understanding of how anger rumination plays in MPR offending will allow for treatment and prevention programs to better address this correlation with MPR proclivity. Furthermore, results demonstrate that further understanding of how anger rumination, within the context of grievance thinking, is needed to potentially fully understand MPR (Barnett, 2011; Wakeling & Barnett, 2011).

Lastly, the four-factor psychopathy model, which consists of Interpersonal Manipulation (i.e., pathological lying, glib, being manipulative), Callous Affect (i.e., lack of remorse and empathy, low affect), Erratic Lifestyle (i.e., impulsivity, recklessness), Antisocial Behavior (i.e., criminal behavior, aggression), was investigated due to its mixed relationship with sexual offending (Darjee, 2019; Hanson & Morton-Bourgon, 2005; Higgs et al., 2019; Moretti et al., 2024; Mouilso & Calhoun, 2013). Psychopathy was not significant in the multiple linear regression, which may be due to the non-clinical and/or forensic sample in comparison to Castellana and colleagues’ (2014) work which found that young offenders scored higher on the PCL-R than non-offenders.

Those interested in MPR offenses seem to reflect multiple psychopathic traits, which supports prior research on sexual offender characteristics. MPR is an extreme form of sexual violence, so individuals with a lack of remorse, high recklessness, and willingness to commit a violent/criminal action do fit a plausible character profile for an MPR offender.

There are a few limitations to the study. First, the study primarily used self-report measures for all variables. The research design was cross-sectional, so causation cannot be determined among any associations. Although self-report measures are efficient, they provoke some concerns. Social desirability bias may have influenced responses (King, 2022), such that loneliness and proclivity for multiple-perpetrator rape are rated in a more positive frame (i.e., answered to perceive self with lower loneliness and proclivity for multiple-perpetrator rape). Furthermore, the M-PRIS measures interest in MPR behaviors which may not necessarily translate into action should these individuals be placed in a similar situation. Research does indicate that there is a relatively high degree of concordance between fantasy and behavior in relation to paraphilic behavior (Joyal & Carpentier, 2022; Seto et al., 2021) and sexual aggression (Birke & Bondü, 2023), this is an untested assumption for MPR proclivity. Therefore, future research on concordance and discordance between proclivities and actual behaviors may be important future directions and should be considered in future studies. The sample size was also relatively small, meaning that some statistical tests may be underpowered and the true effect is not present (Button et al., 2013). Therefore, study results should be considered carefully and future research should strive to collect a larger sample.

Although the study presented interesting findings, there are ways to improve in future studies. To further examine the associations between the dynamic risk factors and proclivity for multiple-perpetrator rape, future studies should employ more sophisticated designs and statistical procedures in order to establish potential causal relationships. Future studies could also interview those charged or incarcerated for MPR offenses to establish their experiences of how these factors may have played a role in their behaviors. Lastly, future studies should obtain a larger sample that would yield more power and a more diverse sample that includes more racial and ethnic individuals.

One implication is for clinicians to target various types of aggression, deviant sexual fantasies, and anger rumination in therapy among those with an interest to commit MPR to potentially reduce the urge to commit the action. Educational and preventive initiatives aimed at addressing sexual violence behaviors may also gain insights into individuals prone to engaging in such behaviors, where these programs seek to diminish the likelihood of MPR occurrences. The results have important implications for clinical, education, and prevention efforts related to sexual violence (e.g., Bringing in the Bystander; Banyard et al., 2004).

Multiple-perpetrator rape is an underreported form of sexual violence (Chambers et al., 2010), and understanding the potential dynamic risk factors of individuals may be a solution to reduce the prevalence. The findings from this study demonstrate that multiple dynamic risk factors correlate to a proclivity for MPR. The results contribute to the sexual violence literature by identifying potential dynamic risk factors (i.e., loneliness, aggressive behavior tendencies, anger rumination, psychopathy, deviant sexual interests, and rape myths acceptance) as major correlates for a proclivity for MPR and reconfirming prior research findings on MPR.