Introduction

Sexual assault is increasingly recognized as a major public health concern within the military [1]. Prevention of these type of crimes is a high priority for the Department of Defense (DoD) [2] and thus, changes have been implemented [3]. Since 2005, sexual assault survivors have had two reporting options: a confidential restricted report, or an unrestricted report, which triggers an official investigation [3]. After making either type of report, a range of services, including support and treatment, are available to survivors. There has also been reform to the Uniformed Code of Military Justice, such as the National Defense Authorization Act of Fiscal Year 14, which eliminated the 5-year statute of limitations for violations of sexual assault and removed Commanders’ ability to consider the character and military service of the accused when determining disposition of the offence [4]. The DoD has also worked to improve training for those responsible for responding to sex crimes (e.g., military leadership, criminal investigators, Sexual Assault Response Coordinators) [3].

Despite these changes, rates of sexual assault remain high in the military and are similar to [6] or higher than rates of sexual assault in civilian populations. Based on the DoD Annual Report on Sexual Assault, there were 6,290 official reports of adult sexual assault by U.S. active duty service members in 2020; the Marine Corps had 5.9 reports per 1,000 service members, the highest rate of all the services [5]. The number of official reports of sexual assault in the military has been steadily increasing in the last decade, which may reflect both greater willingness to report as well as an actual increase in prevalence. Still, total reported incidents substantially underestimate the scope of the problem [5, 6]. The discrepancy between sexual assault experienced, and reported sexual assault cases, is a common phenomenon observed in both civilian and military populations, and further complicates efforts to mitigate these crimes.

The number of military personnel experiencing sexual victimization is particularly concerning given that sexual assault not only increases the risk of adverse short- and long-term health and economic outcomes, but also can impede mission readiness and command morale [1, 7]. Research has linked sexual trauma victimization to increased risk of substance abuse, anxiety, eating disorders, posttraumatic stress disorder, early separation from military service, and post-service unemployment [8–11]. However, the majority of research to-date, has examined either impacts of military sexual violence on survivors, or factors associated with increased risk of sexual victimization. In contrast, limited research has focused on the characteristics of military perpetrators of sexual assault [12]. Better understanding of the characteristics and behaviors of perpetrators of military sexual violence is essential in developing effective prevention programs, and providing additional training or intervention strategies to potential offenders, given that they are responsible for sexual violence.

While evidence suggests sexual perpetrators do not share a single set of traits, previous civilian research has identified an array of predictive characteristics associated with sexual offending [13]. For instance, sexual offenders report poor intimacy skills [14, 15]; deviant sexual interests or preferences [15]; and high levels of social difficulties, avoidance, and loneliness [15–18]. Sexual perpetrators also tend to have lifestyle instability and impulse control issues [13, 15], pro-offending attitudes and behaviors [19], and a poor capacity for victim empathy [20]. Other studies suggest a history of impersonal or risky sexual behavior (e.g., early sexual initiation, sex without a condom) heightens risk for sexual aggression [21–23]. In addition, prior delinquent behavior and cumulative exposure to stress or traumatic events predict future sexual aggression [24–26].

Despite the number of previous studies that have examined the characteristics of civilian sexual offenders, research examining characteristics of military sexual perpetrators is sparse [26–28]. One longitudinal study of male U.S. Navy recruits found 13% self-reported perpetrating sexual assault, although the majority (71%) only perpetrated prior to military service [27]. Service men who reported pre-service sexual assault were nearly 10 times more likely to perpetrate during their first year in the Navy. A follow-on study found that a similar set of risk factors predicted sexual harassment and sexual assault [26]. Importantly, researchers found prior sexual assault perpetration, number of sex partners (impersonal sex), hostility to women, and delinquency/misconduct were predictive of sexual harassment and sexual assault. In another study, researchers reviewed administrative data sources of U.S. soldiers and found that sociodemographic, history of deployment, lower rank, criminal history, substance abuse, and mental health factors predicted sexual assault perpetration [28]. Given the limited number of studies, more research is needed to better understand characteristics associated with sexual misconduct among U.S. Marines.

Each year, the Naval Criminal Investigative Service (NCIS) Criminal Data Analysis Division reports data on five major crime topics, including adult sex crimes and child sexual abuse, committed within the nexus of the Department of the Navy, including active duty U.S. Sailors and Marines [29, 30]. Common characteristics of perpetrators of adult sexual assault include being male, active duty, and in the ranks E3 to E6. However, these official reports only provide details of the offenses, victims, and subjects. Thus, additional analyses are needed to compare factors, such as family history, childhood factors, and life stressors, of military perpetrators with the population at large to better understand military sexual offenders. Furthermore, using self-reported data to predict subsequent sexual offense convictions, adjusting for other factors, may provide a unique opportunity to study the effects of pre-service factors.

This retrospective cohort study addressed this knowledge gap by using pre-service survey data collected from a large sample of U.S. Marines, who were participants of the Recruit Assessment Program (RAP), linked with sexual offense conviction data obtained from a military electronic database. The aim was to determine pre-service factors (i.e., demographics, family history, childhood experience, life events and experiences, and risky behaviors) associated with subsequent sexual misconduct among male Marines. Such findings could aid in: 1) developing or redesigning programs to reduce sexual misconduct in the military, 2) identifying patterns of offending behaviors, and 3) assisting investigators in designing interview questions about service members accused of sexual misconduct.

Materials and methods

Results

Of the 146,307 active duty male Marines, the majority were 18–19 years old (n = 97,520, 66.7%) and non-Hispanic, White (n = 90,877, 62.1%) with a high school education or less (n = 112,304, 76.8%; Table 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Demographics, family history, and childhood experiences by sexual offense conviction among U.S. male marines.

https://doi.org/10.1371/journal.pone.0278640.t001

The life events and experiences as well as risky behaviors are summarized in Table 2. During the study period, 107 (0.07%) Marines in our study sample were convicted of a sexual offense. Most of the convictions were for sexual assault (n = 39, 36.4%) or rape (n = 33, 30.8%), while the remaining were other types of sex crimes, including sexual contact (n = 19, 17.8%), indecent act/exposure/viewing (n = 10, 10.3%), and sexual abuse (n = 5, 4.7%). One third of the victims were children under the age of 16 years (n = 36, 33.6%).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Life events, experiences and risky behaviors by sexual offense conviction among U.S. male marines.

https://doi.org/10.1371/journal.pone.0278640.t002

Based on unadjusted ORs, demographic factors associated with a sexual offense included: 17 years of age (when entering recruit training); American Indian/Alaskan Native, Hispanic, or multiracial/other race and ethnicity; and ≤high school education (Table 1). In regards to family history, participants whose parents earned ≤high school education were proportionally more likely to have a sexual offense conviction. Many of the childhood experiences were associated with greater likelihood of a sexual offense conviction; Marines who reported being raised by one parent, death of a parent, lower family economic status, and some of the ACEs (i.e., emotional trauma, physical abuse, sexual abuse, domestic violence, guardian with mental health problem) were proportionally more likely to be convicted of a sexual offense. None of the life events or experiences were associated with the outcome (Table 2). The risky behaviors associated with a sexual offense were history of suspension from school and having unprotected sex.

The final, reduced multivariable logistic model consisted of a total of 109,956 participants, including 72 with a sexual offense conviction (Table 3). After implementing backward elimination, five variables remained in the final model: race and ethnicity (American Indian/Alaskan Native OR: 5.28; 95% CI, 1.86–14.98; Hispanic OR: 1.83; 95% CI, 1.06–3.18; multiracial/other OR: 3.28; 95% CI, 1.56–6.89), education (≤high school OR: 2.65; 95% CI, 1.21–5.80), parental death (OR: 2.27; 95% CI, 1.16–4.45), unprotected sex (OR: 1.78; 95% CI, 1.03–3.05), and school suspension/expulsion (OR: 1.64; 95% CI, 1.02–2.65). In the supplemental analysis (n = 124,113), including participants with non-missing data for the five predictors in the final model, effect estimates and statistical significance for each predictor were similar except school suspension/expulsion, which was attenuated and not statistically significant (Table 3). When the individual ACE variables were replaced with the ACE score, results remained largely consistent, although school suspension/expulsion was no longer statistically significant. Race/ethnicity did not significantly moderate any of the association between the predictors and sexual offense conviction (all p>.05), thus no additional analyses were conducted to further examine racial/ethnic differences.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Multivariable logistic regression model for sexual offense conviction among U.S. male marines.

https://doi.org/10.1371/journal.pone.0278640.t003

Discussion

This retrospective cohort study identified certain pre-service factors predictive of a sexual offense conviction among U.S. Marines. It expands on previous research that has examined various aspects of sexual trauma within the military context. To our knowledge, this is the first large prospective study of male U.S. Marines to examine pre-service factors associated with subsequent conviction of sexual offense. Of the numerous factors investigated, predictors associated with increased risk of sexual misconduct included: race and ethnicity, education level, parental death, school suspension/expulsion, and unprotected sex in the final multivariable logistic regression model. Results also underscore how few participants were convicted of a sexual offense over a 9-year period (0.01%), highlighting a large discrepancy between self-report estimates of sexual assault perpetration [27] and convictions of sex crimes.

Consistent with previous research, some childhood adversities were associated with a sexual offense conviction. Marines who experienced the death of a parent were more likely to be convicted of a sexual offense. The death of a parent can be one of the most traumatic childhood experiences; parental death is often clustered with other childhood adversities and may lead to additional life stressors [36]. Death of a parent has been found to be associated with a range of adverse outcomes, including mental health problems [37]; Furthermore, some Scandinavian research suggests parental death is associated with perpetrating a violent crime [38]. It may be that loss of a parent impairs the capacity to regulate emotional and behavioral responses to stressors, which increases likelihood to perpetrate violence. Previous research indicates that childhood adversities may contribute to sexual aggression by encouraging hostility toward intimate partners and impersonal sex; however, none of the other adverse childhood experiences were directly associated with a sexual offense conviction [21, 39, 40]. While many of these ACEs, such as childhood sexual abuse and physical abuse, were associated with a sexual offense conviction in our unadjusted analyses, these experiences were no longer associated with the outcome after adjusting for other factors. This may indicate an indirect relationship between childhood adversities and sexual assault perpetration, as has been suggested by past literature [21, 39]. In addition, some evidence suggests that ACEs, which are more prevalent among service members, appear to have less of a long-term impact among male service members compared with male civilians [41].

A history of engaging in risky behaviors often predicts negative outcomes; we found that Marines who reported a history of school suspension or expulsion had elevated odds of having a sexual offense conviction, though not significant in the supplemental models. This finding is consistent with previous research that indicates that prior misconduct or delinquency is associated with sexual assault [26]. Marines who reported unprotected sex had elevated odds of a sexual offense conviction in comparison with those who had sex with a condom, which aligns with prior findings of history of impersonal or risky sexual behavior (e.g., early sexual initiation, sex without a condom) heightening the risk for sexual aggression [21–23, 42]. Specifically, a previous report observed that men with a history of sexual assault perpetration were at increased risk of condom use resistance [23]. Similarly, Frye [42] found that men who perpetrated interpersonal violence were half as likely to have consistent condom use, even after adjusting for other factors.

Evidence suggests that sexual assault is a highly underreported crime with even fewer cases that proceed to the point of charges or prosecution [5]. Thus, the small number of Marines convicted of a sexual offense in this study is not surprising. In both military and civilian contexts, there are complex factors that discourage reporting and provide challenges to prosecution [43]. For instance, sex crimes often occur with no witnesses and under circumstances that make victims embarrassed or fearful of reporting. The military faces additional unique challenges in terms of charging and convicting potential offenders. Up until the end of 2021, commanders were responsible for deciding how to handle alleged sexual offense assailants; they could instigate administrative action, enact nonjudicial punishment, send to a higher authority, or possibly take no action if the victim declined to participate or evidence was insufficient to prosecute [5]. Consequently, sexual assault survivors may be less motivated to report these types of crimes in a system where sex offenders are less likely to be charged and convicted. The majority of adult victims of military sexual offenders are military members [29]. Using 2018 survey data, only 30% of Marines who experienced sexual assault reported it to military authorities [44]. Service members who have experienced sexual assault report numerous barriers to reporting the crime, including, but not limited to: not wanting people to know, concern that nothing will be done, and potential ostracism or retaliation from fellow military members [44, 45].

Because this study relied on military criminal investigative records to assess convictions of sexual offenses, some of the significant effects identified may be influenced by systemic biases in reporting and prosecuting sex crimes rather than differences in military perpetration patterns. For instance, the NCIS annual reports indicate the majority of the officially reported Marine on Marine sex crimes involved Marines of a similar peer group [46], yet a study found 80% of military sexual offenders were of higher rank than their victim [47]. Thus, it is possible that individuals may be less willing to report, charge, and convict a person of a sex crime if the perpetrator is more senior in rank [47]. In the current study, those with ≤high school degree (vs. some college education) and who identified as American Indian/Alaskan Native, Hispanic or multiracial/other (vs. White, non-Hispanic) were more likely to be convicted of a sexual offense. While these results are consistent with one previous study among soldiers that also relied on administratively-recorded sex crimes [28]; they diverge from another study that used self-reported data to identify sexual perpetration, which did not find any significant differences in race, ethnicity, or educational attainment by perpetration status [27]. Taken together, these findings indicate that these demographic characteristics may be more related to charges and convictions of sexual offenses, rather than perpetrations of sexual offenses. It is also possible that those who have not received a college education may be less sophisticated about circumnavigating the system and hence may be more likely to be charged and convicted of a sex crime. In addition, there are racial and ethnic disparities in the U.S. criminal justice system, where people of color or those who identify as American Indian, Black non-Hispanic, or Hispanic/Latino are more likely to be arrested [48, 49], and this bias likely exists within the U.S. military justice system as well. A 2019 U.S. Government Accountability Office report [50] found racial and ethnic disparities within the military criminal investigative organizations. Given growing concerns of unconscious bias within civilian and military criminal and law enforcement sectors, these potential inequalities in the military law enforcement and justice system may need to be further examined.

Previous studies suggest that additional factors (e.g. early sex, criminal history, substance abuse) may be associated with sexual perpetration [22, 28]. In a civilian sample, engaging in early sex was a predictor of self-reported sexual perpetration [22]; a criminal history and substance use diagnoses and treatment during military service were related to cases of sexual assault among U.S. Soldiers [28]. The current study did not find sexual encounter onset, history of arrest, or pre-service alcohol use predictive of a sexual offense conviction. However, the study samples, assessment of predictors, and type of outcomes varied across these studies. Thus, caution must be taken when comparing results across various studies; in addition, replication and expansion of these type of analyses are needed.

Notably, to our knowledge, this is the first study, that has examined predictors of sexual perpetration among U.S. Marines. This study used pre-service self-reported data to assess potential predictors of subsequent sexual offense conviction among a large cohort of Marines, which allowed for the longitudinal assessment of risk. Test-retest reliability analyses found RAP survey data to be reliable [51]. The outcome was derived from official criminal records, which avoids potentially influential self-reporting bias. However, it is important to interpret the current study findings in light of its limitations. Multiple factors known to be associated with sexual assault perpetration were not assessed as part of the RAP survey (e.g., prior sexual assault perpetration, number of sex partners, and interpersonal hostility). Low rates of sexual assault convictions limited the power of our analyses to identify significant predictors, requiring larger effect sizes to achieve statistical significance and preventing the opportunity to parse the outcome by specific sex crimes. While the overall response rate to the RAP survey is high, this study was restricted to male Marine who trained at MCRD, San Diego; female Marines were not training at this location and data were not collected at MCRD, Parris Island during the study period. MCRD, San Diego also primarily trains those with a home residence west of the Mississippi River, who may be demographically different from those who train at MCRD, Parris Island. The predictors assessed in this study relied on self-reported data, which may be subject to recall and reporting bias.

Even though multiple programs within the DoD have been implemented with the intention of preventing incidents of sexual assault, findings from this study and others suggest that more changes are needed to prevent sex crimes, hold perpetrators accountable, decrease barriers for reporting, and increase support for survivors [52]. Our study findings confirm that some demographic characteristics and factors related to family history, childhood experiences, and risky behaviors present an increased risk for a sexual offense conviction among U.S. Marines. It is noteworthy that events that have occurred in the past continue to influence outcomes for personnel during their military service. Effective prevention efforts could benefit from providing confidential supportive assistance to those who have a family or individual background that may present elevated risk. There are efforts underway that use computer-assisted prevention programs to confidentially assess personnel and provide tailored intervention based on known risk factors. Specifically, the Research Triangle Institute is collaborating with the DoD to provide and evaluate a technology-based, tailored sexual violence prevention training to Air Force Cadets [53]. This program seeks to prevent first-time victimization, re-victimization, and sexual misconduct by combining group classroom instruction with tailored, computer-delivered content. Findings from the current study emphasize the potential of technology-facilitated programs that provide private, targeted education; supportive assistance; and prevention materials to individuals based on risk factors. In addition, understanding how these factors influence an offender’s thinking pattern may help law enforcement investigators identify offending motivation and a pattern of offending behaviors. Thus, study results may also assist investigators in developing interview questions that will give them insight into the lives of service members accused of sexual misconduct.

Further, it is essential to continue research that helps better understand sexual assault and convictions of sex crimes in the military; future research may include examining predictors of sexual misconduct among other service branches and investigating the impact of making a restricted or unrestricted report. There have been additional efforts at the national level to further strengthen the investigation and prosecution of reported sexual assault incidents [45]. Specifically, the 2022 National Defense Authorization Act takes the prosecution of sexual assault away from the chain of command and places it under the purview of a service-specific Lead Special Trial Counsel. This counsel is independent of either the accused or potential victim of the crime’s command. These policy changes can assist in bringing more service members accused of sex crimes to court-martial, possibly increasing accountability and prosecution. Executive Order 14062 of January 26, 2022 also modifies the Uniformed Code of Military Justice to create a separate crime category for sexual harassment, increasing the visibility, and perhaps the prosecution, of this type of crime. These policy changes, in addition to other programs already in place, may decrease institutional barriers in reporting and prosecuting sex crimes. Fewer reporting barriers and more prosecution of sex crimes could in turn discourage sexual harassment and sexual assault in the military. Additionally, these military programs and research in this area may provide vital information that is translatable to civilian settings. Ongoing evaluation of the impact of policy changes and implemented programs will be essential to continue to address this persistent problem of sexual assault in the military.

Acknowledgments

In addition to the authors, Recruit Assessment Program team includes Isabel Altarejos, MPH; Zeina Khodr, PhD; Vanessa Perez, MPH; Julia Seay, PhD; Daniel Trone, PhD*; and James Yea, MPH. The authors are grateful of the contributions from Scott Roesch, PhD, Stephanie Gaines, PsyD, and Christopher J. Phillips MD, MPH, the late Principal Investigator of Recruit Assessment Program. The authors also appreciate contributions from the Naval Criminal Instigative Service, Deployment Health Research Department, Millennium Cohort Program and the Birth and Infant Health Research Team. We thank the Recruit Assessment Program participants.

*Lead author of this team (daniel.w.trone.civ@health.mil).

Disclaimer: Valerie Stander, Diane Williams, Keyia Carlton, and Jeffrey Millegan are military service members or employees of the U.S. Government. This work was prepared as part of their official duties. Title 17, U.S.C. §105 provides that copyright protection under this title is not available for any work of the U.S. Government. Title 17, U.S.C. §101 defines a U.S. Government work as work prepared by a military service member or employee of the U.S. Government as part of that person’s official duties. Report number 22–36 was supported by the Military Operational Medicine Research Program and Clinical Investigations Department at Naval Medical Center San Diego, under work unit no. 61133. The views expressed in this article are those of the authors and do not necessarily reflect the official policy or position of the Department of the Navy, Department of Defense, Naval Criminal Instigative Service, nor the U.S. Government. The study protocol was approved by the Naval Health Research Center Institutional Review Board in compliance with all applicable Federal regulations governing the protection of human subjects. Research data were derived from an approved Naval Health Research Center, Institutional Review Board protocol number NHRC.2000.0003. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.