Abstract
Sexual assault is a significant problem within the United States military. Bystander intervention skills training is recognized as a promising strategy for sexual assault prevention within both civilian and military populations. Sexual assault prevention programs which include training in bystander intervention teach individuals to notice situations that may pose a risk for harm and safely act to positively influence the outcome. This study examines correlates of bystander intervention attitudes and intentions among young adult active duty male soldiers (N = 282) between the ages of 18 and 24. Positive bystander intervention attitudes and intentions were associated with lower levels of rape myth acceptance, greater discomfort with sexism, lower likelihood of continuing an unwanted sexual advance after verbal resistance from a partner, greater likelihood of gaining verbal consent from a partner, and greater perceived peer approval for bystander intervention. In a multiple regression, perceived peer approval for bystander intervention and self-reported lower likelihood of continuing a sexual advance after verbal resistance from a partner emerged as significant predictors of positive bystander intervention attitudes and intentions (R2 = .41). Given that perceptions of peer norms are modifiable, these findings highlight the importance of addressing peer norms in bystander intervention training programs for military personnel.
Introduction
Sexual violence is a significant concern for the United States military (Bell et al., 2018; Gidycz et al., 2018; Lofgreen et al., 2017; Wood & Toppelberg, 2017). A survey of over 500,000 U.S. service members found that 5% of women and 1% of men serving in active duty roles experienced a sexual assault in the prior year (Morral et al., 2014). A meta-analysis of 59 studies conducted by Wilson (2016) revealed that 15.7% of military personnel and veterans indicate experiencing sexual assault during military service, with 2.9% of men and 38.4% of women reporting a victimization experience. The prevalence of assault among military service members underscores the importance of developing and evaluating violence prevention programs for service members (Department of Defense, 2019).
Orchowski et al.’s review (2018a) of evaluations of sexual assault prevention approaches for military service members highlight bystander intervention skills training as a commonly implemented approach across various branches of the military. This approach to sexual assault prevention is grounded in Latané and Darley’s (1970) situational model of bystander behavior. According to this model, whether someone intervenes to address a situation that poses a risk for sexual assault depends upon whether an individual notices an event, takes responsibility for doing something to help, has the appropriate skills to intervene, and chooses to step in (Burn, 2009). Numerous studies among civilian populations now document the efficacy of bystander skills training as a strategy for engaging all members of a community in sexual violence prevention (Jouriles et al., 2018). Examples of bystander intervention training programs for sexual assault prevention include Bringing in the Bystander (BITB; Banyard et al., 2007; Moynihan et al., 2015), and the Know Your Power (KYP) social marketing program (Potter et al., 2008, 2009, 2011). Both BITB and KYP have sustained adaptation and evaluation among U.S. Army Europe personnel (Potter & Moynihan, 2011; Potter & Stapleton, 2012). Other programs for military personnel also include instruction in bystander intervention skills (Foubert & Masin, 2012).
Whereas several bystander intervention skills training programs have been implemented and evaluated for military samples, there is a paucity of research examining factors associated with bystander intervention among military personnel. For example, Elliman, Shannahoff, Metzler, and Toblin’s (2018) survey of 286 soldiers found that 22% witnessed at least one situation that posed a risk for sexual harassment or assault. Examples of these situations included witnessing someone saying something sexually inappropriate to someone else or observing an individual attempting to take someone who has clearly had too much to drink home from a bar. Among the individuals who witnessed these scenarios, approximately half of the soldiers intervened (49.2%). Examining data from over 24,000 service members, Holland et al. (2016) found that female service members, those with prior sexual assault training, as well as service members with higher rank were more likely to assume responsibility for helping to address a potential sexual assault. Service members with higher levels of unit morale and greater trust in the military system were also more likely to assume responsibility for intervening in situations that pose a risk for sexual assault. Holland and Cipriano (2019) also found that rank, gender, prior experience of military sexual assault, prior sexual assault prevention training, perceived leader responsiveness, and perceived barriers to help seeking were associated with service members’ own likelihood to encourage a sexual assault survivor to seek support. Taken together, these studies offer a preliminary understanding of factors associated with bystander intervention among military personnel. Research is needed to explore whether modifiable risk and protective factors for sexual violence are associated with bystander intervention among service members.
Numerous studies among civilians highlight ways in which risk factors for sexual violence influence bystander attitudes and behavior (Burn, 2009; Chabot et al., 2009; Frye, 2007). Bystander intervention that addresses sexual assault is less likely among individuals who endorse rape myths (Banyard, 2008; Banyard & Moynihan, 2011; McMahon, 2010), report a high acceptance of rape or sexual violence (Bannon et al., 2013), or believe that their peers approve of sexual violence (Brown & Messman-Moore, 2010). Perceived peer approval for rape is an especially important correlate of bystander intervention (Murphy-Austin et al., 2016; Orchowski et al., 2016), with numerous studies suggesting the extent to which correcting misperceived peer norms can have a significant effect on an individual’s attitudes and beliefs (Orchowski et al., 2020). Studies also suggest that perceived peer support for bystander intervention is correlated with personal bystander intervention (Banyard & Moynihan, 2011; Brown & Messman-Moore, 2010; Fabiano et al., 2003; Stein, 2007). These findings underscore the need for interventions to integrate social norms approaches to violence prevention in with bystander-focused programmatic efforts.
Alcohol use is a well-documented risk factor for sexual aggression (Abbey, 2002; 2011; 2017; Testa & Cleveland, 2017), and can influence the likelihood and processes through which individuals intervene to address sexual assault (Fleming & Wiersma-Mosley, 2015; Leone et al., 2017). One theoretical mechanism that can help explain alcohol’s influence on both aggression and risk recognition is Alcohol Myopia Theory, which posits that when under the influence of alcohol, the drinker tends to focus on the most salient cues in the environment (Steele & Josephs, 1990). Similarly, quantitative research has also explored correlates of bystander intervention among heavy drinking college men and found that heavy episodic drinking was associated with lower endorsement of positive bystander intervention attitudes (Orchowski et al., 2016). Importantly, the association between heavy drinking and attitudes towards bystander intervention was explained by greater endorsement of rape supportive attitudes, suggesting that alcohol use appears to influence bystander intervention through its association with other risk factors for sexual aggression (i.e., rape myth endorsement, hypergender ideological attitudes, etc.).
Purpose of the Present Study
There is a significant need to advance sexual assault prevention efforts among service members and to identify effective strategies that are tailored toward a military environment. Whereas numerous risk and protective factors for sexual aggression—including rape myth acceptance, acceptance of violence, and perceived peer norms—are well documented correlates of bystander intervention among civilians (Banyard, 2008; Banyard & Moynihan, 2011; McMahon, 2010), no study to date provides an assessment of these factors as correlates of bystander intervention among service members. The present study addresses this gap by examining correlates of bystander intervention among a sample of young adult active-duty male soldiers. Males were the target of this research, given data suggesting that men are less likely to intervene to address sexual violence, when compared to their female counterparts (Banyard, 2008). Factors that influence bystander intervention in the context of alcohol use also vary between men and women (i.e., Fleming & Wiersma-Mosley, 2015). As such, the primary aim of the study was to explore the bivariate associations between various risk and protective factors for sexual aggression, as well as positive bystander intervention attitudes and intentions. Several hypotheses were proposed:
We also examined which of the factors examined in Hypothesis one and Hypothesis two were most strongly associated with bystander attitudes and intentions when considered together in a multiple regression analysis. This aim was exploratory in nature, and no specific hypotheses were proposed for this analysis.
Methods
Procedures
Participants were recruited from a large Southeastern U.S. Army military post. Two male research assistants conducted study recruitment from a variety of locations across the military installation (e.g., cafeterias, gyms). The study was advertised as an opportunity to provide information about alcohol use and social activities among soldiers. To be eligible for the study, men needed to be between the ages of 18 and 24, and currently serving on activity duty military service at the military installation. Participants were informed that the study was anonymous, participation was voluntary, and that their responses would not be linked with their name. Participants were also informed that they could skip any item that they preferred not to answer without penalty. Before completing the survey, participants provided consent for research participation via an electronic consent form. All assessments were completed on a tablet computer with ample space for privacy. The survey took approximately 30 minutes to complete, and men were compensated with a $10 gift card for their participation. Given military regulations, all surveys were completed during off-duty hours. The study was approved by the Defense Health Agency Regional Health Command-Atlantic Institutional Review Board.
Participants
A sample of 282 men provided complete responses to the survey items examined in this research and were included in this analysis. The average age of participants was 21.7 years of age (SD = 1.24). Men self-reported their race, with 58.6% self-identifying as White (n = 156), 20.3% as Black (n = 54), 2.8% as “Asian” (n = 8), 1.4% as Native American (n = 4), 1.4% as Native Hawaiian or Pacific Islander, with 13.5% listing their race as other (n = 38), and 5.7% (n = 16) preferred not to answer. Men also self-reported their ethnicity, with 24.1% of participants identifying as Hispanic/Latino (n= 68). Regarding relationship status, 66.7% reported their relationship status as single (n = 188), 25.9% identified as married (n = 73), 4% indicated that they were living with a romantic partner (n = 4), 2.8% identified as divorced (n = 8), and 6.4% did not provide a response (n = 18).
Measures
Participant characteristics
A short questionnaire assessed participant demographics including age, race, ethnicity, and marital status.
Bystander Attitudes and Intentions
Participant’s attitudes towards bystander intervention and bystander intervention intentions were assessed with three items developed for the study. Items were modeled after The Sexual Social Norms Inventory that includes an assessment of personal engagement in bystander intervention (Bruner, 2002), as well as the Bystander Attitudes Scale which is commonly utilized to assess college men’s likelihood to engage in bystander intervention to address sexual violence (Banyard et al., 2007). Questions included: 1) “I would ask if everything is okay if I witnessed a fellow male Soldier pressuring a woman to leave with him”; 2) “I would do something if I saw a fellow Soldier put something in a woman’s drink”; and 3) “I would respect another Soldier who steps in when it looks like a man is trying to hook up with a woman who has had too much to drink.” Participants responded to the items along a six point Likert scale, ranging from “Strongly Disagree” to “Strongly Agree.” Responses were summed to create a total score of bystander attitudes and intentions. Cronbach’s alpha for the scale in the current sample was .77.
Perceived Peer Support and Engagement in Bystander Intervention
Three items aligning with the assessment of personal attitudes and intentions to engage in bystander intervention were utilized to assess perceived peer support and engagement in bystander intervention. Items mirror the items assessing personal bystander intervention attitudes and intentions, which is a strategy for assessing perceived peer norms also utilized by the Sexual Social Norms Inventory that includes items which assess perceived peer engagement in bystander intervention (Bruner, 2002). Questions included: 1) “What percent of soldiers at [this military installation] would ask if everything is okay if they witnessed a fellow male Soldier pressuring a woman to leave with him”; 2) “What percent of soldiers at [this military installation] would do something if they saw a fellow Soldier put something in a woman’s drink”; and 3) “What percent of soldiers at [this military installation] would respect another Soldier who steps in when it looks like a man is trying to hook up with a woman who has had too much to drink.” Participants responded using a 11 point scale, ranging from “0%” to “100%.” Responses were summed to create a total score of perceived peer support and engagement in bystander intervention. Cronbach’s alpha for the scale in the current sample was .72.
Persistence with Sexual Activity Despite Verbal Resistance
Two items assessed participants’ likelihood to persist in an unwanted sexual advance, despite a partner’s verbal resistance. Questions asked about participant’s likelihood to stop sexual activity after a partner expressed that they did not consent (i.e., said “no” or “asked them to stop”). Men responded along a six point scale, ranging from “Strongly Disagree” to “Strongly Agree.” Items included: “I would stop the first time my partner says ‘no’ to sexual activity” and “I would stop sexual activity when asked to, even if I am already sexually aroused.” Items were adapted from the College Date Rape Attitudes and Behavior Survey (Lanier & Elliott, 1997), which was utilized by Allegria-Flores et al., (2017) in outcome evaluation of two sexual assault prevention programs. Items were reverse coded and summed, to reflect an overall likelihood of continuing an unwanted sexual advance without consent. Cronbach’s alpha for the scale in the current sample was .74.
Rape Myth Acceptance
Four items assessed adherence to stereotypes regarding sexual assault. Participants responded to the items along a six point scale, ranging from “Strongly Disagree” to “Strongly Agree.” Questions included: 1) “When a woman asks her date back to her place, I expect that something sexual will take place,” 2) “If a woman dresses in a sexy dress she is asking for sex”; 3) “When a woman fondles a man’s genitals, it means she has consented to sexual intercourse”; and 4) “If a woman lets a man kiss her, it means she wants to have sex,” where Item 1, Item 2, and Item 3 were gleaned from College Date Rape Attitudes and Behavior Survey (Lanier & Elliott, 1997; see also Alegria-Flores et al., 2017). Item 4 was gleaned from the Sexual Social Norms Inventory, which includes items that assess perceived peer engagement in bystander intervention (Bruner, 2002). Items were summed to create a total score of rape myth acceptance. Higher scores indicated greater acceptance of stereotypes about sexual assault. Cronbach’s alpha in the current sample was .73.
Perceived Peer Rape Myth Acceptance
Four items assessed participant’s perception of peer adherence to stereotypes regarding sexual assault. The assessment of perceived peer endorsement of rape myths mirrored the questions addressing personal rape myth acceptance, an approach for assessing peer norms which is followed by the Sexual Social Norms Inventory (Bruner, 2002). Questions included: 1) “What percent of male soldiers at [this military installation] believe that when a woman asks her date back to her place, they expect that something sexual will take place?”; 2) “What percent of male soldiers at [this military installation] believe that if a woman dresses in a sexy dress she is asking for sex?”; 3) “What percent of male soldiers at [this military installation] believe that if a woman lets a man kiss her, it means she wants to have sex?” and 4) “What percent of male soldiers at [this military installation] believe that when a woman fondles a man’s genitals it means she has consented to sexual intercourse?” Participants responded using an 11-point scale, ranging from “0%” to “100%.” Higher scores indicated greater perceived peer acceptance of stereotypes about sexual assault. Cronbach’s alpha in the current sample was .67.
Discomfort with Sexism
Adherence to gender equitable attitudes, also conceptualized as discomfort with sexism, was assessed with two items. Participants responded to the following items along a 6-point Likert scale ranging from “Strongly Disagree” to “Strongly Agree.” Items included “It bothers me when I am with a group of friends and/or soldiers and someone puts down women by making sexual jokes or comments about them” and “I feel uncomfortable when I hear a sexist remark.” These items were generated for the purpose of the current study, informed by existing measures of bystander intervention (Banyard, 2008). Responses were summed, with higher scores reflecting greater discomfort with sexism. Cronbach’s alpha for the scale in this sample was .65.
Utilization of Verbal Consent in Sexual Activity
A single item assessed use of verbal strategies to establish consent with a partner during sexual activity. Participants responded along a six point scale ranging from “Strongly Disagree” to “Strongly Agree” to the question: “During sexual activity, I try to gain verbal affirmation from my partner that what we are doing is OK.” This item was adapted from the Sexual Consent Scale—Revised (Humphreys & Brousseau, 2009).
Engagement in Heavy Episodic Drinking
One item assessed engagement in heavy episodic drinking, defined as consuming five or more drinks in a row, in one sitting. A standard drink was defined as one 12-ounce bottle can or glass or beer or wine cooler, one 5 ounce glass of wine, or a mixed drink containing one show (1.5 ounce) of liquor (NIAAA, 2018). After being provided with the definition of a standard drink, participants entered the number of days in the past month that they engaged in a heavy drinking episode.
Data Analysis
Analyses were conducted using SPSS Version 25 (IBM, Chicago, IL). Study variables were first examined for adherence to assumptions of normality (Tabachnick et al., 2007). A series of bivariate Pearson product-moment correlations were conducted to examine associations between all study variables. To assess multivariate associations between risk and proactive factors for sexual aggression and bystander intervention attitudes/intentions among soldiers, a multiple linear regression was conducted. The multiple linear regression included all variables that demonstrated a significant bivariate association with bystander intervention attitudes/intentions.
Results
Hypothesis 1: Risk Factors for Sexual Aggression and Bystander Intervention Attitudes/Intentions
A series of bivariate correlation analyses documented several associations between various risk factors for sexual aggression and bystander intervention attitudes/intentions (Table 1). Specifically, positive bystander intervention attitudes/intentions were negatively correlated rape myth acceptance, r (280) = −.21, p < .01; and persistence in sexual activity after expression of verbal resistance, r (280) = −.57, p < .001. Perceived peer rape myth acceptance and the number of heavy drinking days in the past month were not significantly associated with bystander intervention attitudes/intentions.
Attitudes Towards Bystander Intervention: Bivariate Correlations with Risk and Protective Factors for Sexual Aggression (N =282).
Note. * p < .05; ** p < .01; *** p < .001.
Hypothesis 2: Protective Factors for Sexual Aggression and Bystander Intervention Attitudes/Intentions
A second series of bivariate correlation analyses next documented several associations between various protective factors for sexual aggression and bystander intervention attitudes/intentions (Table 1). Specifically, bystander intervention attitudes/intentions were positively associated with perceived peer support for bystander intervention, r (280) = .44, p < .001; discomfort with sexism, r (280) = .21, p < .01. Garnering verbal affirmation of consent during sexual activity was also associated with proactive bystander intervention attitudes/intentions, r (280) = .37, p < .001.
Multiple Regression Analysis
Multiple Regression Predicting Attitudes toward Bystander Intervention (N = 282).
Note. *Indicates significance at the p < .05 level; F (5, 276) = 37.78, p < .001, R2 = .41.
Discussion
The present research adds to the literature examining correlates of bystander intervention attitudes among a sample of young active-duty military soldiers. The study hypotheses were supported. Regarding Hypothesis 1, positive bystander intervention attitudes and intentions were associated with lower levels of rape myth acceptance, and lower likelihood of continuing an unwanted sexual advance after verbal resistance from a partner. Regarding Hypothesis 2, several protective factors for sexual aggression were positively associated with bystander intervention attitudes and intentions. Correlates of bystander intervention attitudes and intentions included greater discomfort with sexism, greater likelihood of gaining verbal consent from a partner, and greater perceived peer approval for bystander intervention
Regression analyses revealed that, when considered in the context of other predictors, greater perceived peer approval for bystander intervention, lower likelihood of continuing a sexual advance after verbal resistance from a partner emerged as significant predictors of bystander intervention attitudes and intentions. These data align with recent research by Orchowski et al. (2020) that found that college men who perceived their peers as more likely to intervene were also more likely to engage in bystander intervention themselves. The finding that perceived peer engagement in bystander intervention is associated with personal bystander intervention attitudes also aligns with other studies of civilians (Banyard & Moynihan, 2011; Brown & Messman-Moore, 2010; Fabiano et al., 2003; Stein, 2007).
Perceived peer norms are an important correlate and predictor of bystander intervention attitudes which can be directly addressed and modified through social norms intervention strategies (see Gidycz et al., 2011). Research with soldiers highlights that many men hold misperceived norms regarding what other soldiers think and do in sexual encounters (Berry-Cabán et al., 2020). Specific to bystander intervention, prior research suggests that men do not intervene in situations that pose a risk for sexual assault because they do not believe helping to be an acceptable form of masculine behavior (Berke et al., 2019; Carlson, 2008; Tice & Baumeister, 1985). Other studies suggest that men worry that they would not be supported if they interfere with another man’s opportunity for sexual activity; even if this encounter is one without consent (Carlson, 2008; Oesterle et al., 2018). These data therefore highlight the utility of providing corrective information to soldiers, either by the indirect pathway of reducing men’s sexism, or by addressing misperceptions of bystander intervention more directly. The role of perceived peer norms is an important component of the Integrated Model of Sexual Assault (Berkowitz, 1994; 2002) that has been used to guide several promising sexual assault prevention programs for civilians (Gidycz et al., 2011; Orchowski et al., 2018b; Salazar et al., 2014).
As previously discussed, bystander intervention theory stipulates that to intervene, an individual must first label a situation as problematic (Burn, 2009). It was therefore not surprising that male soldier’s own endorsement of continuing to engage in sexual activity despite a partner’s verbal resistance was associated with lower likelihood to engage in bystander intervention. Although the questions asked did not provide a behaviorally-specific assessment of sexual aggression, as is the case in measures like the Sexual Experiences Survey—Revised (Koss et al., 2007), this variable may nonetheless serve as a representation of personal engagement in sexually aggressive behavior and indicate the need for tailored interventions for high-risk men. Because bystander intervention training is typically considered a universal and community-based prevention approach, these findings highlight how community and group-based interventions may benefit from being paired with individual-level interventions, which can target attitude and behavior change among those with varying levels of risk for perpetrating sexual aggression.
It might seem surprising at first that engagement in heavy drinking was not associated with bystander intervention attitudes/intentions. Although the existing literature regarding the association between alcohol use and bystander intervention is somewhat mixed (Fleming & Wiersma-Mosley, 2015; Orchowski et al., 2018b), there is research suggesting that the association of alcohol with sexual assault is a byproduct of other variables, such as men’s hyper-sexuality. This would suggest that variables other than alcohol use would be more directly associated with bystander intervention, which was our finding. Future research is needed to further explore the direct effect of alcohol use on bystander intervention among male soldiers is needed.
Whereas this study represents an important contribution to the literature, it should be interpreted considering several limitations. First, the study focused only on male soldiers. Research is needed to also explore factors among female soldiers to intervene to help address sexual assault risk among their peer as it is possible that bystander intervention could be associated with a different set of barriers and facilitators. This sample was also limited to men at one Southeastern military post and may not be generalized to other samples, such as men serving at differing military installations or to individuals in other branches of the military. As with many studies of military personnel, the assessment did not examine sexual orientation, which future research may consider examining. It should also be noted that in order for the survey to be administered in a relatively short period during off-duty hours, the assessments for the current research were relatively short, and adapted from longer surveys. Although the study assessments demonstrated adequate to fair internal consistency within the current study sample, follow-up research is needed which administers the full scales assessing each of these constructs, to replicate this study’s findings in this and other military populations.
There are also several pathways to extend the current study. As the present study included young male soldiers, it is unclear if findings would generalize to individuals in leadership roles. The present study also assessed attitudes and intentions towards bystander intervention, rather than actual engagement in bystander intervention behaviors. As highlighted in a recent meta-analysis, there are important distinctions between bystander attitudes, intentions and behaviors (Kettrey & Marx, 2020). Future research will benefit from expanded assessments that include measures of bystander opportunities, bystander intervention tactics, self-efficacy in bystander intervention, perceived barriers and consequences to intervention.
Conclusion
Bystander intervention training is a common component of sexual assault prevention initiatives documented for military personnel (see Orchowski et al., 2018a). Research to date examining modifiable factors associated with bystander intervention among service members is limited. The present findings therefore advance our understanding of the factors that can be targeted in interventions to facilitate bystander intervention among young adult male service members. Findings highlight the potential of targeting perceived peer norms as a means of facilitating bystander intervention. The importance of peer perceptions also highlights the need to develop rank-specific intervention strategies, as rank is an important cultural and differentiating variable in the military. Further, data highlight the need to implement bystander intervention programs in tandem with programs that address men’s personal engagement in sexual aggression.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study is supported by a grant from the Department of Defense, Congressionally Directed Medical Research Programs (W81XWH-15-2-0055).
