Abstract
This mixed-methods research explored the moral motivations of undergraduates who identified as bystanders in a situation of potential sexual assault. In the quantitative analysis, we examined the difference between interveners and noninterveners with regard to their scores on the Moral Foundations Questionnaire–30 Item (MFQ-30), which considers five moral foundations from Moral Foundations Theory (MFT) of care/harm, fairness/cheating, loyalty/betrayal, authority/subversion, and sanctity/degradation, as well as measures of bystander attitudes (BAS-R) and rape-myth acceptance (IRMA, modified). Participants who failed to intervene had significantly higher scores on the loyalty/betrayal subscale of the MFQ-30, and showed a trend toward “conservative” values comprising the latter three MFT foundations. Intervening bystanders were also more likely to endorse bystander attitudes, and less likely to endorse rape-myth supporting beliefs. The qualitative analysis examined brief narratives in which participants described their bystander experience and reasoning in the moment. Analysis found a remarkable flexibility with which each moral foundation could be used to support either intervention or abstention. We argue that emphasizing conservative values (based on loyalty, purity, and/or authority) in addition to the typical liberal (justice-based and anti-harm) reasoning may bolster bystander interventions meant to reach all students.
Introduction
The White House Task Force to Protect Students from Sexual Assault reported that one in five women is sexually assaulted in college, most often occurring in the first and second years (White House Task Force to Protect Students From Sexual Assault, 2014). The Task Force survey of 5,446 undergraduate women additionally found that of the 20% experiencing sexual assault, 6.9% of assaults were physically forced and 16% occurred when the victim was incapacitated (Krebs, Lindquist, Warner, Fisher, & Martin, 2009). Also during the undergraduate years, bisexual and homosexual men experience greater incidence of unwanted sexual contact and attempted rape compared with heterosexual men (Menning & Holtzman, 2014). Such findings underscore the need for successful campus prevention programs to address sexual assault on campuses.
The most common approach to these sexual assault prevention programs has been brief, educational curricula that address potential victims or perpetrators directly (Dills, Fowler, & Payne, 2016; Söchting, Fairbrother, & Koch, 2004). In such programs, sexual assault incidence is meant to be reduced by providing knowledge of the precipitators of sexual assault and its legal and psychological consequences, as well as imparting information about sexism. These approaches have unfortunately produced poor outcomes: The attitudinal changes are often temporary and return to preintervention levels within weeks, and results have been inconclusive on the efficacy for reducing incidence (Anderson & Whiston, 2005; Breitenbecher, 2000; Dills et al., 2016). Due to these limitations, there is a clear need for emerging approaches to expand the understanding of the individual and social factors that influence sexual assault incidence.
Recently, an emphasis on looking “ecologically” at the broader campus community and toward educating bystanders of potential sexual assault has received attention (Dills et al., 2016; Katz & Moore, 2013). The bystander approach has been efficacious in reducing rape-supportive attitudes, and in improving the confidence and willingness to intervene for potential bystanders (Cissner, 2009; Coker et al., 2011; Moynihan, Banyard, Arnold, Eckstein, & Stapleton, 2011; Salazar, Vivolo-Kantor, Hardin, & Berkowitz, 2014). As a leading example, the program Bringing in the Bystander (Banyard, Moynihan, & Plante, 2007) trains students to recognize and intervene, not only in sexual assault but in situations where rape-supportive attitudes are expressed (e.g., sexist language).
Despite the potential of this more recent emphasis on bystanders and, therefore, the entire campus community, the bystander approach seems to inherit some of the limitations of earlier approaches. Namely, bystander approaches bring with them the tacit assumption that by providing adequate knowledge about the issue and teaching steps that are helpful for intervention, this will motivate bystanders to act. In other words, these programs tend to ignore the likely crucial aspect of students’ motivation to act, and in particular, the moral viewpoint that would compel action to prevent sexual assault. We seek to understand the moral motivations of students, whether or not they intervene, as a first step toward investigating alternative or supplemental approaches to existent bystander intervention programs. A deliberate focus on students acting on their ethical principles may be applied to all students whose values likely regard campus sexual assault as a problem, albeit from differing moral perspective (e.g., rape culture, respect for rules or authority, or the upholding of dignity).
Of note, the terms morality and ethics are used somewhat interchangeably in the literature, and both terms will be encountered here. In the psychology of moral development and moral agency, ethics generally refers to specific principles or an individual’s general moral view, while morals refer to beliefs; however, moral, as an adjective, works in the same way as ethical, having to do with values and principles relating to right and wrong (Blum, 1994).
Focus on Bystanders
Research on bystanders emerged in the late 1960s when Darley and Latané (1968; Latané & Darley, 1970) wrote on the bystander effect, which referred to the tendency of onlookers to refrain from action when witnessing potentially dangerous incidents in the presence of others. The classic case that influenced this research was that of Kitty Genovese, whose violent rape and murder in 1964 Queens, New York, was purportedly witnessed by multiple neighbors. The bystander effect theory has been used to explain a wide range of situations, from responses to an epileptic seizure to acts of racism (Nelson, Dunn, & Paradies, 2011).
Several researchers have applied the bystander effect theory to the context of college sexual assault prevention (Banyard, 2008, 2011; Banyard et al., 2007; Bennett, Banyard, & Garnhart, 2014; Burn, 2009; McMahon, Postmus, & Koenick, 2011). This research has focused on barriers and facilitating factors for intervening in sexual assault in a university setting. Burn (2009) summarized barriers relating to both situational and cognitive/individual factors: (a) failure to notice (e.g., distraction); (b) failure to identity the situation as high risk (e.g., ambiguity about consent or danger); (c) lack of information about sexual assault; (d) pluralistic ignorance, that is, abiding by a norm that one disagrees with because of the belief that one’s peers agree; (e) failure to take responsibility (e.g., diffusion of responsibility, that is, believing others will intervene in your stead; perceptions about the worthiness of the victim, such as provocative dress or intoxication; complications surrounding one’s relationship to the potential victim); (f) failure to intervene due to skills deficit (e.g., uncertainty about what to do or say); and (g) failure to intervene due to audience inhibition (i.e., one’s group norms would discourage the intervention). Bringing in the Bystander (Banyard et al., 2007) engages participants in activities that address many of these barriers: lack of knowledge, audience inhibition, lack of skills, community norms, beliefs around effectiveness. An assumption in such a training is that the sexual assault bystander scenario can be made unambiguous enough with sufficient awareness (i.e., of the incidence of sexual assault, preventative steps, or even the condemnation of such behaviors) that action must follow.
Considering a Moral Motivation Approach
The attempt of bystander intervention programs to cultivate group norms that would encourage action can be reframed as an effort to change the moral perspectives of individuals in situations of potential sexual assault. Yet moral psychologists have differed on how moral motivation occurs. Kohlberg (Kohlberg, 1971; Kohlberg, Levine, & Hewer, 1983) took the view of individual progress in cognitive-developmental stages, thereby improving their capacity to assessing moral situations with increasing sophistication. Colby and Damon (1992), following in this tradition, stated that moral cognition has intrinsic motivating power. Blasi (1995), unsatisfied with Kohlberg’s largely untested assumption that moral reasoning would result in action, developed the idea of moral identity. Blasi theorized that the integration of moral responsibility into one’s view of self is a precursor to acting morally. In this way, individuals would feel their identity denied were they to refrain from acting on feelings of responsibility.
If we consider the importance of moral identity, a bystander would need to feel that it would be antithetical to their self-perception were they to abstain from intervening in a scenario that was perceived as morally wrong. Indeed, findings in the field of moral psychology suggest that explicitly labeling an attitude as having a moral basis may increase the strength of the attitude and its resistance to change (Luttrell, Petty, Brinol, & Wagner, 2016; Skitka, 2010; Skitka, Bauman, & Sargis, 2005). Furthermore, in research exploring the gap between moral reasoning and action, the strength of a moral concept has been shown to be associated with the moral responsiveness or action (Blasi, 1980; Walker, 2004). Thus, it is important to understand which moral principles are elicited and challenged in situations of potential sexual assault in order to encourage action.
Moral Foundations Theory (MFT) has emerged as an influential theoretical perspective for understanding morality and moral actions, both among scholarly communities of moral and social psychology and within the mainstream press (Graham, Haidt, & Nosek, 2009; Graham et al., 2011; Haidt, 2012). MFT describes five foundations of morality: care/harm, fairness/cheating, loyalty/betrayal, authority/subversion, and sanctity/degradation. Recently, liberty/oppression was added to the foundations, though this was not included in the instrument at the time of this study and was therefore excluded from our analysis. According to theory, individuals are motivated by affect-driven, implicit value systems established through group membership in society and culture, and evolutionary history.
Relying on Kahneman’s (2003) depiction of a “hot” and “cold” system (systems 1 and 2) whereby fast emotional processing overpowers deliberative and slow reasoning in some situations, Haidt might argue that almost all moral decision making is ruled by system 1. But Kahneman wrote that the cold system, that of reasoning, had some power to overcome the hot system and that certain factors permit or block the more rational system from leading the way. MFT asserts that moral individuals create rational defenses post hoc to support their decisions, that is, serve their “hot” thinking.
But what does this theory make of regret or troubling after-the-event thinking? Haidt (2001) provides an account in MFT of how the “activation” of particular moral foundations can change over time through the feedback of post hoc reasoning and new learning. This is to say that how one comes to understand a situation morally will change through experience, and these shifts may deepen moral conviction but may also result in shifts in the particular foundation that is triggered (e.g., from fairness to disgust). In the context of sexual assault, diverse moral perspectives could certainly be elicited from the perspectives of care, fairness, loyalty (e.g., “bro-code” or “girl-code”), authority (e.g., of the law), and sanctity (e.g., of virtue), all related to the discourse around an incident of sexual assault.
The foundations in MFT are considered universal and adaptive, but the varying emphasis placed on each foundation is influenced by culture. Within the United States, the moral foundations have been found to split on liberal–conservative lines (Graham et al., 2009). Although shared across the political spectrum, care/harm and fairness/cheating are most associated with progressive or liberal values, whereas loyalty/betrayal, authority/subversion, and sanctity/degradation align more with conservatism (Graham et al., 2009). Even so, differing versions of the foundations may be expressed. Haidt (2012) gives the example of sanctity/degradation being linked to sexual (e.g., virginity) or food-based (e.g., kosher, halal) purity in a conservative, traditional sense. But in liberally oriented communities, these same foundations could be expressed in terms of organic foods and products that similarly ward against a sense of defilement or contamination.
The present study explores the moral motivations of bystanders, in situations of potential sexual assault. Following MFT, our hope is that in asking students to reflect on these situations we can still obtain, through their reflections on lived experiences, a close idea of what motivated them to intervene or stand by. While it may be difficult to fundamentally change core belief systems (Haidt, 2001, 2012), young people develop more complex ideas about morality as they age, with the college years being a crucial time for consolidation of moral thinking and the development of moral identity (Blasi, 1980; Rest, 1986; Rest, Narvaez, Thoma, & Bebeau, 1999).
The present study seeks to understand the moral understandings of sexual assault situations that students already hold, with the intention of better representing such perspectives in bystander intervention programs. As this merging of bystander intervention research and moral psychology has not been attempted before, we did not come to this study with predetermined hypotheses. Rather, we sought to explore the nature of student’s moral perceptions of lived bystander events more broadly. The guiding research questions were as follows: How will student’s adherence to the moral foundations set forth in MFT differ by intervention status? How will MFT loadings correlate with those on the measures for bystander attitudes (BAS-R) and rape-myth adherence (IRMA, modified)?
The present mixed methods study contains two phases comprising quantitative and qualitative methodologies. Students filled out a survey that included questions about a remembered bystander event chosen by the participant. They then completed instruments that measured their adherence to the five moral foundations, bystander attitudes, and rape myths. The qualitative phase of the study examined the brief written narratives of lived experience recalled by the participants in the survey—both those who intervened and those who did not—in order to further explore the moral reasoning of interveners and noninterveners. Following Menning and Holtzman (2014) in this journal, as well as other researchers who use qualitative data to add dimension to quantitative findings, we mine this qualitative data for examples of themes and discourse that reflect themoral foundations.
Methods
Participants
The sample was a diverse cohort of undergraduate students from an urban, public university in New England. This was a commuter campus and so student housing is not provided. Of the 269 undergraduates who volunteered, 192 had completed the questionnaire and qualified for the total sample to be analyzed (completion was not required for compensation).
Of the 77 subjects who were excluded from quantitative analysis, 47 did not provide any scenario, and 23 provided a scenario that was unrelated to sexual assault (e.g., peer pressure to drink). The remaining seven subjects provided a scenario, but did not complete the instruments. The scenarios for these subjects were used in the qualitative analysis provided that the content was relevant to sexual assault. Missing values within instruments were addressed with average imputation, which was performed for 23 subjects. The maximum missing number of values was three items on a given instrument (this was for two subjects), and most often only one item was missing.
As detailed in Table 1, the final sample was majority female-identified but with a sizable number of male respondents (female 64.06%; male 34.38%; transgender 1.04%; missing .52%). The sample was representative across race/ethnicity: White, 53.65%; Latino or Hispanic origin, 16.67%; Asian, 16.06%; Black or African American, 8.33%; Other, 3.13%, .5%; nondisclosed, 3.13%; Native American, .52%. Sexual orientation was distributed as straight, 85.94%; bisexual, 8.33%; lesbian or gay, 2.08%; queer, 2.08%; nondisclosed, 1.04%; missing, .52%. As the literature has indicated gender to be influential in bystander behavior (Burn, 2009), we ran a Fisher’s exact test and found gender to be independent of intervener/nonintervener status (p = .318). Finally, age ranged from 17 to 39 (M = 21.78; SD = 4.12), and an independent-samples t test for age was nonsignificant between interveners (M = 21.98, SD = 4.33) and noninterveners (M = 21.42, SD = 3.75); t(163.81) = .94, p = .347).
Bystander Demographics.
Further situational data of the bystander situation are provided in Table 2, which provides the setting of the event and the relationship of the bystander to the potential victim(s) and perpetrator(s). We chose to use the term exploiter for the survey to avoid the legalistic connotation of perpetrator. The bystanders also provided the genders of the potential victims and perpetrators. Potential victims were mostly female (94.5%) and perpetrators mostly male (94%).
Bystander Situation Characteristics.
Procedures and Measures
Subjects were recruited by a variety of methods: flyers, electronic signage on campus, email blasts to undergraduates, and by announcement in the Introduction to Psychology course. The recruitment materials asked students if they had witnessed a “sketchy sexual situation,” in which someone may have experienced unwanted sexual contact such as at a party, club/bar, or other social gathering. They were invited to complete a 20-min online survey and be entered into a raffle to win one of four $100 Amazon gift cards.
Survey participants followed a URL from the recruitment materials to complete the online informed consent and take the survey. They were asked demographic questions regarding their age, gender, sexual orientation, past/current religious affiliation, race, and ethnicity. They were then directed to a brief priming question asking them to think of a situation in which they were bystanders to a potential sexual assault. The exact wording of this prompt is as follows: In high school and college, students who go to parties may be confronted with situations in which a woman or man are in a compromising situation, where, because of drinking or other issues, it appears as if this person is being exploited or taken advantage of or not able to consent to sexual contact. Have you ever seen something like this occur at a party or get together? We are interested in you as an observer rather than you as a person who might have experienced exploitation yourself. You might have known the people involved or not. A person you know might have been the one taking advantage of someone else or the one who was taken advantage of. Think back and try to remember a situation in detail. If there has been more than one situation, choose one that stands out in your memory.
After this prompt, they were asked to provide in open-answer sections of the survey a description of the situation, how they intervened if at all, and list their reasons for intervening or not intervening.
The subjects were then asked to describe their relationship to the potential victim and perpetrator (exploiter), the gender of the potential victim/perpetrator, if anyone was under the influence of substances, the bystander’s age, and to describe the setting. Finally, subjects completed the instruments. Subjects were then instructed to email their request to be included in the raffle or receive course credit if enrolled in the Introduction to Psychology course. They were finally asked if they would like to be interviewed for a second, qualitative study, and thanked for their participation.
Moral Foundations Questionnaire–30 Item (MFQ-30)
This measure (Graham et al., 2011) is based on Haidt’s Social Intuitionist Model of moral reasoning, and measures subjects’ preferences among the five moral foundations: care/harm, fairness/cheating, loyalty/betrayal, authority/subversion, and sanctity/degradation. The scale is divided into two sections: relevance and judgment. In the first, respondents complete a Likert-type scale ranging from 0 (not at all relevant) to 5 (extremely relevant) for moral relevance statements (e.g., “Whether or not someone suffered emotionally”) and some confounds (e.g., “Whether or not someone was good at math”). In the second part, the scale changes to ask subjects to rate their responses from 0 (strongly disagree) to 5 (strongly agree) for moral judgment statements (e.g., “Respect for authority is something all children need to learn”). Mean alpha for the MFT is .73, with the subscales ranging from .69 for Harm, .65 for Fairness, .71 for Ingroup, .74 for Authority, and .84 for Purity. Test-retest Pearson correlations were of similar magnitudes: .71 for Harm, .68 for Fairness, .69 for Ingroup, .71 for Authority, and .82 for Purity.
Bystander Attitude Scale–Revised (BAS-R)
The BAS-R (McMahon, 2010) is a 16-item scale modified from Banyard, Plante, and Moynihan (2005), which assesses bystander-supportive behaviors and attitudes. Respondents use a Likert-type scale (range 1 to 5, from “strongly agree” to “strongly disagree”) to assess the likeliness to engage in bystander-supportive behaviors ranging from opposing sexist language (e.g., “Challenge a friend who uses ‘ho,’ ‘bitch,’ or ‘slut’ to describe girls”) to engaging or intervening in sexual assault (e.g., “Decide not to have sex with a partner if she/he is drunk”). Cronbach’s alpha was .86 in McMahon (2010). McMahon also found a small but significant correlation with the Illinois Rape Myth Acceptance Scale (IRMA), r = .28, p < .001.
Illinois Rape Myth Acceptance–Revised (IRMA-R)
To keep the online survey within a 20-min time limit, we extracted five items from the revised IRMA (McMahon & Farmer, 2011) and throughout the paper refer to these as the IRMA-R Modified. The IRMA itself was a revision of the measure by Payne, Lonsway, and Fitzgerald (1999), which updated the language to reflect that used by college students. The five items do not represent a validated short version but are based on group consensus of the items that best related to rape culture. One item was from the subscale “She asked for it” (If a girl is raped while she is drunk, she is at least somewhat responsible for what happened). Two were from the subscale “He didn’t mean it” (Guys don’t usually intend to force sex on a girl, but sometimes they get too sexually carried away; If a guy is drunk, he might rape someone unintentionally). One was from the “It wasn’t really rape” subscale (If a girl doesn’t physically resist sex—even if protesting verbally—it really can’t be considered rape). The fifth question was from the subscale, “She lied” (A lot of times, girls who say they were raped agreed to have sex and then regret it). These appeared on a 5-point Likert scale ranging from 1 (strongly agree) to 5 (strongly disagree). The five items were summed for a total score.
Regarding the qualitative wing of the study, the subject descriptions of their bystander event and their response were analyzed using the qualitative methodology of Braun and Clarke’s (2006, 2013), which combines thematic and discourse analysis. The method involves close reading of the narratives, preliminary grouping into themes and discourse categories, labeling and defining these, and then revisiting the membership of various narratives within these themes and discourses to further develop the categories. In this case, we decided to use the moral foundations as themes and explore whether starting from a place where the themes were predetermined theoretically, while remaining open to other themes, would be fruitful. We did not merely identify the themes, however, but examined how the themes might be used in a way that supported intervening or not. The richness of the qualitative data surely deserves more space than the end of this article provides; however, for the purpose of expanding upon the moral foundations as motivation for intervening, we provide a selection of relevant examples below. In particular, there was not space to describe both the thematic and discourse approaches as the discourse approach compels us to present a closer reading of some excerpts. But we do present a collection of what students “notice” as part of one discourse we are currently examining, the discourse on drinking. We are preparing a separate paper that delves more deeply into the themes, uses, discourses, and contradictions of each kind of moral motivation from the interviews that were conducted following this study.
Results
For the analysis, we first computed means and standard deviations for the MFQ-30 (instructions from moralfoundations.org are for SPSS, but they were adapted for R statistical software in this study), BAS-R, and the five-question modified IRMA. Independent-samples t tests then compared groups by intervention status. Finally, Pearson’s product moment correlations were run comparing the MFQ-30 scales with the BAS-R and IRMA-R. Means, SDs, and results of the group comparisons are found in Table 3.
Means and Hypothesis Testing for Intervention Conditions.
Note. BAS-R = Bystander Attitude Scale–Revised; IRMA-R = Illinois Rape Myth Acceptance–Revised; MFQ-30 = Moral Foundations Questionnaire–30 Item.
Conservative variable was created for this study, and is not part of the MFQ-30.
One-sided t test.
Regarding the moral foundations from the MFQ-30, noninterveners (M = 2.65; SD = .83) scored significantly higher on the loyalty/betrayal scale than interveners (M = 2.35; SD = .93), t(158.95) = 2.21, p = .029. This result had a large effect size (Cohen’s d = 2.53). Of note, noninterveners showed a trend toward “conservative” values (see Graham et al., 2011) with greater but nonsignificant differences on scales of authority/subversion and sanctity /degradation in addition to the significant result for loyalty/betrayal. Exploring this trend, we grouped together the scales on the three “conservative” foundations into a single variable and performed a directional (one-sided) t test. We found that noninterveners (M = 2.73; SD = .83) scored significantly higher than interveners (M = 2.47; SD = .86), t(147.85) = 1.92, p = .020, which had a similarly large effect size of d = 2.81.
An independent-samples t test indicated that the BAS-R was significantly greater among interveners (M = 4.01; SD = .45) than noninterveners (M = 3.74; SD = .53), t(124.37) = −3.32, p = .001, with a large Cohen’s d effect size of 6.62. Regarding the five questions excerpted from the IRMA, noninterveners scored significantly higher (M = 2.24; SD = .93) than interveners (M = 1.89; SD = .8), t(126.64) = 2.73, p = .007, with a large effect size (d = 1.98).
Table 4 provides correlations between the Moral Foundations and the BAS-R and modified IRMA. A medium to large, positive relationship was found between the foundations of harm/care and fairness/reciprocity with the BAS-R, r(180) = .41, p < .001 and r(180) = .4, p < .001, respectively. In a trend that mirrored our finding that “conservative” values were greater among noninterveners compared with interveners, medium correlations were found between the IRMA-R and the foundations of loyalty/betrayal, r(180) = .31, p < .001, authority/subversion, r(180) = .34, p < .001, sanctity/degradation, r(180) = .3, p < .001, and our own conservative aggregate variable, r(180) = .36, p < .001. The care/harm foundation was negatively associated with rape-myth acceptance on the IRMA-R, r(180) = −.23, p < .01.
Pearson correlations of Moral Foundations and BAS-R, IRMAS-R.
Note. BAS-R = Bystander Attitude Scale–Revised; IRMA-R = Illinois Rape Myth Acceptance–Revised; MFQ-30 = Moral Foundations Questionnaire–30 Item.
Conservative variable was created for this study, and is not part of the MFQ-30.
p < .05. **p < .01. ***p < .001.
With regard to the qualitative analysis, we understood that we were not getting the “hot” thinking in the moment, but following MFT, the reasoning, as it portrays core ethical values and the privileging of some ethics over others, gives some indication of the kind of hot thinking that might have occurred. Of the 225 scenarios described, 85% involved either the potential victim, perpetrator, or both being under the influence of alcohol. The majority of cases were males perpetrating against females (215 cases), along with five cases of female perpetrators of males, three with male against male, and two with female against female. Of the descriptions obtained, 59% described a situation where they, as the bystander, intervened. Interventions included confronting the perpetrator, distracting the perpetrator, forcing oneself into a conversation, asking a potential victim if they were okay and/or needed help, phoning the police/security, keeping an eye on a vulnerable person, and notifying the victim’s friends. Only 22% of the time was the potential perpetrator confronted and rarely did this confrontation lead to a physical altercation. In a few examples, the potential perpetrator called the intervener(s) a name. When a friend was the potential victim, participants intervened 86% of the time. This was only 29% of the time when an acquaintance was involved and 41% for strangers. There appeared to be little difference between males and females regarding intervening for a friend, stranger, or acquaintance.
With regard to care/harm, we uncovered this theme in narratives that focused on when the narrator was preventing someone from being hurt or keeping someone safe. For example, in trying to find a friend they had lost during a party one narrator wrote, “We went downstairs to the basement and saw a guy forcing himself onto her when she was clearly very intoxicated.” In another narrative, a person wrote, In my head I was watching the situation happen and hoping the boy was either a really good guy and good friend of hers that was trying to help her out, but there was another part of me that was wondering where her girlfriends were and wishing she had someone to help her get to her room safely.
We saw these quotes as representing a focus on keeping people safe, suggesting an underlying principle that people should not be harmed, and that it is a moral imperative to intervene when someone is going to be harmed. This could be expressed as simply as, “She was intoxicated. She needed my help.” Sometimes this was qualified with a theme of loyalty: “I don’t want to see any harm come to my friends. It is in my ethical code to help out someone in trouble.”
With regard to fairness/cheating, we expanded this theme to include any discussion of injustice. Thus, when a narrative writer discussed issues of equality, or a power imbalance, or even the idea that someone was being “taking advantage” of by someone else, it was interpreted as reflecting a concern with fairness. For example, in one vignette Underage girls were drunk at a party boat that I was on. They had brought their own alcohol and gotten drunk by themselves. An older man started flirting with them and taking advantage of the fact that they were drunk.
The words “taking advantage” also evoke a moral sense of unfairness. Thus the assessment of power and fairness in a situation can lead to a desire to intervene.
An analysis of what a bystander “notices” is important to the above excerpt and it is also part of the social psychological perspective that informs the bystander intervention approach, in which bystanders are taught to notice things that they might not otherwise notice (Banyard, 2011; Burn, 2009). “Noticing” seems to be a discourse that deserves more unpacking than has been presented thus far in bystander theory, in particular, the relevance of noticing a power imbalance.
We also coded any discussion of “consent” as coming from a fairness perspective. It was clear from the vignettes that students in our study were internalizing the legal language of “consent” to make assessments of these situations: “I knew that she would not have consented”; “she was not in the mind to consent to any sexual acts”; “a consent must be given”; “I do know consent is not really consent when given drunk.” While we considered this kind of thinking to be in regard to fairness, it will be important to consider in a more in-depth analysis whether or not there is a rules-based adherence to legal authority that is part of this discourse. If so, it may be that consent campaigns on college campuses can be effective with individuals exhibiting more of a liberal and social justice–oriented perspective, as well as more conservative students who are likely to be motivated more significantly than their counterparts by respect for law and order (Graham et al., 2009).
Loyalty/betrayal themes arose when participants wrote about “bro code” and not wanting to “cockblock;” however, even though loyalty is seen in the MFT literature as a conservative theme, in our vignettes the theme occurred mostly in vignettes where bystanders were looking out for a friend rather than allegiance to a broader group or institution. For example, “She was my friend and she deserved to have someone help protect her if she ever felt unsafe,” and “I wanted to protect and to help my drunk friend . . . I didn’t know who [sic] the kid was forcing himself on my drunk friend.” The emphasis on friendship in these narratives, and the idea evoked that friendship creates a special obligation in the bystander, may be an especially strong moral motivation in the college years. This is supported by the statistic that our subjects intervened in 86% of the vignettes that included a friend. On the other hand, the problem with this moral motivation could be that strangers are subject to more judgment or analysis with regard to their worth, or the trouble it might take to intervene. When loyalty to friends is prioritized, helping a stranger becomes supererogatory, to put this in moral philosophy terms. Supererogatory moral acts are those that are above and beyond duty, but not an obligation. This idea is supported by our finding that only 41% of the stories regarding bystanding with a stranger showed intervening, or approximately half the rate compared to vignettes including friends.
One noteworthy response within the group of responses coded within the loyalty theme brought up the issue of gender loyalty between women, a kind of answer to the “bro code.” She wrote, “I am a female, I hope another female would do that for me if I was really intoxicated.” This statement implies a gendered loyalty based on an understanding of the world of women. Conceivably, a similar logic could be directed toward college-aged men to emphasize the moral basis of protecting fellow men from becoming exploiters or “getting into trouble,” as it were.
With regard to authority/subversion, we coded this foundation when the narrator expressed a feeling of obligation or duty to a greater authority, like God or family, and when they focused on how the potential victim was treated disrespectfully. For example, a coworker followed a female coworker and attempted to grope her while she was drunk. This intervener wrote, “I felt I had a duty to help her out, I would have felt guilty if I did not do anything.” Another wrote of an “obligation to help others in need, there was no one available to help the lady at the time of the incident, inner sense of duty to help women.” This latter quote may represent a conservative value that for some students is at odds with approaches that try to change rape culture on college campuses by emphasizing gender justice and equality. Because of the word “lady,” it evokes a sense that men ought to protect women as a special class of people who might deserve greater protection because of some sort of purity or frailty. While this value might urge some to intervene, it may in some cases contribute to rape culture where only some women are worth protecting and others are deemed “sluts” and deserving of consequences. This was seen in those who used the more conservative sanctity/degradation foundation to understand what they were noticing.
Sanctity/degradation was coded when a bystander expressed feelings of disgust and/or desire to uphold a standard of decency. This included positioning a woman as a good or bad girl/woman as well as referring to guys as “creeps” or “scum.” One wrote, “It’s a scumbag thing to do.” Another narrator wrote, I did note that this one particular girl who I knew was a sophomore and had no business being there because she frankly was way too young. I watched her all night throw back the beers, the shots, and the endless puff of marijuana, and something told me that she felt more pressured to feel the need to fit in than to say no. Guys thoroughly took advantage of her and touched her in places that were not appropriate. I got sick to my stomach, said something to her and then to the boys and stormed out the party.
He continued, “I don’t condone that behavior, bullying someone who is not capable to defend themselves is gross, simply just wrong.” Another wrote, “The men with the girl were obviously creeps and deserved any blowback they got.” These excerpts show that the sanctity/degradation foundation is not only used to judge women who are in precarious situations at parties and bars, and held responsible for their own assaults, but is also used by some to describe the dirtiness and degradation of assault. In the first quote, the man does not intervene to the extent of stopping the incident, even though he sees the girl as under the influence, “too young,” and feeling “pressure” to fit in. But his getting sick to his stomach was a moral response to the taking “advantage of her” where the men are seen as perpetrators.
Although we looked at individual examples of prioritizing one ethic over another, it was often the case that in explaining their in-the-moment reactions, individuals would name several kinds of concerns: She was helpless. [ . . . ] I knew that she would not have consented. [ . . . ] The guy had no right to act like a creep. [ . . .] They were both friends of mine and I didnt [sic] want to see either of them do something they would regret.
In this description of why this person intervened, it seems that the ethics involved are care (helplessness evokes protection), fairness (no consent, and rights), sanctity (he was acting like a creep), and loyalty (friends). It may be that for some students, the greater number of moral foundations that are elicited in the moment, the greater the likelihood they will intervene.
Although not in the original MFT measure, we also looked for examples of the sixth moral foundation of liberty/oppression that was added to the literature more recently (Haidt, 2012). The theme was coded when narratives focused on the freedom of the potential victim being impinged upon or when bystanders themselves upheld the value of privacy (e.g., a concept that people are free to do as they like). One participant, who did not intervene, wrote, “I think my fear was that she would deny any help and that I would be getting myself involved in someone else’s business where [sic] I knew nothing about.” While not intervening has been explained by social psychology as supported by situational variables, the idea of respect for privacy may be an important moral response to consider for those who wish to change attitudes and compel students to intervene in bystanding situations.
Discussion
While research has shown that situational variables as well as personal variables such as gender, sexism, and belief in rape myths affect people’s willingness to intervene, there may be other factors at work. This study explored the moral thinking of bystanders through the lens of MFT. Noninterveners in our study tended to score higher on conservative foundation subscales, with ingroup/loyalty showing a significant difference between interveners and noninterveners. Noninterveners also scored higher on the five IRMA questions and lower on the BAS-R, which gives some indication that sexism is at play.
The finding of no gender differences between interveners and noninterveners may be explained by our recruitment method. We asked for individuals who had been a bystander in a sketchy sexual situation, and it may be that men who answered our recruitment efforts were more likely to be the ones who were proud of intervening and wanted to write about it, which may not be the case for men in the general undergraduate population. This is supported by qualitative analysis being developed in a separate paper of a “hero” discourse among male interveners (see Brodt, Lamb, Gable, & Pagán-Ortíz, 2016).
Although the quantitative analysis revealed a relationship between prioritizing conservative moral foundations and nonintervention, and also liberal values and intervention, the narratives give examples of how foundational reasoning plays out for individuals in their post hoc reasoning around these events. Reasoning that we (and theorists) identified as representing conservative foundations was voiced by both the interveners and noninterveners. Those who prioritized values of purity and sanctity could justify not intervening by describing the potential victim (women) in degrading terms, but they could also consider the obligation to protect a “lady.” Some interveners also used moral language that suggested a value of “purity” when describing how “creepy” or “sleazy” a potential perpetrator was. When bystander intervention programs use language of equality and justice, or talk about unfairness or of potential harm (foundations of fairness/cheating and care/harm), they may be less efficacious with more conservative students unless appropriately adapted. MFT and our research suggest that if one’s aim is to influence conservative students, it may be effective for a program to focus on the indignity of sexual assault or the creepiness of the perpetrator as well. Likewise, for students who favor authority and respect discourses, an approach that focuses on the law, the rules of the campus, respect for individuals, and/or religious, cultural, or family teachings may be more effective.
While those who scored highest on the loyalty/betrayal subscale of the MFQ-30 were less likely to intervene, we read many vignettes in which loyalty to intervene on behalf of a friend was a reason noted after the fact. We also took note of the loyalty argument that women need to help other women. It may also be the case that a more complex conceptualization of loyalty is needed to understand intervening in situ, and thus the meaning of loyalty might extend beyond what could be captured in the MFQ-30. Indeed, in “sketchy sexual situations,” loyalty, if not paired with care or other foundations, might prevent bystanders from interfering with a friend who is about to perpetrate a sexual assault.
Kugler, Jost, and Noorbaloochi (2014) make the case that the three conservative moral foundations are less morally based and more associated with intergroup hostility and discrimination. If this is true, our argument that bystander intervention programs ought to use these foundations in their arguments or campaigns for intervention is a problematic one, and only more justice-oriented and antiharm approaches should be taken. But we have shown how these three foundations can be quite mutable to be framed in the service of preventing sexual assault.
Regarding limitations, the study design did not explicitly ask whether subjects were repeat interveners or noninterveners. Thus, someone who has intervened multiple times may have reported a situation in which they did not intervene, because that situation was salient and different. Because of this, our comparison between interveners and noninterveners is imprecise. Our study also used the thinking of bystanders after the fact, and although we are presuming that moral foundations influence their “hot” reasoning in the moment, and that individuals can report, after the fact, on “hot” reasoning, this may not be the case. If we agree that moral identity (Blasi, 1980) organizes moral values during the college years, and the research design did not suggest it would be better to describe certain foundations over others, then this supports our assumption that we captured an expression of moral values that is close to the identity of subjects that would likely emerge in a difficult situation. To support this view, it is important to note that the descriptions of the vignettes often contained emotional language.
Regarding the BAS-R, although often used this measure combines attitudes and behaviors and these range from intervening in rape to intervening when someone says something sexist. It appears to be a measure of liberal politics, which was supported in our own analysis that found significant correlations exclusively between the BAS-R and liberal moral foundations (see Table 4). Finally, while the IRMA was modified to keep the time brief, it seemed to work for our purposes. Future work should look into the validation of this shorter version.
MFT has shown that both liberals and conservatives use moral motivations based on justice (fairness/cheating) and care/harm, and conservatives prioritize the remaining foundations. If we want to reach all citizens of a university campus, we might want to find ways to meet the more conservative students where they are morally oriented. While changing rape culture should remain a focus for all universities, we have argued that this is difficult and must be approached from diverse perspectives. Addressing all students’ concerns from their own moral viewpoints and cultures is a way to encourage bystander intervention in sexually exploitative and sexual assault situations.
This study attempted a new approach to examining motivation for bystander intervention using MFT. We found that those who generally endorse conservative moral foundations tend not to intervene in sketchy sexual situations. But our qualitative work revealed that this kind of conservative reasoning can be used in ways that might help harder to reach students become interveners. These results indicate that the field of bystander intervention might benefit from adding a moral motivation approach to current approaches. In so doing, bystander intervention curricula would draw on arguments related not only to the injustice of predatory behavior toward college women and men, and the potential harm to victims, but also values of loyalty, authority, and purity in elicit motivation to intervene across a broad student body.
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: Joseph P. Healey Research Grant (P2016) awarded by the University of Massachusetts Boston (2015-2016 cycle) to the awarded to the second author.