Keywords

The previous chapter reviewed what we currently know about bystander action as a basis for our prevention efforts. What we learn is that there are a number of key variables that can be leveraged to promote helpful bystander action across the continuum of sexual and relationship violence. To learn about the conditions for mobilizing helpful bystanders I drew from three different literatures that span a number of disciplines: research on general helping and low cost helping in clearly defined emergencies like assisting someone who has dropped something, helping someone with a task like crossing the street, helping someone with a broken down vehicle or a medical problem; research on situations requiring “moral courage” (Osswald et al. 2010)—more dangerous situations that carry high potential negative costs for bystanders; and finally the growing research specific to bystanders to SV and IPV. What we learned is that there are many common factors for activating bystanders across these situations—efficacy or confidence to act, awareness of the problem, a sense of responsibility to act. Research on SV and IPV more specifically also highlighted new factors that need to be attended to including peer and community norms, attitudes related specifically to violence including victim blame, community collective efficacy and the importance of behaviors and attitudes modeled by community leaders. What we also learned was that most of the research looked at different pieces of the puzzle in a rather disconnected manner. One study may focus on the presence or absence of other bystanders but not look at how that is also connected to variables like awareness or confidence. Effective prevention work that aims to build bystander action requires that we connect the dots in a roadmap that can help us highlight better how these different pieces or correlates fit together.

4.1 MODELS—Pulling Individual Variables Together

Several theorists in the area of bystander intervention connect variables across the who, what, when, why, and where (Dovidio et al. 1991; Latané and Darley 1970), focusing in particular on the proximal building blocks for bystanders in the immediate situation. These are things we can actually influence rather than more distal factors that may be harder to change. This is important for prevention work. The most well-known of these is the situational model of Latané and Darley and this theory forms the basis of much bystander focused prevention to date. Their model does not seek to explain nuances of how a bystander will take action or how this may differ by situation. Rather it seems to outline a more universal template for taking action or not. First, a bystander must notice the situation and label someone as needing help. The bystander must then feel a personal sense of responsibility for doing something about the situation and have the skills to act. At each place in the model, the variables summarized in the previous chapter can help tip the balance toward or away from action.

A second general model of helping is the arousal cost reward model (Dovidio et al. 1991). What this model highlights is that distress in another person causes unpleasant feelings in observers that bystanders are motivated to reduce. Bystanders seek to do this through actions that will most effectively address the situation with the fewest personal costs. This model integrates the “why” with “when” of helping but also suggests that variability in individual characteristics, the “who,” such as empathy will also be important to the model.

What is important to note is that the situational model is built mostly on helping strangers and discrete, one time incidents that present few potential social costs to bystanders. Models of moral courage provide an alternate view that take into account important issues like negative consequences for bystanders (Osswald et al. 2010). Moral courage is a model that describes helping when an injustice is conftonted (such as defending someone being discriminated against or degraded, treated unfairly because of a minority status, or  whistleblowing about unfair or unethical practices). A key feature of moral courage is that active bystanders face negative consequences, face offenders who are threatening, and bystanders are acting for a greater social good. Moral courage has a different set of correlates than more general helping actions according to research (Osswald et al. 2010). For example, the classic bystander effect of diffusion of responsibility was not found for instances of moral courage. Situations of moral courage triggered heightened awareness and greater emotional arousal than general helping. Positive mood enhances helping but moral courage is enhanced instead by anger, sense of justice, and the presence of strong social norms and examples of others modeling morally courageous acts (see Osswald et al. 2010 for a review). Below I discuss how these foundational theories present limitations for understanding bystander action in the context of SV and IPV.

4.2 Limits of Models for Understanding Sexual and Relationship Violence

The uniqueness and complexity of sexual and relationship violence presents challenges to the bystander models summarized above. Indeed, Pozzoli and Gini (2010) described how taking action to help victims of bullying differed from more general prosocial helping behavior. They state, “Intervening in favor of the victim in the context of peer aggression represents a risky behavior, since the helper confronts a powerful bully and, sometimes, even his/her supporters. Given the particular conditions in which it occurs, intervention in favor of the victim of bullying should be regarded as a complex behavior that include not only the positive perception of the victim, but also a ‘moral’ assumption of personal responsibility to intervene from the defender (p. 816)”. Indeed, each component of the situational model has limitations for understanding helping in this unique and complicated context, a point Osswald et al. (2010) made when calling for new models of behaviors they describe as moral courage. The situational model, since it is the most often cited, is used below to illustrate these limitations in more detail.

The first aspect of the model is to notice and label the situation as a problem. Research is clear that more obvious and collectively defined emergency situations produce more ready bystander actions (Fischer et al. 2011). Sexual and relationship violence, however, represent a continuum of behaviors (McMahon and Banyard 2012) not all of which are agreed upon as harmful. For example, many people do not see harassing comments or unwanted touching as serious but rather may define it as playful or flirtatious behavior. Indeed, popular opinion surveys show that people have a difficult time identifying what dating violence is, for example, and editorials, another indicator of community opinions, often call into question even instances of sexual assault (Chan 2007). This context of disagreement about what behaviors constitute sexual or relationship violence creates problems for bystanders being able to notice and accurately label what they see at the start of the helping process. It may cause doubt about whether intervention is needed. For example, the comments that followed a recent New York Times article about bystander intervention (Winerip 2014) and sexual assault revealed considerable disagreement about whether unwanted advances at a party or in a bar were a problem that should be addressed or part of normal young adult sexual behavior scripts. Scholars highlight both historical and ongoing debates about what consent to sexual activity is and when someone is too incapacitated to give consent (McGregor 2005). Sexual and relationship violence is also not one behavior—it is a continuum of risk factors. So we are asking people to become aware of a variety of things at the same time. These issues present hurdles for bystander awareness that are not a factor for other types of helping situations where bystander action has been explored. Indeed, this is one way that SV and IPV present challenges even for the model of moral courage, a model that highlights that situations that call for moral courage are easily and quickly identified.

Bystanders must not only be aware and label the situation as sexual or relationship violence but they must also feel responsible for taking action. Again, community norms can create confusion for bystanders. Flood and Pease (2009) described the persistence of rape myths. These are attitudes and beliefs about sexual assault. While these have most often been researched in relation to their impact on victims (reducing disclosure, increasing shame and self blame) and perpetrators (condoning their behaviors), these myths can also impact bystanders even beyond identification of the situation as noted above. For example, prevalent norms related to sexual and relationship violence are that it is a private matter or that victims are to blame for what happens given what they wear, what they are drinking, where they are walking, or their choice to remain in a relationship (Goodman and Epstein 2011). These notions of privacy encourage bystanders not to see themselves as responsible for helping as the belief is that couples should be left alone to work out their own problems. What is more, rape myths and belief systems may impact the type of arousal bystanders experience. What happens if they see a victim in distress but rather than taking action to help to decrease that arousal, they reframe the situation, blame the victim and thus determine the person is not deserving of assistance—an option that can be encouraged by community norms that support myths about interpersonal violence (Dovidio et al. 1991). In that instance emotional discomfort has been reduced but not through helping. Myths of false reports and of victim blame work against bystanders taking responsibility for stepping in.

Related to this issue of responsibility is the limitation that much of the research on the situational model is based on helping strangers rather than friends. This makes sense since this area of study began in the social psychology laboratory or in controlled field studies where a believable confederate who posed as needing help was required. This confederate needed to be someone the research participant did not know. In the case of sexual and relationship violence, however, much helping involves friends and as reviewed previously, the quality or helping process may look different for friends versus strangers. Certainly the motives for helping may differ if you are helping a friend versus a stranger, though this is a question that has not been explored in the research literature. Most motives for helping are based on having some sort of relationship with the person such that you expect you will receive something in return or feel related to the person in such a way that their well-being will impact your own well-being. It is less clear how such theories of motivation explain helping strangers.

Latané and Darley’s model also explains one time helping where the end point is the bystander action. But when you look at narratives of people’s experience of helping, particularly in cases of relationship abuse, they usually talked about friends and they often talked about a process of helping or a series of actions that unfolded over time (Latta and Goodman 2011). Friends and family described helping over time that included periods of getting involved and needing to pull back. In my own work I similarly found at times that college students described ongoing instances of bystander action in relation to helping friends who were dealing with sexual assault or relationship violence. This contrasted with more general examples of helping described by these same participants, who tended to describe those as instances of helping both friends and strangers with one time, fairly specific needs like borrowing homework, holding the door open, walking someone home from a party or lending class notes. This fits with work by Osswald et al. (2010) who documented differences in correlates of actions taken when there was a strong potential for negative consequences, critiquing models like the situational model for focusing on situations where there were lower potential intervention costs.

A final aspect of the situational model is evaluating one’s skills to intervene. Part of this piece of the model is assessing the costs and benefits to helping. Again, the circumstances of sexual and relationship violence present unique hurdles for bystanders including the wide range of potential actions to be taken, perceptions of bystanders by others, bystanders’ own confidence and skills as well as safety. As noted earlier, sexual and relationship abuse represent a continuum of behaviors. Thus, helping actions are numerous and can take many different forms from more indirect strategies (calling in professional helpers or trying to distract a person to get them out of the situation) to more direct (confronting perpetrators, trying to remove a victim from a situation). This stands in contrast to other types of helping situations (e.g. someone has a medical emergency so you call 911; someone drops their books so you help pick them up; someone’s car has broken down so you offer to help change the tire) that present a more narrow range of potential helping solutions. How do we help bystanders to sexual and relationship violence expand their ideas for how to help and make choices from among these options that do not leave the bystander feeling overwhelmed?

Finally, some of the situations in which we want people to intervene are dangerous (Hamby et al. 2015; Janson and Hazler 2004). How do we consider safety beyond seeing it as a barrier to helping? For example, across forms of victimizations, victims reported that bystanders were most often hurt in situations of sexual violence (Hamby et al. 2015). I found that college students often cited an unwillingness to take action on campus in relation to party situations where underage drinking occurred. They expressed concern that they would be sanctioned for their own behavior like underage drinking, or the behavior of surrounding peers, if professional helpers (campus security, administrators, and resident assistants) were called in. As one student, “they can help without getting in trouble themselves. I think a lot of people are very willing to help if they don’t see themselves getting in trouble in the outcome.” Yet another student remarked, “We can’t really help this kid ‘cause we’re drinking too,” he said. “So we’re all gonna get screwed.”

The issue is that our current models for understanding helping are not complex enough to map the landscape of sexual and relationship violence that bystanders must navigate. We need a more integrated view that pulls pieces from these different models and puts them together to help us see all of the components we have to pay attention to in order to mobilize bystanders or defenders. We need to start to look at all of these in one place, blending ideas across often siloed topics. The result is a model that can help push our thinking about how all types of bystander action happen, though of current concern is how it better positions us to understand bystanders to SV and IPV. A new model will enable the design of more effective prevention strategies to mobilize them.

4.3 Aspects of SV and IPV Bystander Action that Need Attention in a Revised Model

When we review the more specific research on SV and IPV bystanders, there are three key lessons learned that have implications for a revised model. (1) We need to help individuals develop new helping scripts rather than just relying on previous learned and practiced impulses and strategies to help. (2) We need a more relational model of bystander action given that bystanders most likely know victims and perpetrators in the situations they witness. (3) We need to broaden our consideration of contextual factors, the ecological niche in which the bystander action takes place. Below I explain each of these lessons or challenges posed by SV and IPV action. I then present a model that will blend both elements of previous bystander action models with components that better address these new factors.

4.3.1 Challenging What Helping Means: The Need for New Scripts

Active bystanding to help victims of interpersonal violence carries much greater risks for the helper and involves a more complex set of variables that influence whether the situation will improve or worsen. Thus, an added complication for bystanders is that the usual cognitive scripts they use for pro-social behavior and helping may be of little use (Osswald et al. 2010). Researchers have described the importance of these “behavior social scripts” (Avery et al. 2009). They describe these as guidelines for how one should usually behave in different social situations and assert that some scripts are clearer or better-known than others. Being in situations without a good script can cause anxiety and concern about what to do as, “in unscripted situations, there are fewer norms and cues available guiding individual choices regarding appropriate speech and behavior. This lack of situational signals renders common heuristics governing interpersonal interaction ineffective. The result is a high degree of uncertainty concerning what to say or do (Avery et al. 2009, p. 1389).” To date, however, questions such as these have not been discussed in relation to bystander action for SV and IPV and how prevention work needs to prepare individuals to handle this mismatch between how they usually help others and what the SV or IPV situation may require.

Recent qualitative studies (one on a college campus and one in a rural community) suggested that how individuals described how “helping” in general happens in a community and how helping in relation to SV and IPV happen were quite different (Edwards and Banyard 2014). For example, one college student described, “I wanted to step in and try to talk, but I didn’t know what was going on or what was being discussed so I just didn’t, because I didn’t know my place in the situation.” What is interesting is that in response to questions about helping in general, individuals described small kindnesses, giving someone directions or helping with homework, as well as more significant time investments in listening to a friend’s problems or organizing a community fund raiser for someone with medical bills to pay. When asked about helping in situations of sexual or relationship violence, however, participants often said they were not sure how to help and were concerned about what was best to do or were concerned that helping related to relationship abuse made something private too public or that it might cause negative effects for those who try to help:

That aspect of not wanting to get that public attention drawn into a negative situation, shining a bad light on the family. Everyone in town is known by everyone else, so a negative situation would also be known to everyone else….There’s no education for the community to understand what to do in those situations.

I think a lot of people don’t help either I mean they just are afraid to get involved, they don’t want to poke their noses into other people’s business if they don’t have to because they’re afraid they’ll get sucked into it as well. Because I mean if they get involved you know then they might possibly become victims of ridicule and rumors themselves.

They described trying to help more privately by helping friends find safety or by trying over time to figure out what the victim needed.

Such descriptions suggest that there may be a schema mismatch between how individuals think about helping more generally, and the helping situations they are confronted with regarding sexual and relationship abuse. Sources of this mismatch include that there are a number of helping strategies or options and that bystanders may often need to try a series of actions to end the risky situation. Indeed this is also the promise of knowing perpetrators and victims that bystander action may not just be a one shot opportunity to help but may provide opportunities to loop back and try again or try something different. Bystanders may get second or third chances to do or say something. Further, bystanders may need to try multiple strategies and likely make choices based on the goals they have for trying to influence a situation, what have been termed “different communicative intervention strategies (p. 95)” (White and Malkowski 2014). The prevention curriculum designed by GreenDot, Etc. describes the “3 D’s of bystander intervention” as “Direct, Distract, and Delegate” (GreenDot, Etc. 2015) and to which I would add, “distance.” These categories describe the variety of ways that bystanders can choose to act—by confronting perpetrators, approaching victims, creating distractions that diffuse the situation, getting others to help, or finding ways to remove the potential perpetrator or victim from the situation. This is a greater set of choices then the array of possible actions for helping someone across the street or assisting someone with a medical emergency and thus will require more complex and flexible helping scripts.

Beyond types of actions, bystanders can also take on a variety of roles in relation to SV and IPV, what we might call action roles. This not really been discussed in the bystander literature. Research on bullying uses the term “defenders” to separate bystanders who step in on behalf of victims rather than those who may join in the bullying (Pozzoli and Gini 2013). This defender role seems to capture well the series of actions described in the paragraph above. They are the people who are in the midst of the at-risk situation. As discussed, earlier, however, there are other situations where bystanders are needed and other roles they can play. For example, “supporters” are those bystanders who receive disclosures from victims and are in a position to promote recovery or increase distress through their reactions. They can also be supporters to other bystanders. Bystanders can also play a role as witnesses and may be interviewed by law enforcement or be required to testify in court or in campus judicial procedures (Buzawa and Austin 1993; Shernock 2005). Bystanders can be “resisters” or “dissenters.” We know from research that risk factors for SV and IPV include peer and cultural norms that support the use of coercion in relationships. Dissenters are those who actively speak out against these norms, challenging, for example, comments like the use of the term “rape” to describe a difficult exam in school. These bystanders exercise media literacy and use their voices to challenge media images of interpersonal violence as well as harassing comments (Ryan et al. 2006). Indeed, changing social norms requires in part these dissenting messages from multiple people who model a different descriptive norm—that not all people agree with rape myths or support violence and aggression—in a community, and who can work to change media coverage of stories on SV and IPV that influence how people think about these issues (Bowen et al. 2004; Franiuk et al. 2008). A related role is bystanders as spokesperson/trendsetter. In this role bystanders are change agents or change leaders. They actively pursue more education and awareness about SV and IPV, not just being reactive and responding to problematic situations but demonstrating support for violence prevention—displaying images such as slogans supporting violence prevention, posting positive stories about bystander action on social networking sites, talking to friends, family members, one’s children about what healthy relationships are, encouraging community leaders to make statements related to violence prevention, volunteering for violence prevention or at agencies that support victims. These are the group of “innovators and early adopters” discussed by Rogers (2002) as the 15 % of people who embrace new ideas ahead of others and become role models for new community norms not just reacting against old ones. It is likely that the correlates of these different roles and the pros and cons of being in these roles will differ.

Lacking cognitive frameworks for how to help that work across these roles and actions may leave even motivated defenders immobilized. Understanding this conceptual mismatch between the way we usually think about helping and the resources we draw upon to do so and the helping scripts we need for sexual and relationship violence may provide new ideas for prevention. As will be discussed further below and in next chapters of this book, one solution to the mismatch of helping scripts is to ensure that prevention strategies include time to articulate and practice new helping scripts that can be flexibly used across bystander roles.

4.3.2 Emphasizing Relational Components of Bystander Action

Just as victims often know their perpetrators, given the context of SV and IPV, bystanders often know victims, perpetrators or both. In several studies of SV the position of the bystander in relation to the victim and perpetrator makes a difference for potential bystanders themselves in terms of attitudes and intention to help (Bennett et al. 2014). Bystanders, according to this work, are considering their position in relation to who needs help or who is creating the risky situation (Nicksa 2014). This raises key questions for our model of bystander action including: What do we need to add to a model that mostly describes helping strangers rather than helping friends? How do we better understand the impact of consequences of bystander actions given their connections to those involved and how those consequences (positive and negative) impact future attitudes and behaviors?

4.3.2.1 Creating New Scripts and Harnessing Potential Over Time: Moving Beyond Linear Views of Bystander Action

Helping in situations of SV and IPV may not be a one point in time event. We know that friends are the most likely individuals to receive a disclosure of SV (Banyard et al. 2010). Friends or family members may witness abuse consisting of many different incidents and behaviors and patterns of control over time in a relationship. Friends and family  interviewed in one study described a process of becoming aware of the abuse in the relationship, trying different strategies to engage and help the victim, and even periods of disengagement from helping (Latta and Goodman 2011). Even sexual violence, which can be a single incident, may involve multiple occasions in which victims need help or support as they choose to disclose or not, seek formal services, or file a complaint with authorities. For example, a college student described the challenges of helping a friend in an abusive relationship, “Um well I’ve heard about it for a couple months now so like it’s kind of hard to say the same things over and over and she doesn’t seem to get the message but I still feel like I have to try to help her.” There are multiple and yet linked opportunities for action by bystanders (Latta and Goodman 2011). While Latané and Darley’s research does not describe helping specifically as linear, more recent uses of their work does seem to conceptualize this more step-wise path. We need to develop a model of bystander behavior as a process that may loop and coil over time to better account for the often relational context of bystander action in SV and IPV. We need a model that more transparently and intentionally describes the relationship between bystanders and other actors in the situation, between different correlates of bystander action and that follows these over time. Such a view could help potential bystanders anticipate these connections. For example, as noted in Chap. 3, we know very little about what is helpful to bystanders regarding helping friends.

From a research perspective we need more longitudinal studies of bystander action. How do actions link together over time? How often do bystanders help friends once with a single incident and how often does their helping consist of a series of actions over hours or days or months? Does the concept of linearity really apply or do bystanders find themselves going over and over the same actions, rather than through steps set out in a straight line? Do they revisit particular steps but not others as they help friends at different points along the continuum of risk for sexual assault? Research to date is relatively silent on these questions.

4.3.2.2 Bystander Behavior Doesn’t End with Taking Action: Starting to See Consequences

Very little is currently understood about the impact of helping, both intended (did the action help?) and unintended (did the bystander get hurt or experience retaliation of some kind?). The relational nature of action to address SV and IPV foregrounds the issue of consequences. While Ullman (2010) has extensively studied victims’ perspectives of reactions they receive to disclosure, we do not know about victims’ perceptions of the range of bystander actions, nor have we studied bystanders’ own perceptions of what happened once they intervened. Did they make the situation better? Did they make it worse? Did the identified victim feel positive about receiving help? Did the identified perpetrator retaliate? What about others who might have observed what went on and who might support the bystander or retaliate against them? Research on bystander behavior has considered how bystanders weigh potential pros and cons to stepping in as well as investigating barriers to action and highlighting types of helping that are likely to carry negative effects (Banyard 2008; Bennett et al. 2014; Dovidio et al. 1991; Osswald et al. 2010). These barriers or costs, however, have been studied mainly as perceptions of consequences or anticipated consequences when we talk about consequences at all and even these have not been well catalogued. We know little about describing actual consequences or how a bystander’s status or position may influence those. For example, what are the actual negative and positive results of bystander action? As one student in my own study of bystander behavior remarked, “‘What are people going to think about me?’ Um, you know, ‘What if the people involved know my friends so they’re going to talk about me?’ So, definitely self-image, what other people think about you, um, I think just fear is a big thing and just not wanting to get involved. Not wanting to get into a mess.” Another stated, “I kinda have just a really strong trust with the majority of my friends and I also don’t want them to be harmed so I would put myself out there for a friend but on a stranger, I would really, I would be a little hesitant only in the sense that I would wanna scan the situation to make sure that I would not be put in harm.”

While bystanders in many situations may have questions  about consequences, there are also likely unique consequences for SV and IPV that have not been well explored. Discussions of moral courage note that what distinguishes this form of helping are the potential negative consequences for bystanders but what those consequences look like are not specified (Osswald et al. 2010). One of the few studies of bystanders and multiple forms of victimization found that bystanders can be harmed by stepping in. Sexual assaults seemed to incur higher rates of bystander harm compared to other forms of victimization at least as reported by victims (Hamby et al. 2015). What is more, bystanders to SV and IPV may be more likely to know the victims and perpetrators involved in the situation since SV and IPV is most often perpetrated by someone the victim knows and in social and neighborhood contexts where bystanders are known as well. These relationships raise the stakes as bystanders may experience including harm to friendships, working relationships, or team unity.

We still know little about what bystander actions are the best to include in someone’s skill set. Better understanding consequences of bystander actions will permit a more nuanced approach to training bystanders as potential prevention allies and helping bystanders stay safe from physical harm and other consequences (including more emotional consequences to relationships, socials standing, or sanctions by the community for a bystander’s own behavior such as underage drinking). For example, one college student noted; “They were fighting it was late at night and he-he got very physical with her and at that point just kind of like “hey”, you know, “are you alright?” And um I was walking towards them anyways to go past and um yeah I stepped in between them and um then they started turning on me…so I just kinda I-I walked away and called the campus cops. The girl initially turned around and said uh you know don’t help me…That kind of made me kind of question whether or not I should of done it.”

Future research is needed to answer the following questions: What negative and positive thoughts do bystanders have not only before taking action but while taking action? How did the people directly involved (the victim and perpetrator) react? Was the victim relieved? Was the perpetrator angry? What were the reactions of other bystanders in the situation? Were they supportive? Dismissive? Did they have a negative reaction to the bystander trying to help? How did the situation turn out? Was harm avoided for the victim? Did anyone get in trouble? Was a relationship damaged? Were supportive or derogatory social media posts made? Did the bystander end up being interviewed or called as a witness to a crime? Most importantly, how does experiencing any of these consequence then feed back into future decisions to take bystander action or not?

4.3.3 Embracing a Larger Ecological Model: Revisiting Community and Cultural Factors that Influence Bystander Actions

A revised model explaining bystander action also needs to more intentionally consider context. We know that a variety of community factors are a part of why violence happens and community interventions need to be part of the prevention equation (Casey and Lindhorst 2009; Pinchevsky & Wright, 2012). We also know that pro-social behavior or general helping is influenced by broader cultural and societal factors that are not captured in the microcosm of the immediate helping situation. For example, levels of helping differ by country and culture, by rural versus urban areas (see Banyard 2011 for a review: Pozzoli et al. 2012). More specific to bystander action and relationship abuse, actions are higher in communities with greater resources and greater collective efficacy (Edwards et al., 2014). At the more micro-level, teacher and school attitudes and parent and peer norms affect bystander behavior for bullying, and students are more likely to report a risk of violence on campus if they have greater trust in campus authorities (Hektner and Swenson 2012; Pozzoli and Gini 2012; Sulkowski 2011). For example, tests of Latané and Darley’s situational model among primary and middle school students for defending against bullying showed the need for an expanded model to explain patterns in the data (Pozzoli and Gini 2013). In particular, inclusion of measures of peer and parent norms were important such that perceived pressure from parents and peers to help support victims predicted both attitudes and actions to defend the victims (see also Rigby and Johnson 2006).

How can we expand our understanding of the context of bystander intervention to better include different communities and ecological niches? In chapters two and three we saw how bystander research has looked mainly at the intra-individual and relational levels of the social-ecological model or the theory of triadic influence. We have done little to really investigate other aspects of context that might matter. Exploring these questions and their potential answers provides the foundation for an expanded model of bystander action. There are three components of a closer consideration of context for bystander action. The first is a consideration of the relationship of the individual bystander to his or her community and how the intersection of her/his identities in that community create different potential consequences or action choices (for example, is the bystander underage for drinking alcohol versus of legal age, is the bystander from an underrepresented group in that community?). The second is developmental context. The third is ecological niche or how cultural and community norms that may vary by ecological niche open possibilities or present restrictions for different bystander actions and how we help potential bystanders consider those.

4.3.3.1 Considering the Position of the Bystander: Social Position and Power not Just the Presence or Absence of Others

Knowing what to do or say and choosing to do so may depend on who the bystander is. As can be seen from the review of the bystander literature in the previous chapter, to date models of bystander action generally ignore the position/location or status of the bystander (one exception is Osswald et al. (2010) who discuss the social position of the bystander or type of government under which they act as aspects of moral courage but this discussion is not detailed and describes a more macro-sociological level than what I explore here). Whether the bystander is in the presence or absence of other bystanders is more often the focus of what is studied as is what is going on inside the head of the bystander. Some studies have considered the position of the victim (member of one’s in-group or out-group, member of a racial minority) (Laner et al. 2001; Levine and Thompson 2004). Little considered is the position or status and power of the bystander themselves. In part this may come from the fact that early studies of bystander action were conducted around helping strangers, individuals who would be unlikely to know the status of the helper. Even in this literature clues exist about the relevance of social position. For example, researchers found greater defending by secondary school students with higher social status among their peers (Caravita et al. 2012). It may be easier to step in if one occupies a more high status or high power position, such as being older students on a college campus. In my own study, students described how a combination of experience, knowledge, but also power (e.g., being of legal age for drinking) contributed to this increased comfort. For example, “I definitely would be more willing [to help] than I used to be. I haven’t been in many situations where that would come into play but I would be more than willing than I would freshman year… Being older comes with maturity completely.” Another student remarked, “If I have any difference in opportunity to help, it’s not because I’m older. It’s because I’ve joined more clubs—I’m an officer in several clubs. I’m in more of a position of power so I can help.”

To date, power, position, and status have been missing in a consideration of the context of bystander action in general and almost never explored in relation to sexual and relationship violence. There are many different ways that people can have power or be lacking in power. Bystanders have their own position in their communities—they have access to power or are lacking power in different circumstances given their membership in different social identity groups, their visibility as community leaders or not, how long they have been in the community. For example, in one research study about whether bystanders were likely to report child neglect, more highly educated community members and those in roles where they might come in contact with child maltreatment (teachers, medical professionals) were more likely to take action to address suspected child neglect (Fledderjohann and Johnson 2012). The impact of differences in position needs to be researched more carefully and a consideration of a bystander’s position of power or lack of power should be thought through in prevention training.

For example, we can hypothesize that lack of access to power including being part of an underrepresented group in a community, presents particular barriers to helping because it may make a bystander vulnerable to more negative consequences. A bystander who is a member of an underrepresented group may step in but then be erroneously labeled as part of the problem or one of the perpetrators. A bystander’s position may make her or him more or less likely to be supported and joined by others in taking action. Thus, a bystander’s position in his or her community likely impacts the range of action options that seem possible or safe. For example, what are the implications for students from underrepresented cultural groups who choose to step in on a campus that is predominantly Caucasian? How are sexual minority students supported or seen as bystanders? Are individuals from marginalized groups or underrepresented groups at risk for more negative consequences or retaliation within the community when they step out as bystanders? Are they less likely to be supported by others? Are they more likely to be mistaken for being perpetrators rather than bystanders? What is more, a bystander’s position in the community likely influences their sense of trust in other community members and in formal community helpers. To the extent that a bystander feels isolated or marginalized, they may be less likely to call police or campus authorities to help, or may feel less able to call on other community members to join with them in taking action. One study of college students found students with lower trust in campus authorities were less likely to say they would report suspicions of violence on campus (Sulkowski 2011).

4.3.3.2 Including a More Thorough Consideration of Time as Part of Bystander Action Processes

We also need a more developmental view. We may ask, what set of variables or resources, perhaps collected earlier in the lifespan, are needed to form the foundation for bystander action that may be taken later? How do we take a view of bystander action as well as a view of prevention that extends across the lifespan? Do bystander behaviors and the variables that impact them look different across the lifespan? Research suggests that helping skills develop over time, are influenced by developmental contexts like parenting (Carlo et al. 1999) and that certain aspects of bystander prevention may be more or less appropriately developmentally timed. Related to bullying, bystander intervention training seems to be more effective for high school students rather than younger students (Polanin et al. 2012). Is this finding due to age or developmental effects or is it that older high school students have higher social status and are more embedded in their communities? One college student described it this way, “I do feel a little more… Due to my experiences, I do feel like I’m a little more easily swayed to help a situation… I do feel more comfortable intervening after four years of being here [on campus].” How do we explain these developmental trajectories and how does that help us potentially link strategies to develop active bystanders over time?

We can start with a broad developmental question—how do we learn how to help? It seems to come in part from family contexts, we learn from our parents and caregivers; positive supportive parenting has been linked to prosocial tendencies (Carlo et al. 1999), is also linked to empathy and perspective taking (Espelage et al. 2012) and to moral development (Thornberg and Jungert 2013). Indeed, researchers of child development discuss the importance of early nurturing contexts for prosocial behavior (Biglan et al. 2012) while others talk about educational contexts that promote social and emotional learning (SEL) (Dulak et al. 2011). Studies show that SEL skills can be taught and are especially effective when taught by classroom teachers—perhaps because there is more diffusion of the innovation when actual school personnel are trained to do it (the personnel are then trained and may alter other aspects of classroom and their work, not just the curriculum presented to the students on SEL) (Durlak et al. 2011). This work suggests that we start developing active bystanders early rather than waiting until late adolescence or early adulthood.

Research also considers how bystander action may look different at different points in time, though the variety of measures used and variables studied makes comparisons challenging. We can also ask how correlates of helping may change over time. Studies of defenders in bullying showed the importance of moral responsibility, which might be similar or linked to to the sense of responsibility found by those who study college students (Banyard and Moynihan 2011; Pozzoli and Gini 2013). Children in middle school samples also showed effects of peers on bystander action, similar to the peer influences found among college students (Fabiano et al. 2003; Brown et al. in press; Pozzoli and Gini 2013). These results suggest patterns of developmental similarity in the correlates of helping. However, studies also show that models of defending behavior among bystanders to bullying did not fit younger children and adolescent samples equally well suggesting developmental differences (Pozzoli and Gini 2013). There is still much we need to understand about how our models of bystander action may look similar or may need to be different across the lifespan.

4.3.3.3 Ecological Niche: Community Norms for How it is Appropriate to Help

Finally, chapter three summarizes a few studies that support cross-cultural differences in quantity of helping. What is less understood is how communities and cultures vary in how helping happens or in what types of helping actions are seen as most appropriate. Prevention programs involve limited time to train participants. Thus, better tailoring the bystander action strategies that are practiced and demonstrated in prevention programs to match the range of options that are culturally salient to particular communities might increase prevention effectiveness. One ecological niche variable that has been more studied is gender and it can serve as an interesting example. We know from research that women and men often differ in what sort of situations they feel most comfortable helping in. In my own prevention work I have found, for example, that women discuss using distraction or distancing to address a risky situation while young men describe more direct intervention. In spite of actions they have taken in the past, young women ask for more opportunities to practice skills in prevention sessions to build confidence as active bystanders. On the other hand, young men need more encouragement to develop and use skills for diffusing, distancing, or delegating rather than just directly jumping into a situation that could be unsafe.

4.4 Re-envisioning a Model of Bystander Behavior: From Helping Factors to Bystander Action Coils

Prevention efforts will be improved to the extent that we can better hone bystander action skills by being able to map out bystander actions that are most likely to be helpful and to minimize danger and costs to all in situations across different ecological niches. The process of helping can be expanded by envisioning notions of feedback loops, what I describe as bystander action coils, instead of a linear, single incident series of steps or stages (Chaudoir & Fisher 2010). The analogy of coils can better describe bystanders who help individuals with whom they have relationships over time and across situations. I draw this notion of process, and organization of some of its components from work by researchers confronting similar limitations in understanding how people choose to disclose or conceal a stigmatized identity (which includes the identity as a victim of SV or IPV) (Chaudoir & Fisher 2010). While Chaudoir and Fisher’s model is specific to disclosure, not bystander action, it provides an interesting scaffolding for an expanded bystander model. I draw also from the Haddon Matrix model for analyzing injury prevention in public health (Barnett et al. 2005; Runyan 1998, 2003). The Haddon Matrix visually outlines a number of key dimensions of prevention including a consideration of factors across the ecological model (intra-individual variables, aspects of the person or vehicle of the problem, aspects of the immediate physical situation, and broader norms that make up the social context) and across time before, during, and after the injury event. Figure 4.1 presents a picture of what a revised model of bystander action could look like. The model blends work from across research on helping and bystander action in general with SV and IPV specific research. It more intentionally describes a process model of bystander action. It borrows the organizing framework from Chaudoir and Fisher (2010) and from the Haddon Matrix (Runyan 2003) but adapts it for cataloguing in greater detail the additional components of helping that make bystander action for SV and IPV so complex.

Fig. 1
figure 1

Bystander action coils

Full size image

The revised bystander action coil model includes specifications of the decision making process; contextual factors that impact decision processes as well as how the event may be related to outcomes; characteristics of the event itself; and outcomes of bystander action. Each of these represent a loop in the coil of bystander action and these loops potentially repeat over time developmentally as new skills and experiences are added to one’s resources, attitudes, and behavior toolkit. The idea of multi-component loops allows for more clear specification of additional factors like the bystander’s social position and context that while likely relevant to all bystander actions, are made particularly salient when considering SV and IPV. The first loop of the coil (action coil 1) represents an individual’s internal decision making process. It is reviewed in detail in Chapter three as it includes the key components from previous bystander research using the situational model and in this way is not new (Latané and Darley 1970). This includes both the who of bystanders as well as motivations or why people help (to promote relationships or to act in a moral way that a person sees is “right,” to improve one’s self esteem or contribute to one’s sense of self as a helpful person) or avoid action (desire to avoid risk to safety, to relationships). This component also includes an analysis of costs and benefits also described in other models (Dovidio et al. 1991) including consideration of barriers that are specific to SV and IPV situations (Bennett et al. 2014; Burn, 2009).

A variety of contextual variables or processes, the second action coil, can also be described and are in need of more research. This coil includes many of the factors described above as key components of an expanded model of bystander action that comes from research on SV and IPV. For example, perceived peer norms, whether related to the acceptance of coercion in relationships or norms about the appropriateness of helping, impact bystander action (Brown et al. in press; Brown and Messman-Moore 2010; Fabiano et al. 2003). Cultural norms and contexts are an extension of this and also impact individual decisions about whether and how to help (Pozzoli et al. 2012) and need to be a more explicit part of models of bystander behavior. They are included in this contextual processes coil. Expanding beyond friend norms, perceived relationships with community leaders, particularly trust in authorities may also make it more likely that bystanders will enlist the help of formal helpers such as school administrators (Sulkowski 2011), though to date we know little about what promotes this trust. The perceived severity of the situation also seems to impact attitudes and behavior (Bennett et al. 2014; Fisher et al. 2006) as are a bystanders’ own past experiences with helping. This action coil should also include a consideration of the social position of the bystander. The position of the bystander connects to other coils in that it likely impacts all aspects of the decision process, mediating processes, and how the event itself unfolds.

The next component in the model (action coil 3: the event) is the coil that makes up the event itself. This encompasses the how of bystander action—what individuals choose to do and how they choose to act. Finally, this component also includes a consideration of the when and where or the context of the situation—the position of the bystander (how much power do they have in the situation? In the broader community?), the number of bystanders present and the extent to which these bystanders know one another (Levine and Thompson 2004). It includes relationships among the parties in the situation, such that a bystander may feel safer to intervene if they know both parties but more likely to help if they only know the victim (Bennett et al. 2014). This coil also takes into account that bystanders play different roles depending on the type of SV and IPV they encounter (McMahon and Banyard 2012).

Finally, the fourth coil is about the outcomes of bystander action. This coil includes both current outcomes in the proximal situation but also past experiences of bystander action that likely impact choices and actions in the present. We need to know much more about the consequences of bystander action for victims (are they helped? Harmed?); for perpetrators (do they stop their behaviors? Is their behavior sanctioned in some way? If so, by whom or what?); and for bystanders (are they retaliated against for their actions?). Hamby et al. (2015) found that having a helpful bystander present was associated with more positive mental health outcomes for victims but that these positive effects were also more likely reported by victims if bystanders were not perceived as having been harmed. Indeed victim outcomes were less associated with the presence or absence of a bystander and more about whether the victim perceived that the bystander was harmed by being in the situation. Hart and Miethe (2008) conducted interesting analyses of helping ratios. They examined in what contexts across different types of crime bystanders are likely to be most helpful. For example, high helping ratios were for robberies and sex offenses that took place in daylight and involved strangers. Having a model that specifies consequences can promote further research on questions like: How do bystanders’ past experience of helping affect present and future action? Are bystanders who experience negative reactions from victims, perpetrators or onlookers less likely to help in the future? Are they likely to keep helping but try a different type of action? What actions produce high positive outcomes for SV and how do these differ for IPV?

More information about all of these coils, and placing what we know about bystander correlates together in this process model will enable us to better train and advise potential bystanders about costs and benefits of helping, how best to evaluate their own safety, and generate actions that may work from their particular location.

4.5 Key Points Summary

  • Research on bystander action often implicitly describes it as a linear process.

  • Missing from current research is in-depth consideration of how helping unfolds and changes over time, including how consequences impact future helping.

  • Examining instances of SV and IPV that bystanders confront makes clear that the relational context of helping is important.

  • While the social-ecological model is a foundation for violence prevention work, the broader levels of bystanders’ ecology have been less considered.

  • The current chapter creates a new scaffold for bystander action research, “Bystander Action Coils,” that is likely applicable to all forms of bystander helping but better nests all of the factors that we know are particularly important for SV and IPV responses.

4.6 Practice Implications

A number of practice implications and questions follow from this expanded model of bystander action. It calls for practices that consider the position of the bystander, developmental trajectories, and broader community contexts. Many of these will be more fully addressed in Chapters five and six. First, the new model suggests that we should talk to prevention participants about action as a process that may look different in the case of sexual and relationship violence from what they usually think of and do for helping. We must help participants anticipate these differences and identify new skills and scripts they may need to have. Helping is a process that may require periods of engagement and disengagement as friends experiencing SV or IPV may need different things at different times.

We need to understand much more about the outcomes of bystander actions, including negative repercussions and the variables that might mediate these outcomes. Are certain groups of people in a community more or less likely to be supported for taking bystander action? These topics can be part of discussions when training bystanders. Bystanders may need space to come back and have conversations about negative consequencees they experienced or actions they tried that were more or less successful. Following from this, we need to ask what more we should be doing to provide ongoing support and safety nets for bystanders. We know that helping in the context of sexual violence and relationship abuse is complicated, what policies and community resources can support these actions? Bystanders need the chance to build skills over time but also to loop back, check back in and discuss and receive support and further training once action has been tried and consequences experienced. This requires more than one-time training.

More linear bystander models that focus on getting potential bystanders from awareness to action (even if some looping back in thought processes happens in the middle) translate into prevention programs that are more closed ended, one point in time trainings. What the action coils model suggests instead is that we need to move from encouraging bystander action to cultivating bystander tenacity. Instances of relationship abuse or stalking intervention may take time. Many different actions may need to be tried over time and bystanders will not meet with success each time. We do not want bystanders to become discouraged especially since successes may not be clearly visible. We want bystanders to think about helping as a process over time, to think about meeting victims or potential victims where they are in one moment while being prepared to help again. Prevention resources need spaces to connect back with bystanders over time after initial education efforts. Ongoing training fits better with an action coils model.

We seem to assume that bystander intervention is helpful. But this may not always be the case (Hamby et al. 2015; Planty, 2002). Bystanders may take action that is not helpful to victims or actions that bring retaliation or negative consequences onto themselves or other bystanders. An expanded bystander model sets the stage for research on consequences of bystander actions so that our prevention efforts can more carefully advise potential bystanders about how they may best be helpful, how they can best minimize risks to themselves and others rather than taking the approach of just increasing action at all costs. We need to see which programs already include discussions of bystander safety in their programs (i.e. Eckstein et al. 2014) and whether what they are doing is enough or if the focus on safety ought to be a more expanded part of programming. This is especially important as more and more online bystander programs are created where direct interaction with participants is reduced and as campuses are mandated to include some form of bystander training in response to amendments to the federal Violence Against Women Act legislation.

We need to recognize that bystander action does not take place in a vacuum. It is impacted by contexts including community norms and policies. We should be attending as much to making changes in those areas as we are to changing the skills of individual bystanders (Lippy & DeGue 2014). Bystander action, like any other pro-social behavior, also does not come about overnight. It must be built on a foundation of positive development that can be strengthened by interconnected prevention efforts across the lifespan (Banyard 2013; Hamby and Grych 2013). For example, Peer Solutions (Peer Solutions 2010) takes a comprehensive, youth development approach to prevention that builds to bystander action through a series of interconnected protective factor strengthening components not all of which are bystander focused. More detail on specific prevention implications are discussed in chapters five and six. The purpose of this chapter was to show how an expanded model of bystander action in relation to sexual violence and relationship abuse can improve the logic models we use to describe both why we think bystander prevention may work and what components of our prevention tools are needed for it to be effective.