Keywords

Looking across the research literature on bystander behavior we find a number of theories and empirical studies that explain helping and bystander action. Variables that inhibit or facilitate action become the building blocks of our logic models and learning objectives for prevention curricula and tools. However, to the extent these theories and research questions are limited; they may also restrict our impact. In this chapter I summarize key empirical literature about the who, what, where, and when of bystander action and helping. These factors include cognitions, emotions and intentions that make up bystander attitudes and influence a range of bystander behaviors, the “attitude system” of bystander intervention (Zimbardo and Leippe, 1991, p. 33). They also include important relational/social and contextual factors (Levine et al. 2012).

Each section begins with a review of the helping and bystander literature more broadly (including the study of helping in emergency situations, informal control of criminal or norm violating behavior including bullying, and the study of relationships) and then summarizes available findings that are more specific to sexual and relationship violence. In order to best organize this broad range of findings I draw from the social-ecological model of Bronfenbrenner (1977) that outlines many different levels at which variables may influence human behavior (factors within the individual, factors related to close relationships with family and peers, factors related to local settings including workplace, school, and community, and finally aspects of the wider society and culture) (see Banyard 2011 for a review and use of this framework to organize research on bystander intervention). I also use Flay et al. (2009) Theory of Triadic Influence that starts with the ecological model but goes beyond it as well.

Flay et al. sought to integrate different theories of health behavior change under one model, identifying different layers of variables that would pertain to any given health issue. This model begins with the Social Ecological Model, grouping causal factors under intra-personal, social-situational, and cultural environment/community headings (p. 455). Under these headings however, is a consideration of time, with some variables being farther away from the current decision to engage in a behavior like helping. These factors are what Flay et al. call “underlying causes and predisposing influences.” These set the foundation for the behavior but are influences that were likely set in motion earlier in an individual’s history. The best time to address these variables and shape the healthy outcomes we would like to see is when they are developing, through infancy and early childhood prevention work (for example via social emotional learning curricula that promote empathy and perspective taking; Durlak et al. 2011). Flay et al. also describe “proximal immediate predictors (p. 455)” that are the variables that affect behavioral choices more immediately in the moment including temporary situational characteristics. For example, being one of many bystanders in a large crowd leads to diffusion of responsibility and less bystander action (Latané and Darley 1970). In what follows in this chapter I try to indicate factors related to bystander action that are both distal and proximal and that span the ecological model.

Prevention is much easier to do with factors closer in time to when the attitudes or behaviors develop. That is, it is much easier to change something that is in the process of developing rather than a belief, behavior, etc. that has become an ingrained habit for an individual or community. Thus, considering the range of variables related to intervention can help us target different leverage points at different places in the lifespan. For example, bystander intervention for bullying among school children may work best when leveraging aspects of moral development that are part of what is developing at that point in the lifespan. Work with college students might need to instead harness motivations related to taking care of relationships, fitting in with peer norms, or forming an identity as a helpful person, developmental concerns for that age group that may be a better source of prevention motivation. This chapter dissects these distal and proximal factors in the small compartments of who, when, why, and where bystander actions occur.

3.1 Who

When asking the question about “who” in bystander intervention we can look both at who is the bystander—examining personal characteristics that may make them more or less likely to step in—and who is the person needing help—characteristics that may make receiving help more or less likely. Both have been the focus of research though most studied have been general characteristics of prosocial people with minimal attention to who is helped.

3.1.1 Who Helps?

3.1.1.1 Lessons from Helping in Contexts Other Than SV and IPV

There are several different layers to understanding who provides help. At the innermost layer, evolutionary theories and research on biological and genetic foundations of helping suggest that the foundation of helping behavior may be hardwired, particularly via empathy (Penner et al. 2005) suggesting both that all individuals have the capacity to help and that biology may explain some of the variation we see in how much people help others. Carlo and colleagues (Carlo and Randall 2002; Carlo et al. 1999; Eisenberg et al. 1999, 2002) define helping as related to personality, what they call “prosocial tendencies.” These aspects of helpful or prosocial behavior begin in childhood and are somewhat stable into early adulthood. Planned forms of helping (volunteering or watching a neighbor’s house or planning in advance to give a friend a ride) showed modest and mixed correlations with personality traits like social responsibility and mastery, though not for all samples (Amato 1990). In the context of bullying, personality measures like empathy, extroversion and openness were related to different types of bystander action (Freis and Gurung 2013). Working for social justice such as challenging racism (a form of looking out for others) has been linked to the openness dimension of the Big Five (Osswald et al. 2010). Studies have supported the idea that increased empathy is related to greater helping and tha at bystander’s physical strength is also a factor (Coke et al. 1978; Fischer et al. 2011).

Demographic characteristics are also part of this intra-individual layer of influences. Other reviews highlight personal demographics like gender (Banyard 2011; Eagley and Crowley 1986; George et al. 1998). For example, gender scripts influenced the types of actions men and women chose with men more likely to help in heroic, assertive and public ways and women more likely to help through nurturing and caring in social networks. The result of this study was that while overall, men and women did not differ in the amount of helping they did, they were different in when they helped, with each gender more likely to help in circumstances congruent with gender role beliefs (Carlo and Randall 2002; Dovidio et al. 2006; Eagly 2009). What is more, research also suggests that correlates of helping behavior vary by gender and that different aspects of masculinity affect confidence about intervening and concerns about negative outcomes of bystander action (Carlson 2008; George et al. 1998; Leone et al. 2015). The gender of an individual bystander also intersects with the gender of other present bystanders (Levine and Crowther 2008) such that men are more likely to intervene when other bystanders are women and women are less likely to help in the presence of men, likely due to activation of internal gender scripts about women needing help and men seeing their role as leaders even in helping situations.

Personal experiences with an issue also impact action. More highly educated community members and those in roles where they might come in contact with child maltreatment were more likely to take action to address suspected child neglect (Fledderjohann and Johnson 2012). Individuals who have experienced trauma or victimization themselves are also more likely to help others (Christy and Voigt 1994; Frazier et al. 2013). On the other hand, a study of IPV specifically found those with personal experience were less likely to take bystander action by reporting to police (Gracia and Herrero 2006).

More consistent results have been found for moral development and coping as correlates of defending behavior for bullying. Bystanders who stepped in to help peers who were being bullied rather than passively standing by scored higher on an assessment of moral responsibility (this was particularly true for adolescent samples) and used more problem-focused coping (rather than distancing or internalizing coping strategies) (Caravita et al. 2012; Pozzoli and Gini 2010). Studies on moral courage (instances when individuals step in to address human rights and other social justice issues rather than just more low-cost helping instances) showed participants who had anger at injustice and strong ethical standards and sense of moral justice (Osswald et al. 2010). These studies have in common a focus on internal and rather stable qualities of an individual that may impact their likelihood of helping others. They are distal variables that shape the lens through which key intentions, attitudes, cognitions, and emotions more proximal to helping are filtered (Zimbardo and Leippe 1991). They may best be part of the focus of prevention and youth development work early on.

There are also a constellation of attitudes that research shows are linked to bystander actions. These include self-efficacy and sense of responsibility. Latané and Darley’s (1970) classic research on the bystander effect was grounded in the notion that individuals in larger groups experience “diffusion of responsibility” that works against taking action. Bystanders in large groups felt others would step in instead. Furthermore, research consistently shows that people who feel more confident in their ability to help are more likely to do so.

3.1.1.2 Lessons from Research on SV and IPV

Similar patterns of factors at the individual level have also been found for bystanders to SV and IPV including gender, personal experiences and attitudes. For example, self-efficacy is linked to greater bystander action related to SV and IPV (Banyard 2011; Lazarus & Signal, 2013). As with helping more generally, some studies find gender differences in how men and women indicate they will take action in situations of SV and IPV (Chabot et al. 2009; Nicksa 2010, 2014). More nuanced measures of constructs like masculinity show that men who believe in gender norms about it being important for men to be strong and tough had greater concerns about negative consequences for stepping in as a bystander to SV. Men who believed that it is important for men to be respected by others reported greater bystander confidence (Leone et al. 2015). Indeed, aspects of gender role stress as assessed in this study were related to different relationships between traditionally studied bystander variables like confidence and perceived pros and cons for intervention. Women were more confident about the helpfulness of the support they provided to a friend who told them about an unwanted sexual experience, though they also reported feeling more emotional distress than men, and were more likely to endorse helpful responses to IPV survivors (Ahrens and Campbell 2000; Banyard et al. 2011; Beeble et al. 2008; West and Wandrei 2002). Studies show that personal experiences with child maltreatment of IPV were related to greater bystander action or intent in some studies (Chabot et al. 2009; Frye 2007) but not others (Gracia and Herrero 2006) with differences likely due to the type of bystander action being assessed. This is relevant to understanding gender since men and women have different risk of exposure to various types of interpersonal violence across the lifespan.

Attitudes specific to SV and IPV are also important for bystander action in these contexts. College students with greater sense of confidence or efficacy in themselves as helpful bystanders had greater intention to intervene and reported greater levels of bystander action (Banyard and Moynihan 2011), though gender also intersected with attitudes as woman displayed greater knowledge about sexual assault and lower acceptance of myths about rape (e.g., Suarez and Gadalla 2010; West and Wanderei 2002). In a community sample from an international study, willingness to report IPV was associated with lower tolerance of IPV (Gracia and Herrero 2006). Indicators of different stages of readiness to change, especially awareness and sense of responsibility related to sexual and relationship violence were related to bystander intentions and behaviors (Banyard et al. 2010; Banyard and Moynihan 2011; Gracia et al. 2009). Importantly, bystander prevention with young adults seems to increase efficacy for addressing sexual and relationship violence (Banyard et al. 2007; Cares et al. 2015).

3.1.2 Who Is Helped?

The “who” of helping and bystander action also requires us to ask who is helped? Interpersonal violence presents complicated situations where both victims and perpetrators can be the receivers of bystander action. Further, the person at whom the bystander action is directed could be a friend or stranger. We know that more general pro-social helping behavior is more likely to be provided to friends and family than strangers (Amato 1990; Penner et al. 2005), yet laboratory studies of bystander intervention usually use strangers as research confederates who help to stage the helping dilemma. To what extent does this make a difference in the barriers and facilitators of action? For example, studies of hypothetical crimes found bystanders less likely to report crimes when they knew the perpetrator (Nicksa 2014). Another vignette study found female college students had more intention to help children while men were more likely to help women (Laner et al. 2001).

3.1.3 Perpetrators Versus Victims

Nearly all of the psychological literature on bystanders has been about understanding help provided to those who need it. The outcome variable in research studies has been whether help was provided to someone having a medical emergency, stepping in to an argument, or offering instrumental help. For example, the arousal cost reward model of helping explains how the emotional arousal created when someone needs help compels us to take action on their behalf (Dovidio et al. 1991, 2006). This research is most relevant to helping victims. Thus, most of the research summarized in this chapter focuses on assisting potential victims. For example, rape myth acceptance, one indicator of victim blaming attitudes, was associated with lesser intent to help as a bystander (McMahon 2010). Perceptions of greater danger to a person in need also facilitated helping across studies (Fischer et al. 2006, 2011). Other researchers have described instances of moral courage, situations that differ from ordinary helping because there is a higher than normal chance that the bystander will experience negative consequences from their actions (a term that me be appropriate for understanding situations where there is risk for SV or IPV), anger seems to be a particularly activating emotion.

On the other hand, social control research in sociology is concerned with how communities as a whole respond to criminal behavior or deviance and express disapproval (Charuand and Brauer 2008). This theory seems most germane to interventions to address the perpetrator. Some similar and some different variables have been researched in this context. Social control is related to how a bystander thinks and feels about what they observe: how much do they see the behavior of the other person as deviant? How much do they see themselves as responsible for doing something about it? How legitimate do they think it is to exercise informal social control? How much does the behavior affect their own self interest—to what extent are they personally affected or harmed? How important is the norm that is being violated (some people may get more upset about environmental issues like littering or not picking up dog waste while others are more angered by hygiene issues like spitting)? How much does the situation bring up feelings of anger or disgust/distain? Higher levels of these thoughts and emotions are related to greater indications of willingness to exert social control against the person violating the social norm via deviant behavior.

3.1.4 Friends Versus Strangers/Ingroup Versus Outgroup

3.1.4.1 General Helping Situations

In the general bystander literature and work specific to sexual violence, people are more likely to help friends than strangers (Amato 1990; Bennett et al. 2014; Katz et al. 2014). Levine and colleagues have looked beyond the distinction between friends and strangers to what happens with bystander action if victims are in-group or out-group members. Even unknown victims who are perceived as in group members (e.g. fans of the same sports team or sport, part of the same campus) are more likely to be helped (Levine et al. 2005; Levine and Crowther 2008). This suggests that how the status of the person in need of help is perceived by the bystander can be influenced by information or framing from the social context. Levine and colleagues have discussed this in terms of social identity theory—that how we see our membership in groups and the connection of victims and perpetrators to those groups influences actions in both general helping and in situations where fighting occurs (Levine et al. 2012).

Interestingly, there may also be different barriers to helping friends and strangers. One study of helping in a sexual assault situation showed that sense of responsibility increased action to help strangers but was unrelated to helping friends. Further, feeling uncertain about the helping skills you have was a barrier to helping strangers but not a barrier to bystander intervention with friends (Bennett et al. 2014). Another study found greater sense of responsibility and empathy to help friends in a SV party vignette and these variables explained participants’ greater likelihood of helping friends compared to strangers. Perceived barriers such as victim blaming attitudes and concerns about what others might think about a bystander stepping in did not seem to differ by whether the person needing help was a friend or stranger Katz et al. (2014). As noted above, intent to help varies by who the bystander knows (Bennett and Banyard 2014; Nicksa 2014). Correlates of each may be different and empirical models to date seem to do better at explaining factors related to helping strangers.  

3.1.4.2 The Case of Violence

A complication in the case of SV and IPV is that frequently bystanders know both victims and perpetrators, and victims and perpetrators may also know each other. For example, individuals who overheard a potential relationship abuse situation were less likely to offer help if they believed the man and woman knew each other (Shotland and Straw 1976). In one study of two vignettes of a hypothetical sexual assault, college students were more likely to identify the situation as a problem if they knew the victim and less likely to label the situation as a problem if they knew the perpetrator. However, participants felt safer taking action if they knew either the victim or perpetrator compared to a situation involving strangers (Bennett and Banyard 2014). Participants had the greatest intent to intervene using tactics that both helped the victim and confronted the perpetrator when they only knew the victim. Participants who knew only the perpetrator in the scenario had lower intent to provide help to the victim, indicated greater intent to confront the perpetrator, and lower intent to contact outside resources (Bennett et al. 2015).  

3.2 What or How

3.2.1 Lessons Learned from General Helping and Informal Social Control

A key piece of bystander action is having the skills to help (Burn 2009; Latané and Darley 1970). What then do bystander skills and actions consist of? How is it that bystanders intervene? Research is clear that both intent to take action (an attitude) and the actual action itself (behavior) are important components of bystander intervention. Behaviors have been assessed among bystanders in a number of ways. Social psychology most often uses laboratory studies. A confederate poses as someone in need of help. Bystanders are brought into the situation under the pretense of filling out surveys or some other behavior and an observer notes whether the bystander takes action to help the confederate. Describing the type of action taken is usually not the focus of study or only a small set of helping behaviors are called for by the situation (helping to fix a flat tire, helping to pick up dropped items). Nonetheless, different classifications of helping exist. For example, Amato (1990) distinguished between planned and spontaneous helping and found different correlates of each. Moral courage has been described as its own type of altruism distinct from helping. It involves addressing injustice and assisting people who face discrimination and unfair treatment because of less powerful social status. In these situations taking action may have high costs for the person who steps in and little personal benefit (for example, individuals who helped Jews in Nazi Germany or someone who steps in to defend a gay man who is being physically harassed for his sexual orientation) (Osswald et al. 2010). Practitioners in bystander prevention have created the “3 D’s” to describe general categories of bystander action as direct action, distracting, or delegating (GreenDot, Etc. 2015).

3.2.2 Considering the Specifics of SV and IPV

Bystander intervention in the case of sexual and relationship violence is nearly always spontaneous and less amenable to the lab studies used most frequently in other studies of helping and prosocial behavior more generally (Banyard et al. 2014; McMahon and Banyard 2011). Several methods have been used to try to document bystander action to address SV and IPV. Harari et al. (2001) were able to stage a sexual assault in a parking lot and observe whether bystanders stepped in to help. They operationalized helping minimally—as either approaching the couple, seeking out a policeman nearby or walking away with no intervention. Parrott et al. (2012) have developed an interesting lab model of sexual aggression where individuals were bystanders in the lab and observed decisions about showing a woman in another room sexually explicit media against her wishes. Speaking up against this behavior was measured as the bystander action. This is one of the few ways to date that seeks to assess bystander action for sexual or relationship violence directly.

Most bystander research uses self-report measures that describe a variety of actions across a continuum of situations (Banyard et al. 2014; McMahon and Banyard 2011). These actions can be direct within the situation (such as interrupting someone taking an intoxicated person away from a college party and up to their room or directly talking to a victim and trying to connect them with resources) or more indirect (enlisting friends to take someone home from a bar or encouraging friends of a victim to reach out to offer support) (Berkowitz 2009).

A study of teens found that most who had a friend in a violent relationship offered some sort of help and support, with talking to friends and offering advice or suggestions or encouraging their friend to leave the relationship being most common (Fry et al. 2013). A community sample of neighborhood residents’ bystander actions related to intimate partner violence found several clusters or dimensions of helping including strategies focused on victims, focused on perpetrators, focused on neighborhoods or on formal helping systems (Frye et al. 2012). Community members reported differences in how possible it would be for them to prevent partner violence, reporting it would be easiest for them to provide help to victims or to access formal services (Frye et al. 2012). Preliminary studies found that factors related to different types of bystander action may vary but little research has explored or described these patterns (Banyard and Moynihan 2011). Thus, we do not yet know whether we need to teach different things to promote actions in low risk versus high risk situations or to encourage supportive behaviors toward victims. For example, Slater et al. (2013) found that in-group members were more likely to use more direct and confronting strategies to break up a fight while out-group members relied more heavily on trying to diffuse the situation with comments.

A number of challenges exist when trying to understand types of bystander action for SV and IPV. It is difficult to separate descriptions of the situation (at a party where someone’s personal space is being violated versus hearing catcalls shouted from a passing car to a woman on the street) from types of bystander response as they are linked. Researchers often measure both at the same time, making it difficult to clearly summarize what we know about the “what” of bystander intervention as distinct from the “when” (a topic considered in more detail in the next section). We also know little about what actions are most helpful. This is a key question for prevention education as it would help us focus resources on skill building for the most effective and safe responses to sexual and relationship abuse. Planty (2002) indicated that victims most often reported that bystanders to crime made the situation better but few studies have investigated this question. Hamby et al. (2015) found that the helpfulness of bystanders varied by victimization type. What is more, whether a victim perceived the bystander to be helpful or not was more important in the link between bystanders and better victim outcomes that just whether or not a bystander was present. Bystander safety was also associated with more positive victim outcomes. So it is not just about whether a bystander takes action. It is also about what bystanders do, how what they do is perceived by the victim, and whether bystanders are themselves hurt in the process (Hamby et al. 2015). This is another critical component of the “what” of bystander intervention that has been under researched.

3.3 When

In order to describe when bystanders intervene I consider several topics. The first is describing the types of situations that constitute sexual and relationship abuse. Next we need to consider opportunity, an understudied topic in bystander research—to what extent do individuals find themselves in situations where they have the chance to help? Finally, I review literature about the nature of the situation—its status as a perceived emergency or not, for example. These are key proximal factors for bystander action no matter when an individual encounters it in the lifespan. In this area there is a growing base of research specific to SV and IPV and thus that is the bulk of the literature on which I draw except when noted.

3.3.1 Considering the Type of Situation

A number of factors impact when people step in to help and using the theory of triarchic influence most of these are more proximal, situational perceptions. These include whether the situation involves an emergency, perceptions of danger to those in need of help, as well as the presence and number of bystanders in the situation.

3.3.1.1 Notes from the Study of Prosocial Behavior More Broadly

As noted in earlier sections of this book, research related to bystander action spans several categories of behavior. Osswald et al. (2010) and Greitemeyer et al. (2006) distinguished between instances where helping is needed and instances of moral courage. Helping involves instances where there are likely few negative social consequences while moral courage are situations where action is needed but there is little potential benefit to the bystander and potentially many negative consequences. An accident or medical emergency is an example of helping while stepping in to challenge someone who is racially harassing someone is more about moral courage (Greitemeyer et al. 2006). Participants in several studies were asked to describe either a time when they helped or not or exhibited moral courage or not and then answered questions about correlates of this action or inaction. When looking at correlates of action separately for helping and courage, helping was significantly associated with greater empathy and quickness in perceiving a need. Both helping and courage actions were related to greater awareness of an emergency. Courage but not helping actions were related to greater felt responsibility, felt anger, perception of norms, and feeling one had skills to take action. These correlates reflect more distally developed variables like empathy and those more specific to the immediate situation like emotional reactions and norms.

3.3.1.2 The Complexity of Situations of SV and IPV

Sexual and relationship abuse span a number of different types of situations, and each of these types often unfold over time. Some instances where bystanders help with SV and IPV may look more like what Osswald et al. call helping but other instances are more clearly about moral courage. What is more, bystanders have the chance to take action before, during, or after an assault (McMahon and Banyard 2011). Bystanders might step in when they notice an escalation of risk factors, they might step in when an assault is taking place, they may choose to help after an assault, when a survivor seeks support or assistance or when a perpetrator discusses his actions. Most researched is how friends, family and professional helpers (law enforcement, advocates, medical and mental health professionals) respond when victims come forward to disclose what happened to them (Campbell et al. 1999; Fry et al. 2012; Ullman 2010). It is clear from this work that victims receive an array of both positive/supportive responses but also negative victim blaming comments and that negative responses in particular can increase a victim’s distress after an assault. We know most of this from the perspective of victims who are clear about the importance of being believed, encouraged, and helped to find resources. We know much less about what enables bystanders to provide these responses at each of these more specific time points.

We also need to explore more about how taking action to help may need to differ between SV and IPV. For example, interviews with friends and family members of IPV survivors documented the long process involved with trying to support these individuals through abusive relationships that may go on for years and periods of leaving and reconnecting (Latta and Goodman 2011). Indeed, what survivors find helpful from bystanders may change depending on their own perceptions of the relationship they are in Edwards et al. (2012). The challenges of engagement and disengagement for bystanders may look different in instances of sexual violence that do not occur within the context of a long-term relationship. Support may need to be more short term and immediate but may need to include interfacing with different services systems and resources as a survivor seeks medical attention and makes choices about pressing criminal charges. Bystanders are also allocated different levels of responsibility for intervention related to these problems. In the case of child maltreatment, bystanders are often mandatory reporters who need to advise authorities as part of their actions. This has also become the case for “responsible employees” on college campuses [faculty and staff who are required under new Title IX provisions to alert campus authorities so that an investigation can be pursued (White House Task Force to Protect Students from Sexual Assault, 2014, notalone.gov)]. In Vermont all citizens are required to report instances of physical danger.

3.3.2 Understanding Opportunity to Intervene

Researchers have begun to study opportunity as its own aspect of bystander intervention. To some extent, this presents a methodological puzzle in that measuring how much helping someone does has to be put in the context of how much opportunity they have to take action (McMahon et al., in press). Studies showed that college students, for example, often reported high levels of opportunity to take action against sexual and relationship abuse (McMahon et al. 2015). More specficially, first year students reported having many chances to take action in low risk situations such as when someone was making harassing comments. Once opportunity was accounted for, college students were most likely, however, to actually take action in high risk situations where they worried an assault might be about to occur and least likely to take action proactively when given the chance to learn more about sexual violence and how it can be prevented [McMahon et al. 2015, in press]. Among a community sample of adults over 40% reported observing child neglect during the past year (Fledderjohann and Johnson 2012). Thus, research suggests that opportunities to be an active bystander to violence are plentiful.

3.3.3 Numbers of Other Bystanders

The classic work of Latané and Darley (1970) showed that bystanders were less likely to help when additional bystanders were present, what they termed the “bystander effect.” They described several attitudes that influence this inaction. Diffusion of responsibility refers to a bystander feeling less motivation to step in when many other bystanders are present. Any one individual feels that others could step in which reduces any one person’s motivation to do so. A second process is “evaluation apprehension.” This attitude leaves bystanders concerned about how others’ will see them if they take some sort of public action. Bystanders may be worried about making a mistake or doing something that others will judge negatively. As a result, bystanders may choose to do nothing in the presence of others, a construct that Latané and Nida also referred to as “audience inhibition” (Fischer et al. 2011; Latané and Nida 1981). Finally, there is the concept of “pluralistic ignorance” whereby bystanders who see others doing nothing, infer that those other bystanders do not see the situation as a problem and thus adopt this view themselves, reducing their sense that something needs to be done (Fischer et al. 2011; Latané and Nida 1981). This work reminds us that whether or not bystanders step in has to do with both the bystander and the people around him or her, proximal aspects of the situation.

Other work says its not just about the numbers but about the composition of the crowd. Levine and Crowther (2008) explored how gender roles can promote and hinder pro-social responses to both hypothetical vignettes and actual need for help. Across a series of studies they used both an imagined scenario of a man attacking a women and a staged opportunity for participants to actually provide help to a research confederate who posed as someone upset and needing support. While both men and women were exposed to the same situation where helping was needed, they varied the gender composition of the group of bystanders surrounding the research participant. They found that men increased helping when other bystanders were women while women decreased their helping when other bystanders were men. Levine and Crowther hypothesized that the results were due to the operation of gender norms such that women deferred to men in a group, assigning them more responsibility to help because of stereotypes about male assertiveness or heroism. Men in the presence of female bystanders responded to these same gender role beliefs and assumed leadership for providing help. Women were more likely to help in same-sex groups. In another study using innovative methods, coding CCTV footage of public aggression, researchers were able to document how third parties were able to lessen violence and aggression and how it was collaborative groups of bystanders rather than individuals who were successful at this suggesting the importance of bystanders responding when others could also be enlisted to help (Levine et al. 2011).

3.3.4 Perceptions of Emergency and Danger

Research on bystander intervention more generally finds that individuals are more likely to intervene if they identify the situation as more of an emergency. An important factor is whether the situation is dangerous just to the victim or to victim and others including bystanders. Bystanders are more likely to step in when they also feel at risk. In these circumstances they are also more likely to enlist others to help and more likely to see situation as a problem that needs to be addressed (Fischer et al. 2011). This may be because dangerous emergencies create a great deal of arousal (anxiety, concern, stress) that motivates bystanders to work with others to more effectively help to reduce the danger including potential costs to themselves (Fischer et al. 2006, 2011). More specific to SV and IPV is the variable of severity. Research finds greater intent to intervene in more severe situations (Bennett et al. 2015) though Gracia et al. (2009) did not find an effect for severity. People who think IPV happens often in a community, a potential marker of perceived severity, were more likely to report IPV positively (Gracia and Herrero 2006).

3.3.5 Perceived Barriers to Bystander Action

People engage in a costs/benefits analysis when deciding when to help. Social psychologists have described this in terms of “rational choice,” that we choose the option that has the best outcomes (Paternotte 2011). Thus, bystander research has also catalogued barriers to intervention (Bennett et al. 2014; Burn 2009). These barriers include not being aware of the situation or not labeling it as a problem, and being concerned that the costs of action will be too high (either because of physical safety concerns or because of concerns that others will not support bystander action). It can also be the case that bystanders lack confidence or lack skills to know what to do. In one qualitative study of college students the following quotes were common:

“it was difficult because it wasn’t clear what was happening at first. I’d never been in a situation like that before. I just didn’t know what to do so I was just kind of freaking out.”

“I was shocked and didn’t know what to do and couldn’t believe that something like that would happen to someone I’ve known almost my entire life”

“I was like, this is a big problem, I don’t know what to do and I didn’t want to confront him to his face because he was very drunk and quite large”

One understudied aspect of barriers to intervention is the use of substances including alcohol. While a large volume of research links alcohol and sexual assault and alcohol use is also a risk factor for IPV, studies are relatively absent about how alcohol may make bystander action more or less likely. Some studies from the substance field may be useful here as they look at factors related to taking action or not related to friends who are drinking too much. This research that found that negative social norms (that is, social norms that promoted drinking as a positive behavior) were related to lower intentions to intervene related to alcohol use (Mollen et al. 2013). In the sexual assault field, research is clear that victim blame is higher and perpetrator blame less if the victim has been drinking, and these factors may lessen bystander action (Bieneck and Krahe 2011). Anecdotally, in my own work, participants who are young and on college campuses discussed their concerns about being an active bystander if they were underage and had been drinking, concerned that they would be punished for their own behavior if formal helpers were involved in the situation. This suggests alcohol use by perpetrators may lessen bystander actions. On the other hand, given what we know about the effects of alcohol on cognitions and decision making, it may also be the case that alcohol use may make bystanders themselves more likely to disregard their own safety and step in or perhaps to make different decisions about how to intervene. Though not specific to SV or IPV, a recent study used focus groups with community members who were part of community night life at the bars and pubs in the UK (Levine et al. 2012). These are social contexts in which alcohol is heavily involved. Yet participants described systems of bystander action and informal social control that operated in those contexts. This suggests that alcohol use, a risk factor for aggression that may require bystander action, does not interfere with some systems of informal social control by bystanders (Koelsch et al. 2012; Levine et al. 2012).

3.4 Why

Closely tied to when is why bystanders help. Classic social psychology research on bystanders focused more on what proximal conditions make action more or less likely rather than unpacking the motives of those who take action. There are some theories in the broader helping and altruism literature that speak to this question and researchers on bullying among children have offered hypotheses as well. Using the theory of triarchic influence, these variables are more distal to the bystander action performed—they involve aspects of individual motivation that likely develop early in the lifespan.

In studies of bullying among children, researchers described individual motives for helping in terms of moral responsibility. Girls who defended others against bullies showed higher or more developed ethical reasoning skills (Caravita et al. 2012). Researchers studying prosocial behavior more generally discuss more evolutionary motivations to help promote one’s own and one’s family and ingroups’ survival (Penner et al. 2005 for a review). These are more distal factors, traits that need to be nurtured in early relational environments (Biglan et al. 2012). This work suggests that we need to begin building bystanders early in the lifespan (Carlo et al. 1999, 2003). Given recent research that shows many youth have been exposed to violence prevention messages (Finkelor et al. 2014) it will be interesting to see how children who get early bullying bystander prevention are primed to receive messages later about bystander behavior and sexual and relationship violence. To date we do not have answers to how these more distal experiences impact actions later on.

In between these distal motivations and more proximal variables described below, is why we help because of relationship oriented motivations of reciprocity and commitment (which in part explains why there is greater helping provided to friends) (Zimbardo and Leippe 1991 for a review). We help others so that they will help us in return later. In our quest to be accepted by others, we observe how they act and try behaviors we have seen others do (Fabiano et al. 2003; Stein 2007; Penner et al. 2005 for a review). Thus, part of why we help is because we see others who have stature within our community or sub-community modeling and endorsing helping attitudes and behaviors (Kelly 2004). This has been studied in terms of our perceptions of what we think others are doing (descriptive norms) and what we think others think we all should be doing (injunctive norms). Such norms can have a powerful impact on what we think and do (Fabiano et al. 2003; Paul and Gray 2011) and some social marketing campaigns that aim to model positive social norms have shown success in changing attitudes that support bystander action to address SV and IPV (Potter and Stapleton 2012). Bystanders are motivated to act if they feel others share similar views. Research shows that norms affect attributions of blame to victims of IPV and one study found greater intent to help a victim of IPV after exposure to pro-helping norms, while decreases in intent to help were found after exposure to anti-helping norms (Baldri and Pagliaro 2014). Further, researchers in the communication field highlight how individual differences in the communication strategies people choose when trying to influence others come from variations in judgements about what constitutes effective communication styles.  Differences in past relationships ( a distal variable) and variations in immediate goals related to communication affect these varied perceptions (White & Malkowski, 2014). This work reminds us that a key component of understanding the “why” of bystander responses is to more carefully observe the motives or goals that may drive a bystander’s selection of one communication or action strategy over another. More research on this topic is needed.

Peer norms are a part of this story as well. In the bullying literature, defenders of victims selected friends who were similar to them and that may have helped promote bystander action (Ruggieri et al. 2013) and classrooms where students perceived that teachers condemned bullying and thought it was a problem had lower rates of bullying and greater defending behavior (Hektner and Swenson 2012; Pozzoli et al. 2012; Sapouna et al. 2010; Veenstra et al. 2014). Among college students peer norms in favor of taking action against sexual violence were not related to greater bystander action overall, though it was related to great intent to help. However, there was a link between more positive norms and greater behavior among subgroups of students, particularly among Black men in college (Brown et al., in press). The role of peer norms and relationship variables will be discussed in more detail in Chap. 5 where I consider how to create change among bystanders. Interestingly, in a community sample in Spain, people who reported exposure to greater public discussion of IPV were more positive about reporting it (Gracia and Herrero 2006).

One theory that addresses more proximal variables related to motivation is the Arousal Cost Reward Model of helping (Dovidio et al. 1991, 2006). Emotional arousal is a source of motivation. Seeing someone in distress causes an uncomfortable level of emotional arousal. People are motivated to reduce uncomfortable arousal and will help others as a way of doing this if the costs are not higher than the benefits. There is empirical support for this model, though other researchers found emotions like love are linked to greater helping then distress or feeling solidarity (Lamy et al. 2012). Aspects of the situation can also influence access to different thoughts and some researchers have found playing prosocial video games of listening to prosocial music increases prosocial thoughts which are a mediating link to actually doing more helping (Greitemeyer and Osswald 2007, 2011).

3.5 Where

3.5.1 Cultural and Geographic Variability: The Potential Importance of Ecological Niche

Bystander behavior also occurs within a broader community context, what we might think about as ecological niches. We know that proximal situational factors like the presence of other bystanders impacts when someone will take action but to what extent do more distal setting characteristics have an impact on helping? These may be objective measures of the setting (rural versus urban) or perceptions of community cohesion, belonging, collective efficacy or they may be aspects of the cultural values a community holds. Understanding how helping differs by ecological niche may help us to better adapt and translate prevention tools to motivate bystander action in these different locations. To answer these questions researchers on prosocial helping more generally have compared helping rates across countries and across rural and urban settings. Relatively little of this work has been done related to SV and IPV specifically.

One way that communities differ from each other is in their physical characteristics and social processes. For example, one international study found less general, low-cost helping (when someone dropped a pen, when someone hurt themselves, or when a blind person needed help crossing a street) in wealthier countries (Levine et al. 2001) while another study focused on the crime of pick pocketing found more bystander intervention in a more advantaged community (Zhong 2010). More specific to IPV, community-level poverty was unrelated to bystander action, though at the level of individual income, more well to do individuals were less likely to help. Characteristics of the setting, such as poverty level, may matter mainly because of the perceptions and relationships that are affected within the community. Pinchevsky and Wright (2012) discussed how communities with high levels of economic disadvantage and where people move around a lot created the conditions for low collective efficacy, low social capital, and less communication. People were struggling to survive and had little time or energy for working on common goals with community members. Residential instability makes building relationships harder, though studies have been mixed with regard to whether community social processes like collective efficacy are related to perpetration rates and bystander intervention (Edwards et al. 2014; Frye 2007; Rothman et al. 2011). Recently, across different communities and different types of interpersonal violence, variables like collective efficacy, cohesion and trust in community authorities were related to greater bystander action or willingness to help (Edwards et al. 2014; Fledderjohann and Johnson 2012). Greater community support and collective efficacy were related to victims’ perceiving bystanders as helpful and more safe (Banyard et al., in press).

Rural and urban differences have been found in helping, with greater altruistic behavior in rural communities (Rushton 1978), though others did not see an impact of city size on helping (Levine et al. 2001). Studies of SV and IPV have considered the unique challenges for survivors in rural communities but I could locate no studies that specifically compared urban and rural communities on bystander action to address IPV and SV. This may be an important direction for future research.

Culture, both in terms of race and ethnicity but also in terms of sets of values of different groups of people is yet another way that location or niche may matter for bystander intervention (Ferrans et al. 2012). Again, much of what we know comes from investigations of helping that does not include SV or IPV. House et al. (2013) examined the development of cooperation and helping behavior across a number of countries. Interestingly, they found similarity in helping and cooperation in early childhood across cultures. Differences then began to emerge and grow through middle childhood and into adulthood, suggesting to the authors the importance of cultural socialization in creating differences particularly in helping that carried potential costs. This is consistent with discussions of moral courage, helping that carries potential costs to the bystander, which is described as influenced by social and political contexts that impact the access to power and support a particular bystander may have (Osswald et al. 2010).

Communities in cultures that place more emphasis on the well-being of the group versus the individual are associated with greater helping in some studies (Levine et al. 2001) but such in-group focus also seems to inhibit helping strangers (Knafo 2009). In relation to SV and IPV, one recent study suggests that stronger ethnic identity was related to greater intent to help in SV and IPV situations at least among college students (Lee 2014). Other studies found that race and culture may impact correlates of bystander action as well. In relation to bullying, while the level of bystander action was similar in two different countries, the correlates (who, where, when) were different between Italy and Singapore (Pozzoli et al. 2012) and types of helping were different between Estonian and Russian-Estonian teenagers (Tamm and Tulvost 2015). Among a sample of U.S. college students, Black students engaged in more bystander behaviors to address SV and these behaviors were more influenced by supportive bystander peer norms than White students (Brown et al. 2014). This shows that bystander opportunities and challenges may be framed by culture in many different ways that we do not yet fully understand but that have implications for adapting our prevention strategies so that they are more culturally competent.

3.5.1.1 Online Versus in Person

Increasingly aspects of SV and IPV are extending into the online environment and thus so is potential bystander action (Bastiaensens et al. 2014). Research in this area has focused mainly on victimization and perpetration in online environments, suggesting that this is an area where bystander action could be helpful. Another line of research has focused on using electronic media to promote helping, through online bystander trainings for college students (e.g. Kleinsasser et al. 2015; Salazar et al. 2014) or through research that showed how prosocial video games or music increased helping by getting people thinking about things like helping and empathy and having those processes more front and center in their thinking (Greitemeyer 2011; Greitemeyer and Osswald 2010). However, bystanders are also active online and can choose to take action there. One study documented a variety of strategies students used online to confront bullying and harassment including telling the bully to stop, offering comfort to the victim, and trying to change the topic (Freis and Gurung 2013). Research also shows that some of the similar barriers to action exist online including that large social networks can easily create diffusion of responsibility (Blair et al. 2005; Martin and North 2015). While online resources for bystander education are proliferating, more research is needed about how to help bystanders take action using a range of social media and online environments.

3.5.1.2 Moving from Reactive to Proactive: A Different Setting for Bystander Action

Most of the settings where bystanders act involve an instance of SV or IPV. Yet another “when” of bystander response is proactive intervention when there is no risk at all. This involves getting more information, pursuing education, volunteering to raise awareness about sexual and relationship violence, displaying a logo or slogan that promotes violence prevention messages, starting conversations with friends and family about anti-violence messages, writing a letter to the editor to comment on a media story, encouraging community leaders to talk about SV and IPV, or working to enact new policies or laws that work against sexual violence and relationship abuse. Rogers (2002) diffusion of innovation theory reminds us that “innovators and early adopters”, the first 15 % of a population to adopt new ideas or behaviors, have a powerful influence on the remainder of the community. Though much prevention and intervention work in the SV and IPV field relies on peer educators and community volunteers who help crisis center staff answer hotlines, plan events, and teach prevention messages in schools we know llittle about the effectiveness of such efforts and how to enhance actions that go beyond only reacting to risk in the moment (Anderson and Whiston 2005). Among college students this is when bystanders are least likely to get involved if given the opportunity (McMahon et al. 2015). A literature in social psychology on volunteerism, or the more public, scripted, planned type of helping that happens over time finds individuals with higher social capital, who have social or organizational support for their work and feel satisfied with the roles and work available to them are more likely to sustain this type of action (Amato 1990; Penner et al. 2005). If part of prevention work is getting more community members involved in prevention and intervention efforts, we need more research to better understand what may motivate them to do so. For example, several studies looked at men’s involvement in SV and IPV prevention and what motivated them to become engaged (Barone et al. 2007; Casey and Ohler 2012; DeKeseredy et al. 2000). What is more, such bystanders have the potential to harness the power of social media to mobilize others to take action (Baek 2015 for research on use of social media to influence political mobilization and voting as an example).

3.6 Summary

  • Factors that influence bystander action include aspects of the self that may form early in life and characteristics of the more immediate situation.

  • A variety of factors across the social ecology help explain when individuals will act or not.

  • Common themes across types of helping include: perceived efficacy, sense of responsibility, awareness of a problem, emotional arousal.

  • Examining research specifically on bystanders related to SV and IPV also reveals new factors that are either unique to these problems or have not yet been studied in terms of bystander action more generally such as peer norms, relationship to people involved in the situation, and victim blaming attitudes.

  • Prevention strategies need to teach flexibility to equip bystanders to manage the complex set of variables at play in any one situation.

3.7 Implications for Practice

Research on bystander action to date provides a number of lessons for prevention, In particular, prevention tools should focus on all of the correlates that research commonly shows help increase bystander action. Prevention tools should promote awareness. They need to build knowledge about what sexual and relationship violence are. Underlying this knowledge is also providing information about consent—what it is and how to actively seek and receive it (Borges et al. 2008). People need to feel responsible and a key piece of this is helping people see that the problems of sexual and relationship violence happen where they live. For example, what students said was most memorable about an educational program on one campus were the local community stories and statistics that made the problem relevant to their own particular experiences (Banyard et al. 2005).

Bystanders also need confidence to take action and be surrounded by others who model and support helping. The foundation for the particular actions that a bystander chooses come from developmental moments early in the lifespan—empathy, prosocial personality tendencies, moral development—that need to be the focus of prevention early on (Biglan et al. 2012). Cultivation of these seeds of helping will affect a variety of bystander actions, not just those for SV and IPV. Bystander action is also motivated by aspects of the current situation and broader community contexts that require prevention efforts at the level of the community to modify aspects of these situations. Bystander focused prevention should have at its core, activities that attend to the variables that appear most consistently in the literature including boosting confidence, increasing awareness, and building skills specific to SV and IPV situations. Prevention strategies need to be built on logic models and strategic plans that take into account the variety of factors across the social-ecological model that impact bystander action.