Beyond Sex: Likelihood and Predictors of Effective and Ineffective Intervention in Intimate Partner Violence in Bystanders Perceiving an Emergency

Abstract

Within the framework of the bystander model of intervention, we examined specific correlates and the likelihood of effective and ineffective intervention strategies of bystanders to an instance of intimate partner violence (IPV) identified as an emergency. We measured psychological variables associated with general prosocial behavior (including sex, instrumentality, expressiveness, empathy, personal distress, dispositional anger, and perceived barriers) as influential predictors in four IPV intervention behaviors (i.e., calling 911, talking to the victim, talking to the perpetrator, and physically interacting with the perpetrator). One hundred seventeen college community members completed preintervention measures, watched a film clip of IPV which they identified as an emergency, reported their likelihood of becoming involved and utilizing intervention behaviors, and identified perceived barriers to intervention. Participants were more likely to indicate using effective over ineffective intervention tactics. Lower perceived barriers to intervention predicted greater intervention likelihood. Hierarchical regression indicated that men and individuals higher in anger and instrumental traits were more likely to report that they would engage in riskier ineffective forms of intervention. Implications regarding bystander training and associations to intervention in related forms of violence including sexual assault are discussed.

Keywords

bystander intervention intimate partner violence (IPV)emergency detection prevention

Intimate partner violence (IPV) remains a prevalent problem in the United States and throughout the world (World Health Organization [WHO], 2013). Each year, in the United States alone, there are millions of IPV victims (Centers for Disease Control and Prevention [CDC], 2003; Prah, 2006). A recent study revealed that 10% of a representative U.S. sample of randomly dialed participants reported that they have struck their partner (Klevens, Simon, & Chen, 2012). The health ramifications of IPV are so severe that the WHO (2013) declared the prevention of violence as a public health priority and strongly promotes research assessing intervention as a critical factor in the reduction of IPV.

McMahon and Banyard (2012) generated a conceptual framework of proactive and reactive opportunities for intervention in preventing the specific IPV of sexual assault. They identified primary, secondary, and tertiary prevention behaviors. Prevention behaviors that could occur before the assault are considered primary, those that happen during an assault are secondary, and behaviors occurring after an assault are tertiary. Victims are unlikely to report victimization and, when they do, it is usually to untrained, informal sources such as friends and family (Kaukinen, Meyer, & Akers, 2012; Walsh, Banyard, Moynihan, Ward, & Cohn, 2010); thus, tertiary reactive responses (McMahon & Banyard, 2012) are probably not the most effective to reduce IPV. Primary intervention offers some promise of forestalling IPV among educated individuals who can see the connections of subtle precursors to potential IPV. Secondary intervention during an actual assault, especially if it is deemed an emergency, offers the most explicit opportunity for involvement in reducing IPV.

Bystander Approaches to IPV Prevention

McMahon and Banyard’s (2012) framework is based on the recent initiative to prevent IPV by increasing the likelihood of secondary intervention among bystanders (Banyard, 2008; Chabot, Tracy, Manning, & Poisson, 2009; Katz, 2014). Bystanders are individuals who witness an act and decide to help or do nothing (Banyard, 2008; Latane & Darley, 1970; Volokh, 1999). Approximately one third of all instances of IPV take place in the presence of a third party or bystander (Planty, 2002).

Multiple social-psychological models offer explanations for bystander intervention. According to Latane and Darley’s (1970) multistep situational model of bystander intervention, an individual must first notice the event, deem it to be an emergency, take responsibility, devise a plan, and implement his or her plan. Another model that offers a less altruistic explanation of bystander behavior is the Arousal Cost Reward Model (Piliavin, Rodin, & Piliavin, 1969). According to this model, individuals observing an emergency experience emotional arousal and are motivated to intervene to reduce their arousal and assess the associated personal costs and benefits. Both models suggest that perceived barriers or costs may reduce the likelihood of intervention.

While some researchers have focused on characteristics of the IPV situation (Banyard, Eckstein, & Moynihan, 2010; Latane & Darley, 1970), other researchers have examined personal characteristics of bystanders as predictors of bystander intervention (Banyard, 2008; Casey & Ohler, 2012). Many studies to date have been retrospective and based on varying naturally occurring situations. In a study by Banyard (2008), bystander behavior was operationalized by how many behaviors from the Bystander Behavior Scale (Banyard, Moynihan, & Plante, 2007) participants reported engaging in the last 2 months. In a smaller sample, interview-format study, trained male allies were asked to discuss situations within the last 2 years in which they could have spoken up regarding another male’s sexist or problematic behavior (Casey & Ohler, 2012). In these studies, few of the examined correlates predicted bystander behavior, and bystander behavior was limited by actual exposure to the situation in question (e.g., asking an acquaintance who seems upset if he or she needs help).

Intervention Behaviors

Victims report that indirect and direct interventions can both be helpful. Direct interventions are aimed at the victim or perpetrator. These include talking to the victim or the perpetrator and physically attacking the perpetrator. Indirect interventions are not necessarily aimed at, or evident to, the victim or perpetrator. Indirect interventions include third party reporting to the authorities such as calling 911 (Chabot et al., 2009; Volokh, 1999; West & Wandrei, 2002).

In a large-scale study, victims of violent crimes reported that bystanders were helpful, or more likely to help than hurt the situation, but many had no impact (Hart & Miethe, 2008). In these reports, the perceived effectiveness of bystander intervention was subjectively assessed by the victim as helping, hurting, or having no effect but specific types of bystander behavior were not identified. Victims of IPV indicated that reporting to formal helpers and direct interaction with the victim are effective but direct, especially physical, intervention toward the perpetrator can be aversive and ineffective (West & Wandrei, 2002). It is clear that some attempts to help victims may backfire and antagonize the perpetrator, especially if the bystander does or says something to “make the offender angrier” (Hart & Miethe, 2008, p. 638). Such behaviors could result in injury to the bystander as well as the victim (Klein, 2012). Direct interventions that threatened perpetrators without eliminating contact with the victim are associated with increased homicide (Dugan, Nagin, & Rosenfeld, 2003). Thus, it is imperative to identify predictors of effective and ineffective bystander intervention and to reduce the likelihood of aversive, threatening tactics directed at the perpetrator.

Psychological Predictors

Prior research on helping and prosocial behavior indicates that psychological variables may be influential in IPV intervention. Sex differences in IPV bystander behavior similar, and contradictory, to those found in a meta-analysis by Eagly and Crowley (1986) have been noted. Piliavin et al. (1969) found that 90% of first-time helpers in a naturalistic setting, where the victim was a male who collapsed on a subway (i.e., it was a staged accident that did not involve a perpetrator who caused harm to the victim), were male. In this study, bystander intervention was identified only as direct assistance to the victim; cell phones did not exist, so calling 911 was not an option. Banyard (2008) found that female undergraduates were more likely to report they had engaged in a variety of bystander behaviors than males. She explains that this sex difference may be because many of the behaviors were directed at friends and acquaintances and they took place in safe settings. Chabot et al. (2009) found that men and women were equally likely to intervene in hypothetical cases of IPV but men would choose more direct and risky tactics than women. In the current study, we test for sex differences in four types of intervention behaviors, including calling 911 and interacting with both the victim and perpetrator in a controlled emergency scenario with a victim and a perpetrator.

In 1986, Eagly and Crowley suggested that sex differences in helping behavior may be explained by gender roles. Almost 25 years later, Eagly (2009) reiterated the importance of examining the influence of gender identity, especially expressiveness and instrumental traits, and social expectations on prosocial behavior. Tice and Baumeister (1985) found that high instrumental, but not expressiveness, scores were negatively associated with intervening to help a choking victim when they believed that others were present. In violent situations, perceptions of, and adherence to, gender identity are important in men’s decisions to intervene directly with a perpetrator when other witnesses are present (Carlson, 2008; Leone, Parrott, Swartout, & Tharp, 2016). It seems that gender identity may be associated with the bystander effect and ineffective helping behavior but little research has examined how gender identity is related to intervention decisions when the bystander is the sole witness to IPV. Thus, an examination of instrumental and expressive traits along with biological sex in IPV intervention is important (Chabot et al., 2009; Eagly, 2009).

Personality variables associated with emotional arousal have been identified as correlates to prosocial behavior. Empathy is the dispositional tendency to experience emotions of compassion when observing suffering of others (Davis, 1983) and is related to greater prosocial behavior (Eisenberg & Miller, 1987). Recently, Fries and Gurung (2013) found that empathy was associated with greater likelihood of intervening in online bullying. Using film clips as stimuli, Sze, Goodkind, Gyurak, and Levenson (2012) established that greater empathic concern was indicative of greater charitable contributions. Fries and Gurung and Sze et al. also noted that personal distress, the disposition to experience greater discomfort and anxiousness when observing another suffering, was linked with greater prosocial behavior.

Dispositional anger is another possible predictor associated with emotional arousal. Anger has been shown to be coupled with lower prosocial behavior in urban teens (McMahon et al., 2012). It is also correlated with a variety of direct and indirect aggressive acts toward others (Archer & Webb, 2006) but its relevance to IPV intervention tactics has not been examined.

A variety of cognitive and psychological barriers could impede interventions. Perceived barriers, or costs, to intervention (e.g., personal injury or embarrassment) reduce general bystander intervention (Burn, 2009; Latane & Darley, 1970; Piliavin et al., 1969). Burn (2009) found that barriers specific to each of the five steps of Latane and Darley’s situational model of bystander intervention could decrease intervention to prevent sexual assault. Exner (2011) also found that barriers limited undergraduates’ likelihood of acting to reduce sexual assault. Both studies were specific to sexual assault and examined primary intervention behaviors but neither assessed how barriers might be related to specific intervention behaviors. In this study, we generate and assess barriers specific to four secondary IPV intervention behaviors to expand our understanding of the perceived psychological deterrents of action.

Purpose and Hypotheses

In the current study, we examine correlates, or attributes of bystanders, that are associated with secondary intervention behavior decisions (Arluke, 2012; Banyard, 2008; Casey & Ohler, 2012; McMahon & Banyard, 2012) while controlling for Steps 1 (noticing an event) and 2 (interpreting it as an emergency) in Latane and Darley’s (1970) situational model of bystander intervention. The psychological correlates we examined include sex, empathy, personal distress, instrumentality, expressiveness, anger, and perceived barriers to intervention. Our overall purpose was to evaluate the specific self-reported IPV intervention likelihood of secondary reactive prevention among individuals who responded to a standardized IPV emergency. Understanding the use of effective and ineffective behavior tactics may inform and enhance bystander training for helpful intervention (Alfredsson, Ask, & von Borgstede, 2014). The generation of a reliable measure of perceived barriers to IPV intervention may also be useful for further research into secondary intervention research.

We tested the following hypotheses:

Hypothesis 1: Based on victims’ reporting of more helpful than hurtful bystander intervention, we predict that self-reported likelihood of effective intervention tactics will be greater than ineffective behaviors.

Hypothesis 2: Given previous findings that men are more likely to engage in riskier tactics, we expect that men will report greater likelihood of ineffective behaviors directed at the perpetrator and women will indicate greater likelihood of effective involvement.

Hypothesis 3: Existing models of bystander intervention emphasize the role of barriers in decision making, so we believe that greater perceived barriers will be associated with lower intervention likelihood.

Hypothesis 4: Because researchers have identified the aforementioned psychological variables as correlates of prosocial behavior, we expect that they will differentially predict effective and ineffective intervention likelihood. Greater distress, empathy, and expressiveness will be more indicative of effective intervention, and instrumentality and anger will predict greater direct ineffective intervention.

Method

Participants

One hundred thirty college community members of a rural Northeastern town participated in the study. Thirty percent of the participants were 25 years of age or older with a sample mean age of 26.36 years (SD = 11.07). Thirty-seven percent of the participants were male, 63% were female; one individual did not identify sex. The sample was predominantly Caucasian (89%) with 5% African American, 1% Asian/Pacific Islander, 1% American Indian/Alaskan Native, and 4% mixed race; three individuals did not indicate race. Five percent of the sample stated that they were of Hispanic origin. We excluded nine participants because they did not indicate that they had detected an emergency during the film.¹ We omitted four additional participants because of technological issues (i.e., their survey data were not saved, they did not view the film, or they did not receive directions to indicate an emergency during the film). Because we intended to examine personal predictors for individuals who perceived that they had observed an emergency, our final number of participants was reduced to 117.

Measures

Participants completed an online survey using SurveyGold Software. The measures are described below.

Interpersonal Reactivity Index (IRI)

We employed two 7-item subscales of the IRI (Davis, 1980, 1983), including Empathic Concern (EC) and Personal Distress (PD). These two subscales are widely used to measure dispositional affective empathy. EC assesses feelings of sympathy and concern that an individual may feel for another, for example, “I often have tender, concerned feelings for people less fortunate than me.” PD measures an individual’s self-oriented feelings of personal anxiety and unease, for example, “In emergency situations, I feel apprehensive and ill at ease.” The IRI is scored on a 5-point scale with 1 being does not describe me well and 5 being describes me very well. Higher score indicates greater empathy or distress. In our sample, the reliability for the EC subscale was .82 and that for the PD subscale was .77.

Personal Attributes Questionnaire (PAQ)

PAQ (Spence & Helmreich, 1978) is a 24-item self-concept scale. It is used to measure instrumentality and expressiveness. The Instrumental scale has eight traits including independence and self-confidence and is associated with dominance and self-assertion or agency. The Expressive scale contains eight traits such as gentle and helpful and is indicative of nurturance and interpersonal warmth or communion. Participants indicated their agreement to each statement using a 5-point Likert-type scale, where 1 = strongly disagree and 5 = strongly agree. The PAQ offers adequate internal consistency with reliability scores from .78 to .85. In our sample, the reliability of the Expressive scale was .81 and the Instrumental scale was .47.

Buss and Perry Aggression Questionnaire

We utilized the seven-item Anger subscale of the Buss and Perry Aggression Questionnaire (Buss & Perry, 1992) to assess participants’ dispositional anger. This subscale measures affective processes and is correlated with a variety of aggressive acts (Archer & Webb, 2006). Sample items include “I have trouble controlling my temper” and “I flare up quickly, but get over it quickly.” The items use a 5-point scale with 1 being strongly disagree and 5 being strongly agree. Our sample had a good reliability (α = .81).

Perceived Barriers Scale

We created a scale to assess the extent to which participants believed that various reasons would reduce their likelihood of getting involved in situations like the one portrayed in the video, Dangerous Intentions (Uno, 1995). Specific items for the scale were generated by the researchers as possible deterrents associated with Steps 3 (responsibility), 4 (plan generation), and 5 (action) of Latane and Darley’s (1970) situational model of bystander intervention. Steps 1 and 2 (i.e., noticing the event and deeming it an emergency) were already accounted for by the study design. In addition, we modified some items from Banyard (2008) and Burn (2009) to assess self-efficacy associated with Step 4 and negative outcomes associated with Step 5. The scale contains 13 items, as noted in Table 1. These include the items “It is none of my business,” indicating a lack of responsibility to act, “I don’t know what I could do,” as a lack of self-efficacy in generating a plan, and “I am afraid I would get hurt,” as a perceived reason not to act. For each item, participants used a 5-point scale, where 1 = strongly disagree and 5 = strongly agree. An overall sum score was computed by adding all the scores together. The scale had good reliability (α = .81). The strong negative correlation with overall likelihood of intervention (r₁₁₄ = −.50, p < .001) provides some evidence of validity.

Table 1.

Mean Ratings of Individual Items and Composite Score for the Perceived Barriers Scale and Intervention Likelihood Behaviors (N = 116-117).

Scale	M Rating	SD
Barriers (1 = strongly disagree, 5 = strongly agree)
I would need help from others to be effective.^a	2.67	1.17
I am afraid I would get hurt.^a	2.43	1.31
Someone else would be more helpful.	2.20	1.04
I would be embarrassed if I made a mistake.^b	2.16	1.11
I would be too nervous.^a	2.16	1.18
I don’t know what I could do.^a	1.96	1.00
It is none of my business.^b	1.84	0.94
I didn’t really know what was going on in the situation.^b	1.79	0.85
It is their relationship and problem to solve.^b	1.79	0.89
It would not make a difference.^c	1.52	0.82
There is nothing that could be done in this case.	1.44	0.65
I don’t think anyone would get hurt in this situation.	1.34	0.60
It does not seem that serious.	1.30	0.62
Barrier composite (13 items, α = .81)	1.89	0.53
Intervention likelihood (1 = very unlikely, 7 = very likely)
Overall likelihood of involvement	6.33	1.08
Talk to victim	6.22_a	1.10
Call police/911	6.17_a	1.31
Talk to perpetrator	4.86_b	2.06
Physical interaction composite (three items, α = .88)	4.49_b	1.81

Note. Different subscripts indicate that means are significantly different at the p < .001 level.

These items are only correlated with physical interaction likelihood.

These items only correlated with call police/911.

This item was significantly correlated to all four types of interventions.

Types of interventions

Participants indicated their overall likelihood of getting involved if the scene had been a real life situation and answered questions regarding their likelihood of engaging in six specific direct and indirect intervention behaviors. Participants reported their likelihood of intervention from 1 to 7, where 1 = very unlikely and 7 = very likely. This response pattern allowed each participant to report the likelihood for each intervention behavior and offered improvement over the prior response option of identifying only the most likely action of the four (Chabot et al., 2009) allowing for within-subject comparisons. The items included one indirect item (i.e., call 911), two direct verbal items (i.e., talk to victim and talk to abuser), and an enhanced direct physical composite (α = .89) of three behaviors (i.e., try to stop abuser, restrain abuser, and attack abuser). We tested these four intervention types because of their identified distinctiveness, classification as effective or ineffective tactics, and their prior use as intervention methods (Banyard, Moynihan, Cares, & Warner, 2014; Chabot et al., 2009; Hart & Miethe, 2008; West & Wandrei, 2002).

Procedure

Participants were recruited via posters, on and off campus, to participate in a research study and earn a US$10 gift card. Participants reported individually and voluntarily consented to participate in the study. We informed them that they would watch a video containing violence similar to that which might appear on TV or in a movie. In this simulation paradigm, participants watched a 4-min scene of IPV between a husband and wife, with a male perpetrator and a female victim, from the film Dangerous Intentions (Uno, 1995). Prior to watching the video, they completed the self-report measures of personality. Emergency detection was assessed by pressing a button if, and whenever, they detected an emergency situation while watching the video. After watching the video, they indicated the likelihood that they would get involved in the situation and use the six intervention behaviors and then they completed the Perceived Barriers Scale. Participants were debriefed and received a US$10 gift card for their time.

Results

Manipulation Check

We confirmed that the video was realistic and representative of IPV. On a scale of 1 (never) to 5 (always), when asked how often they believed what they were “watching in the film seemed unreal,” participants’ mean score was 1.77 (SD = 0.94) and 80% stated rarely or never. The mean score for “I did not believe that the scene depicted was indicative, or representative, of domestic violence” was 1.37 (SD = 0.89) with 93% of the sample stating rarely or never.

Intervention Likelihood

As predicted, the likelihood of the behavioral interventions varied as depicted in Table 1. The most commonly reported intervention tactic was to talk to the victim (M = 6.22, SD = 1.10) and the least likely was to physically interact with the perpetrator (M = 4.49, SD = 1.81). The effective behaviors of calling 911 and talking with the victim did not differ significantly from each other but they were both significantly more likely (p < .001) than the two ineffective behaviors (i.e., talking to and getting physically involved with the perpetrator). Likewise, the two ineffective behaviors did not statistically differ from each other.

Contrary to expectations, men (M = 6.02, SD = 1.34) were no more likely than women (M = 6.28, SD = 1.30) to call 911, t(113) = −1.00, p = .32, d = 0.19, nor did men (M = 6.09, SD = 1.15) indicate they would talk more with the victim than women (M = 6.29, SD = 1.08), t(113) = −0.93, p = .35, d = 0.18. However, men (M = 5.56, SD = 1.68) indicated a greater propensity than women (M = 4.42, SD = 2.15) to talk to the perpetrator, t(105.04) = 3.16, p = .002, d = 0.61, and men (M = 5.71, SD = 1.08) reported greater likelihood of physical intervention than women (M = 3.72, SD = 1.74), t(112.85) = 7.58, p < .001, d = 1.46.

Perceived Barriers

In general, as seen in Table 1, bystanders disagreed that the 13 statements on the Perceived Barriers Scale would prevent them from intervening in situations like that depicted in the video. Participants rated a “lack of seriousness” (M = 1.30, SD = 0.62) as the least likely reason to prevent intervention and “needing help from others to be effective” (M = 2.67, SD = 1.17) as most likely to affect action. In Table 2, as expected, mean scores on the Perceived Barriers Scale correlated negatively with all four intervention likelihoods. Individuals with greater perceived barriers reported lower likelihood of engaging in any of the intervention behaviors.

Table 2.

Intercorrelations Between and Means and Standard Deviations of Predictors, Barriers, and Intervention Likelihoods (N = 109-117).

Measure	Predictors							M	SD
Measure	1	2	3	4	5	6	7	M	SD
Predictors
1. Sex (Male = 1, Female = 2)	—	−.12	.22*	−.16	.36***	.13	.10	—	—
2. Instrumentality		—	−.03	−.05	−.03	−.34***	−.35***	20.93	3.60
3. Expressiveness			—	−.05	.65***	.28**	−.03	24.19	4.45
4. Anger				—	−.12	.16	.08	2.38	0.71
5. Empathy					—	.29**	−.11	3.98	0.66
6. Personal distress						—	.26**	2.51	0.75
7. Barriers							—	24.59	6.94
Intervention likelihood
Indirect (call 911)	.09	−.12	.09	−.00	.15	.16	−.28**	6.17	1.31
Talk to victim	.09	.03	.15	.01	.12	.14	−.18*	6.22	1.10
Talk to perpetrator	−.27**	.17	−.02	.01	−.01	−.18	−.23*	4.86	2.06
Physically interact	−.54***	.34***	−.12	.20*	−.26**	−.18	−.37***	4.49	1.81

p ≤ .05. **p < .01. ***p < .001.

Given the significance of the overall barrier score to the four intervention behaviors, we conducted post hoc correlational analyses, as noted in Table 1, to ascertain if there were any noteworthy associations between individual items on the Perceived Barriers Scale and specific intervention behaviors. We found that the item “It would not make a difference” was particularly important as it was the only one that correlated to all four actions. Physical intervention and call 911/police were each uniquely correlated to four different items suggesting differential cognitive reasoning associated with intervention decisions. There were no unique associations between barrier items and the behaviors of talking to the victim or the perpetrator.

Relationships Among the Variables and Predictors

Intercorrelations, means, and standard deviations for sex, personality measures, barriers, and the four intervention propensities are presented in Table 2. Women had greater expressiveness and empathy scores. Instrumentality was negatively correlated to personal distress and perceived barriers but was not associated with sex. Expressiveness correlated positively to empathy and personal distress. Barriers were negatively correlated to all intervention likelihoods. Men were more likely to engage in both ineffective behaviors but personality predictor variables were only related to physical intervention likelihood.

Psychological Predictors of Intervention Likelihood

We planned, a priori, to conduct individual hierarchical regressions to determine whether the addition of perceived barriers and the five personality variables improved the prediction of the four intervention behavior propensities beyond that afforded by sex differences. Due to the unexpected lack of relationship between the personality predictors and the three intervention behaviors of calling 911, talking to the victim, and talking to the perpetrator, none of the three models resulted in significant increases (p < .05) attributed to personality variables in R² after Step 2. At best, the model predicting participants’ likelihood of calling 911 only accounted for 17% of the variance in behavior. Thus, we present only the regression analysis predicting physical intervention likelihood.

We conducted a hierarchical regression analysis to determine whether perceived barriers and personality variables contributed to the prediction of respondents’ likelihood of physical intervention above and beyond that accounted for by sex differences. We entered sex as a predictor in Step 1, and perceived barriers and the personality variables of instrumentality, expressiveness, anger, empathy, and personal distress as predictors in Step 2. The overall regression model, including all seven predictors, was significant, R² = .46, F(7, 106) = 12.70, p < .001. As shown in Table 3, sex explained 28% of the variance in physical intervention propensity, F(1, 112) = 49.99, p < .001. Perceived barriers and personality variables accounted for an additional 17% of the variance, F(6, 106) = 5.65, p < .001. Sex (sr² = .14), perceived barriers (sr² = .07), instrumentality (sr² = .03), and anger (sr² = .02) significantly predicted physical intervention propensity; empathy was marginally significant (p = .056, sr² = .02). Results suggest that men, individuals with lower perceived barriers, higher instrumentality, and greater anger are more likely to decide to intervene physically. Physical intervention likelihood could be predicted well from the set of variables with 46% of the variance in intervention propensity accounted for by the regression.

Table 3.

Summary of Hierarchical Regression Analysis for Variables Predicting Direct Physical Intervention Likelihood (N = 113).

Variable	Model 1			Model 2
Variable	B	SE B	β	B	SE B	β
Sex	−1.98	0.30	−.53***	−1.51	0.29	−.40***
Barriers				−0.96	0.27	−.29***
Instrumentality				0.10	0.04	.20*
Expressiveness				0.04	0.04	.10
Anger				0.42	0.19	.16*
Empathy				−0.55	0.28	−.20
Personal distress				0.05	0.20	.02
ΔR²	.28			.17
F for change in R²	49.99***			5.65***

Note. Overall final model: R² = .46, F(7, 106) = 12.70, p < .001.

p ≤ .05. **p < .01. ***p < .001.

Discussion

In general, participants reported that they would get involved. They were more likely to call 911 or talk with the victim than to interact with the perpetrator. This is a positive finding as these forms of bystander involvement are most helpful (West & Wandrei, 2002). However, participants still reported that they were likely to utilize ineffective interventions in perceived instances of IPV. It should be mentioned that the study was designed so that participants viewed themselves as the only bystander and they identified the situation as an emergency. These two conditions may have reduced diffusion of responsibility and increased the likelihood of intervention (Fischer, Greitemeyer, Pollozek, & Frey, 2006; Latane & Darley, 1970). It is also possible that there were demand characteristics associated with completing the personality measures prior to the video and intervention decisions. Future researchers may want to manipulate the number of bystanders to assess bystander apathy in cases of IPV and test for possible order effects associated with the timing of completing the personality measures.

Our findings partially support prior research that the sex of the bystander is an important factor in intervention. Sex differences in bystander intervention emerged for some, not all, forms of intervention. Similar to Eagly and Crowley (1986), Carlson (2008), and Chabot et al. (2009), the men in our study indicated being more likely than women to engage in riskier acts such as getting verbally and physically involved with the attacker. These sex differences remained statistically significant even after controlling for personality variables that might partially account for the differences. Risky involvement may be associated with different perceptions regarding safety. In post hoc analyses, women (M = 2.73, SD = 1.30) reported that they were more afraid than men (M = 1.95, SD = 1.17) that they would get hurt, t(95.75) = −3.28, p = .001, d = 0.63. Women (M = 2.85, SD = 1.22) also indicated that they would need significantly more help from others than men (M = 2.40, SD = 1.03) to be effective, t(100.45) = −2.14, p = .035, d = 0.40. These sex differences in perceived barriers associated with personal safety offer some explanation for why men may be more likely to engage in risky intervention behaviors and to act alone than women. Also, individuals who perceived themselves as being stronger were identified as more likely to intervene in cases of violence (Laner, Benin, & Ventrone, 2001). Educational appeals to encourage men to pay attention to the risks of physical intervention and the benefits of seeking additional help, regardless of their perceived strength, could help to protect them against harm. Because women generally appear to make safer choices, our findings reinforce the importance of using different intervention strategies and training for men and women (Exner, 2011; Katz, 1995) to reduce direct involvement with the perpetrator.

There were no sex differences in decisions to utilize the more effective intervention tactics of calling 911 and talking with the victim. It could be that the perceived risk of involvement for these behaviors is lower, and they are viewed as equally safe and easy forms of intervention for men and women (Burn, 2009; Latane & Darley, 1970; Piliavin et al., 1969).

Our findings expand on the role of gender in intervention decisions. Prior studies examined sex differences but have not identified the extent to which instrumentality or expressiveness may predict propensity to intervene in IPV. We found that expressiveness was associated with being female but instrumentality was not correlated with sex. The lack of association of instrumentality with sex may be explained by cultural shifts in gender values and socialization. As an individualist society, the United States places greater value on instrumentality traits (e.g., independence and competition) than expressive traits, and there is an increase in the endorsement of instrumental traits among women (Twenge, 1997; van den Bos et al., 2010). Instrumentality, but not expressiveness, demonstrated a moderate effect on predicting physical intervention. This distinction may be explained by the action-oriented instrumental nature of instrumentality versus the interpersonal communality of expressiveness (Spence & Helmreich, 1978). Instrumental characteristics include acting quickly, confidently, and independently in pressure situations and would be expected to correlate with intentional behavior in a perceived emergency as we measured in our study. Also researchers examining bystander behavior in instances of sexual assault found perceived norms to be important in predicting intervention behavior (Banyard et al., 2014; Fabiano, Perkins, Berkowitz, Linkenbach, & Stark, 2004: Gidycz, Orchowski, & Berkowitz, 2011). Dodge and his colleagues (Duke University, 2015) found that children who had acquired a hostile attribution bias were more likely to act aggressively toward others. Thus, an increased propensity among individuals high in instrumental traits to be aggressive toward the perpetrator may be attributed to expectations that others would behave in a similar immediate forceful manner. Additional research should examine the influence of instrumentality on IPV intervention because it is associated with dangerous intervention behaviors (Carlson, 2008).

As indicated by large effect sizes (Warner, 2013), perceived barriers are strong indicators of participants’ self-reported likelihood of getting involved in an emergency such as IPV. Similar findings are reported by Arluke (2012) who identified the influential role of perceived barriers in intervention in instances of animal cruelty. In his study, adolescents with higher perceived barriers reported lower intervention behaviors. Thus, barriers predict both intentional and retrospective interventions in cases of observed abuse. It makes sense that educational initiatives should identify perceived barriers and incorporate these perceptions into individualized training programs. In particular, reducing inaccurate perceived barriers (e.g., “It would not make a difference”) could enhance overall intervention in IPV. Focusing on increasing perceived responsibility (e.g., “It is their problem to solve”) and reducing confusion may enhance the likelihood of calling emergency responders. Also increasing perceived barriers to risky physical attacks (e.g., “I am afraid I would get hurt”) and a misperceived sense of invincibility of acting alone could reduce injury and engagement in direct physical intervention.

We created the Perceived Barriers Scale as a tool to identify and examine rationalization for a lack of bystander intervention in IPV that was not just specific to sexual assault. In our age-diverse sample, the measure had good reliability. The scale offers potential usefulness as a measure for assessing cognitive deterrents to intervention that may prevent Steps 3, 4, and 5 in Latane and Darley’s (1970) situation model and influence the cost–reward analysis as described in the Arousal Cost Reward Model (Piliavin et al., 1969). Additional studies with varied IPV situations could offer further validation and refinement of the scale. Likewise, additional perceived barriers associated with Steps 1 and 2 may need to be added to assess all five steps of the bystander intervention model if these steps are not accounted for in future studies. Also, prior studies have found that individuals’ relationships with the victim and perpetrator may be important barriers associated with Step 3. Adding some of these additional situational variables as well as other arousal items beyond our sole measure of “nervousness” may improve the applicability of the Perceived Barriers Scale.

As expected, direct physical intervention was predicted by greater dispositional anger and instrumentality. These findings imply that bystander training tailored to individuals’ anger may be helpful; thinking about others and our reactions to their plight should inform bystanders to intervene appropriately and not just to reduce their own negative emotions (Piliavin et al., 1969). Based on the findings of Dodge and colleagues (Duke University, 2015), socialization of children to be more empathic and less aggressively impulsive may address this point as well.

Unexpectedly, some personality variables were not related to the other intervention decisions. There was a noted trend for lower empathy to predict greater physical involvement likelihood. This may be because individuals with lower empathy experience less compassion for both the victim and perpetrator and may not be deterred by potentially causing physical harm. However, greater empathy and personal distress did not predict any intervention propensities. The discrepancies in our findings and previous studies (Fries & Gurung, 2013; Sze et al., 2012) may be attributed to the difference in the perceived danger of engaging in the behaviors. Individuals’ emotional dispositions of empathy or personal distress may be more influential in nonemergency situations that offer safer opportunities for action. In the bullying study, intervention was assessed by a variety of behaviors generated by the participants (Fries & Gurung, 2013). Therefore, the likelihood of intervention may have been greater in their study because of the larger number of possible intervention behaviors. Additional research with more response options may offer greater understanding of the relation between these personality components and intervention behavior. It is also possible that these measures of affective empathy may not be as influential in predicting helping behavior as previously thought. Marianovic, Struthers, and Greenglass (2012) found that high levels of cognitive, but not affective, empathy were predictive of helping behavior associated with natural disasters. Einolf (2008) found that empathic concern provided little, if any relationship, with 14 different helping behaviors. Given these findings, future researchers examining IPV intervention may find that cognitive measures of empathy better explain intervention decisions.

More research into all prevention types, not just secondary reactive, as examined here, is warranted (McMahon & Banyard, 2012), and the development of specifically tailored bystander training for effective intervention is needed (Behnke, Ames, & Hancock, 2012). Casey and Ohler (2011) interviewed self-identified allies and noted that only 26% of their sample reported intervening all or most of the time in situations they could provide primary or secondary reactive behaviors. Perhaps the best depiction of the challenge of actual intervention is expressed in the words of a participant who disclosed that “intervening is challenging, complex, and can be an ongoing struggle, and source of guilt . . . being a ‘positive’ bystander is a difficult undertaking under the best of circumstances” (Casey & Ohler, 2011, p. 77). It is especially noteworthy that Casey and Ohler’s (2011) participants were trained allies. If trained individuals find it hard to speak up in conversations with friends, then it is even more challenging for those untrained or those not ready for change to intervene effectively. These challenges further highlight the importance of continued research to identify variables associated with positive and lasting change in bystander attitudes and behavior (Banyard et al., 2010; Brown & Messman-Moore, 2010).

Limitations

The sample was recruited from a college community, so our participants may be relatively more educated than the general population. Our findings may also be limited to residents from the northeastern region who are fairly homogeneous in race. Also, the film clip we used, while providing control as an IPV stimulus, was restricted, in that it was of a middle-age, Caucasian, heterosexual couple where the husband verbally attacked the wife. While our participants found the film to be convincing, this may be, in part, because they found it to be similar to situations they could imagine occurring in their own lives (Levine, Cassidy, Brazier, & Reicher, 2002).

Our participants responded to a realistic but hypothetical instance of IPV where there were no actual consequences for their behavior. It is possible that actual perceptual attitudes in the heat of the moment (i.e., individual perceived barriers) would be greater, and responses to a real life, similar situation may not directly mirror those found in the study.

Conclusion

Our study highlights distinctions between effective and ineffective types of interventions in IPV. Men and women are equally likely to indicate using effective intervention behaviors but there are distinctions in predicting ineffective intervention tactics. Men are more likely to engage in ineffective interventions, and individuals with greater perceived barriers are less likely to engage in interventions. Also, risky physical intervention is predicted by instrumentality and anger. Our findings suggest that educators planning bystander training programs should attend to bystander sex, barriers, and personality as they prepare their informational tools. In particular, anger management could be promoted in those identified as higher in dispositional anger, and discussions of accurate norms regarding instrumentality could reduce dangerous interventions. More personally tailored programs may require some additional costs, but the increased quality of intervention outcomes would be beneficial. In addition, teaching bystanders to be safe and helpful has both short- and long-term positive physical and mental health outcomes for victims (Hamby, Weber, Grych, & Banyard, 2016). Given the growing body of empirical evidence linking exposure to IPV trauma with long-term negative health consequences, effective secondary reactive bystander intervention offers promise in reducing IPV and public health care costs (McMahon & Banyard, 2012; WHO, 2013).

Footnotes

Authors’ Note

The content is solely the responsibility of the authors and does not necessarily represent the official views of New Hampshire IDeA Network of Biological Research Excellence (NH-INBRE).

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by a grant from the New Hampshire IDeA Network of Biological Research Excellence (NH-INBRE) National Institutes of Health (NIH) Grant 1P20RR030360-01 from the INBRE Program of the National Center for Research Resources.

Notes

Author Biographies

Heather Frasier Chabot, PhD, is a professor at New England College in Henniker, New Hampshire. She completed her graduate work in social psychology at the University of New Hampshire in 1999.

Melissa L. Gray, BA, is a graduate of the psychology program at New England College.

Tariro B. Makande, BA, is a graduate of the psychology program at New England College.

Robert L. Hoyt, BA, is a graduate of the psychology program at New England College.

References

Alfredsson

Ask

von Borgstede

(2014). Motivational and cognitive predictors of the propensity to intervene against intimate partner violence. Journal of Interpersonal Violence, 29, 1877-1893. doi:10.1177/0886260513511696

Archer

Webb

I. A.

(2006). The relation between scores on the Buss-Perry Aggression Questionnaire and aggressive acts, impulsiveness, competitiveness, dominance, sexual jealousy. Aggressive Behaviors, 3, 464-473. doi:10.1002/ab.20146

Arluke

(2012). Interpersonal barriers to stopping animal abuse: Exploring the role of adolescent friendship norms and breeches. Journal of Interpersonal Violence, 27, 2939-2958. doi:10.1177/0886260512441073

Banyard

V. L.

(2008). Measurement and correlates of pro-social bystander behavior: The case of interpersonal violence. Violence and Victims, 23, 85-99. doi:10.1891/0886-6708.23

Banyard

V. L.

Eckstein

R. P.

Moynihan

M. M.

(2010). Sexual violence prevention: The role of stages of change. Journal of Interpersonal Violence, 25, 111-135. doi:10.1177/0886260508329123

Banyard

V. L.

Moynihan

M. M.

Cares

A. C.

Warner

(2014). How do we know if it works? Measuring outcomes in bystander-focused abuse prevention on campuses. Psychology of Violence, 4, 101-115. doi:10.1037/a0033470

Banyard

V. L.

Moynihan

M. M.

Plante

E. G.

(2007). Sexual violence prevention through bystander education: An experimental evaluation. Journal of Community Psychology, 35, 463-481.

Behnke

A. O.

Ames

Hancock

T. U.

(2012). What would they do? Latino church leaders and domestic violence. Journal of Interpersonal Violence, 27, 1259-1275. doi:10.1177/0886260511425246

Brown

A. L.

Messman-Moore

T. L.

(2010). Personal and perceived peer attitudes supporting sexual aggression as predictors of male college students’ willingness to intervene against sexual aggression. Journal of Interpersonal Violence, 20, 503-517. doi:10.1177/0886260509334400

10.

Burn

S. M.

(2009). A situational model of sexual assault prevention through bystander intervention. Sex Roles, 60, 779-792. doi:10.1007/s11199-008-9581-5

11.

Buss

A. H.

Perry

(1992). The Aggression Questionnaire. Journal of Personality and Social Psychology, 63, 452-459. doi:10.1037/0022-3514.63.3.452

12.

Carlson

(2008). I’d rather go along and be considered a man: Masculinity and bystander intervention. Journal of Men’s Studies, 16, 3-17. doi:10.3149/jms.1601.3

13.

Casey

E. A.

Ohler

(2012). Being a positive bystander: Male antiviolence allies’ experience of “stepping up.” Journal of Interpersonal Violence, 27, 62-83. doi:10.1605/01.301-0017741112.2011

14.

Centers for Disease Control and Prevention. (2003). Costs of intimate partner violence against women in the United States. Atlanta, GA: National Center for Injury Prevention and Control. Retrieved from http://www.cdc.gov/violenceprevention/pub/ipv_cost.html

15.

Chabot

H. F.

Tracy

T. L.

Manning

C. A.

Poisson

C. L.

(2009). Sex, attribution and severity influence intervention decisions of informal helpers in domestic violence. Journal of Interpersonal Violence, 24, 1696-1713. doi:10.1177/0886260509331514

16.

Davis

M. H.

(1980). A multidimensional approach to individual differences in empathy. JSAS Catalog of Selected Documents in Psychology, 10, 85.

17.

Davis

M. H.

(1983). Measuring individual differences in empathy: Evidence for a multi-dimensional approach. Journal of Personality and Social Psychology, 44, 113-126. doi:10.1037/0022-3514.44.1.113

18.

Dugan

Nagin

Rosenfeld

(2003). Exposure reduction or retaliation? The effects of domestic violence resources on intimate partner homicide. Law & Society Review, 37, 169-198. doi:10.1111/1540-5893.3701005

19.

Duke University. (2015, July 14). Kids expecting aggression from others become aggressive themselves: Children taught to be vigilant for hostility from others are prone to aggressive behavior. ScienceDaily. Retrieved from www.sciencedaily.com/releases/2015/07/150714160921.htm

20.

Eagly

A. H.

(2009). The his and hers of prosocial behavior: An examination of the social psychology of gender. American Psychologist, 64, 644-658. doi:10.1037/0003-066X.64.8.644

21.

Eagly

A. H.

Crowley

(1986). Gender and helping behavior: A meta-analytic review of the social psychological literature. Psychological Bulletin, 100, 283-308. doi:10.1037//0033-2909.100.3.283

22.

Einolf

C. J.

(2008). Empathic concern and prosocial behaviors: A test of experimental results using survey data. Social Science Research, 37, 1267-1279. doi:10.1016/j.ssresearch.2007.06.003

23.

Eisenberg

Miller

(1987). The relation of empathy to prosocial and related behaviors. Psychological Bulletin, 101, 91-119. Retrieved from http://search.proquest.com/docview/203482864?accountid=42685

24.

Exner

(2011). Implications for sexual assault prevention: College students as prosocial bystanders. Journal of American College Health, 59, 655-657. doi:10.1080/07448481.2010.515633

25.

Fabiano

Perkins

H. W.

Berkowitz

A. B.

Linkenbach

Stark

(2004). Engaging men as social justice allies in ending violence against women: Evidence for a social norms approach. Journal of American College Health, 52, 105-112. doi:10.1080/07448480309595732

26.

Fischer

Greitemeyer

Pollozek

Frey

(2006). The unresponsive bystander: Are bystanders more responsive in dangerous emergencies? European Journal of Social Psychology, 36, 267-278. doi:10.1002/esjp.297

27.

Fries

S. D.

Gurung

R. A. R.

(2013). A Facebook analysis of helping behavior in online bullying. Psychology of Popular Media Culture, 2, 11-19. doi:10.1037/a0030239.

28.

Gidycz

Orchowski

L. M. A.

Berkowitz

A. D.

(2011). Preventing sexual aggression among college men: An evaluation of a social norms and bystander intervention program. Violence Against Women, 17, 720-742. doi:10.1177/1077801211409727

29.

Hamby

Weber

M. C.

Grych

Banyard

(2016). What difference do bystanders make? The association of bystander involvement with victim outcomes in a community sample. Psychology of Violence, 6, 91-102.

30.

Hart

T. C.

Miethe

T. D.

(2008). Exploring bystander presence and intervention in nonfatal violent victimization: When does helping really help? Violence and Victims, 23, 637-651. doi:10.1891/0886-6708.23.5.637

31.

Katz

(1995). Reconstructing masculinity in the locker room: The Mentors in Violence Prevention Project. Harvard Educational Review, 65, 163-174. Retrieved from http://nec.gmilcs.org/login?url=http://search.ebscohost.com/login.aspx?direct=true&;AuthType=cookie,ip,url,cpid&custid=danforth&db=ehh&AN=9507051918&site=ehost-live&scope=site

32.

Katz

(2014). Mentors in violence prevention: History/overview. Retrieved from www.jacksonkatz.com/aboutmvp.html

33.

Kaukinen

K. E.

Meyer

Akers

(2012). Status compatibility and help-seeking behaviors among female intimate partner violence victims. Journal of Interpersonal Violence, 28, 577-601.

34.

Klein

(2012). Responding to intimate violence against women: The role of informal networks. New York, NY: Cambridge University Press.

35.

Klevens

Simon

T. R.

Chen

(2012). Are the perpetrators of violence one and the same? Exploring the co-occurrence of perpetration of physical aggression in the United States. Journal of Interpersonal Violence, 27, 1987-2002.

36.

Laner

M. R.

Benin

M. H.

Ventrone

N. A.

(2001). Bystander attitudes toward victims of violence: Who’s worth helping? Deviant Behavior, 22, 23-42. doi:10.1080/016396201750065793

37.

Latane

Darley

J. M.

(1970). The unresponsive bystander: Why doesn’t he help? New York, NY: Appleton-Century-Crofts.

38.

Leone

R. M.

Parrott

D. J.

Swartout

K. M.

Tharp

A. T.

(2016). Masculinity and bystander attitudes: Moderating effects of masculine gender role stress. Psychology of Violence, 6, 82-90.

39.

Levine

Cassidy

Brazier

Reicher

(2002). Self-categorization and bystander non-intervention: Two experimental studies. Journal of Applied Social Psychology, 32, 1452-1463. doi:10.1111/j.1559-1816.2002.tb01446.x

40.

Marianovic

Struthers

C. W.

Greenglass

E. R.

(2012). Who helps natural-disaster victims? Assessment of trait and situational predictors. Analyses of Social Issues and Public Policy, 12, 245-267. doi:10.1111/j.1530-2452415.2011.01262.x

41.

McMahon

Banyard

V. L.

(2012). When can I help? A conceptual framework for the prevention of sexual violence through bystander intervention. Trauma, Violence, & Abuse, 13, 3-14. doi:10.1177/1524838011426015

42.

McMahon

S. D.

Todd

N. R.

Martinez

Coker

Sheu

Washburn

Shah

(2012). Aggressive and prosocial behavior: Community violence, cognitive, and behavioral predictors among urban African-American youth. American Journal of Community Psychology, 51, 4067-4421. doi:10.1007/s10464-012-9560-4

43.

Piliavin

I. M.

Rodin

J. A.

Piliavin

(1969). Good Samaritanism: An underground phenomenon? Journal of Personality and Social Psychology, 13, 289-299. doi:10.1037/h0028433

44.

Planty

(2002). Third-party involvement in violent crime, 1993-1999. Washington, DC: U.S. Department of Justice.

45.

Prah

P. M.

(2006). Domestic violence. CQ Researcher, 16, 1-35. Retrieved from http://library.cqpress.com/cqresearcher/cqresrre2006010600

46.

Spence

J. T.

Helmreich

R. L.

(1978). Masculinity and femininity: Their psychological dimensions, correlates, and antecedents. Austin: University of Texas Press.

47.

Sze

J. A.

Goodkind

M. S.

Gyurak

Levenson

R. W.

(2012). Greater emotional empathy and prosocial behavior in late life. Emotion, 12, 1129-1140. doi:10.1037/a0025011

48.

Tice

D. M.

Baumeister

R. F.

(1985). Masculinity inhibits helping in emergencies: Personality does predict the bystander effect. Journal of Personality and Social Psychology, 49, 420-428. doi:10.1037/0022-3514.49.2.420

49.

Twenge

J. M.

(1997). Changes in masculine and feminine traits over time: A meta-analysis. Sex Roles, 36, 305-325. doi:10.1007/BF02766650

50.

Uno

M. T.

(Director). (1995). Dangerous intentions [Film]. Los Angeles, CA: Columbia Broadcasting System.

51.

van den Bos

Brockner

Stein

J. H.

Steiner

D. D.

Van Yoeren

N. W.

Dekker

D. M

. (2010). The psychology of voice and performance capabilities in masculine and feminine cultures and contexts. Journal of Personality and Social Psychology, 99, 638-648. doi:10.1037/a0019310

52.

Volokh

(1999). Duties to rescue and the anticooperative effects of law. Georgetown Law Journal, 88, 105-114. Retrieved from http://www.lexisnexis.com/lnacui2api/api/version1/getDocCui?lni=3Y6G-GXF0-00CV-50PX&;csi=7337&hl=t&hv=t&hnsd=f&hns=t&hgn=t&oc=00240&perma=true

53.

Walsh

W. W.

Banyard

V. L.

Moynihan

M. M.

Ward

Cohn

E. S.

(2010). Disclosure and service use on a college campus after an unwanted sexual experience. Journal of Trauma & Dissociation, 11, 134-151. doi:10.1080/15299730903502912

54.

Warner

R. M.

(2013). Applied statistics: From bivariate through multivariate techniques (2nd ed.). Thousand Oaks, CA: Sage.

55.

West

Wandrei

M. L.

(2002). Intimate partner violence: A model for predicting interventions by informal helpers. Journal of Interpersonal Violence, 17, 972-986. doi:10.11770886260502017009004

56.

World Health Organization. (2013). Violence against women: Intimate partner and sexual violence against women. Retrieved from http://www.who.int/mediacentre/factsheets/fs239/en/index.html