Revisiting the gender-relations debate in the violent murder of Kitty Genovese: Another side of gender-bias favoring women in bystander reactions to emergencies

Highlights

• Gender does not always affect bystanders' reactions to physical assaults.

• Bystanders are more sympathetic of female victims of physical assault than male victims.

• Dangerous emergencies do not always affect diffusion of responsibility as extant research suggests.

Abstract

The murder of Catherine (Kitty) Genovese in New York in 1964 by Winston Moseley has generated numerous academic publications. One of the major focal points in the debates is the role of gender in bystanders' reactions to violent incidents. Some analysts drew on experiments that found that men did not intervene in incidents involving a man as a perpetrator and a woman as a victim to explain the lack of intervention in the incident by the so-called 38 bystanders falsely reported by The New York Times in 1964. This current article analyzed three videos containing four different assaults that occurred on the busy streets of Argentina, the United States and the United Kingdom to assess whether the gender of the perpetrators and victims affected bystanders' reactions or not. In Incident 1 and Incident 2 involving men as perpetrators with female victims, none of the male and female bystanders physically intervened. In Incident 3 involving a man as a perpetrator and a woman as his victim, both male and female bystanders intervened to save the victim. However, in Incident 4 involving a woman as a perpetrator with a male victim, nobody intervened instead, some of the bystanders laughed at the male victim. The article concludes that whilst gender seemed to have determined intervention in Incident 3 (saving a female victim from a violent man) and its lack in Incident 4 (leaving a male victim to save himself from a violent woman), other factors could be responsible for lack of intervention in Incident 1 and Incident 2 and these include the duration of the assault, the level of violence applied by the perpetrator and bystanders' perception of their own safety. The implications of the bystanders' reactions were highlighted.

Introduction

Catherine (popularly known as “Kitty”) Genovese was a 28-year-old woman who was murdered by a 29-year-old serial murderer, Winston Moseley in a knife attack in Kew Gardens, Queens, New York in the early morning of March 13, 1964 (Gallo, 2014). This incident received increased media attention but the caption of those media reports caused a lot of damages and panic. The original reports by The New York Times in 1964 were instrumental in this panic as will be demonstrated below. The decades-old publication that partly caused the damages and panic is now available online as a digitalized version. The first report was published on March 27, 1964, on the front page with the following sensational headline: 37 Who Saw Murder Didn't Call the Police: Apathy at Stabbing of Queens Woman Shocks Inspector (The New York Times, 1964). Another report by The New York Times in 1964 recently cited by The New York Times (2016) also had the following caption: For more than half an hour 38 respectable, law-abiding citizens watched a killer stalk and stab a woman in three separate attacks in Kew Gardens.

The story of Genovese's murder on March 13, 1964, as described by The New York Times (2016) more recently, particularly the claim that 38 people witnessed or observed the incident and ignored a murderous attack shocked “the national conscience” and eventually led to “an avalanche of academic studies, investigations, films, books, even a theatrical production and a musical”. The story “has been told and retold” and has appeared in “news accounts, editorials, best-selling books and graphic novels, popular songs, movies, and theatrical productions, as well as in textbooks and academic articles” since it was originally published in 1964 (Gallo, 2014, p.273). Academics, preachers and commentators sought explanations for the reported lack of intervention in the incident (Darley & Latané, 1968). Manning et al. (2007, p.555) observed that this incident “paved the way for the development of one of the most robust phenomena in social psychology”, which is the “bystander effect” research pioneered by John Darley and Bibb Latané. References to the bystander effect “can be found in nearly every introductory (social) psychology textbook” (Fischer et al., 2011, p.518; bracket in the original). The bystander effect is a belief that the presence of others during an emergency could lead to no one helping out in such emergency as everyone expects that someone else at the scene could help (Darley & Latané, 1968). The research on this as well as the proposed psychological processes will be discussed in the latter part of this article.

The story of 38 witnesses has been described as “a kind of modern parable – the antonym of the parable of the good Samaritan” (Manning et al., 2007, p.555). As will be demonstrated in the later paragraphs, the claims in the original report were flawed. More recently, The New York Times (2016) has detailed many problems with the claims made in the original publications in 1964. It acknowledges that “the portrayal of 38 witnesses as fully aware and unresponsive was erroneous” and that all other claims in the original article about witnesses and their perceptions were “grossly exaggerated” (The New York Times, 2016). It has been acknowledged that only “a handful of people” saw the first attack clearly (The New Yorker, 2014) and that two of them called the police (The New York Times, 2016). A seventy-year-old woman attended to Genovese and “cradled” her in her arms until the arrival of the police; Genovese died as she was being transported to a hospital (Ibid.). The accounts of the actual witnesses to the incident were examined by Manning et al. (2007) who found that none of these few witnesses knew that it was a serious matter and that none was aware that Genovese was actually being stabbed. The witnesses were never silent as claimed in the original reports because when Genovese first screamed, Robert Mozer yelled at Moseley “Leave that girl alone!”, which led to Moseley fleeing the scene before returning later to meet Genovese again in the vestibule (The New Yorker, 2014).

The recent recognition of the errors in the original publication suggests that those who contributed to the numerous debates on this issue were simply reacting to a false claim made by a journalist who was probably desperate to sell the news. There will be more details on this in the main body of the article. Gender was repeatedly emphasized as part of the reasons for non-intervention. Cherry (1995), for example, attributed the acclaimed non-intervention of the bystanders to the gender of the victim as a woman and the perpetrator as a man. She referred to a previous study by Borofsky et al. (1971) to support the idea that men are unlikely to help a woman under an attack by a man. This study will be discussed in depth in the main body of this article. Borofsky et al. (1971), however, did not make any references to the Genovese incident in their work.

Cherry's (1995) attributions can be blamed on the erroneous report by The New York Times because she might not have made this claim if she had known that the claim that 37 or 38 bystanders observed the incident and failed to intervene, was false. The news reports of this incident by The New York Times and the numerous debates in which gender had consistently featured had caused a lot of damages, including fear among girls and women. Such fear can be found in Gallo's (2014) article where she uses herself as an example of one of those young women who had lived in fear as a result of the misrepresentations of the Genovese incident:

As a thirteen-year-old small-town girl with dreams of living a grown-up life in New York City, I was riveted by the story of her death in my local newspaper. She has haunted me for nearly five decades. In 1964 and beyond, Kitty Genovese provided a stark lesson of the cost of female independence, a reminder of the brutal price that could be exacted from women who boldly inhabit public space on their own terms (Gallo, 2014, p.277).

Although the gender of Genovese was claimed to have motivated the acclaimed lack of intervention (e.g., Cherry, 1995), this was also exaggerated as the few bystanders (never 38 bystanders) did not have a clear idea of the situation. Some people “thought they had heard lovers or drunks quarreling” (The New York Times, 2016). A quarrel and a gruesome murder are completely two different events. It cannot be assumed that people should ignore a gruesome murder in the same way as a quarrel. Further, the people who ignored the quarrel could not be claimed to be gender-biased if they could also ignore a quarrel by alcoholics who might not necessarily be a couple or people of opposite genders. These are part of the reasons I believe that the claim that the gender of Genovese as a woman was the reason for non-intervention by bystanders is unsubstantiated, therefore, requires further investigation based on evidence, rather than a fabricated news report.

Nevertheless, studies have investigated the contributions of gender on bystander intervention in different ways but these studies were mostly laboratory experiments (e.g., Austin, 1979; Borofsky et al., 1971), self-report surveys (e.g., Brewster & Tucker, 2016) and meta-analysis (Fischer et al., 2011; Latané & Nida, 1981). Latané and Nida's (1981) review of over 50 studies on this found conflicting evidence on the influences of gender on bystander intervention. Further, the influences of the methods used to study this phenomenon on the outcomes of the research investigations have been questioned. For example, Latané and Nida (1981) acknowledged the questioning of the validity of the so-called inhibition by some scholars who considered the phenomenon to be restricted to laboratory settings and not natural environments. Whilst Latané and Nida termed the belief by these scholars “unfounded” and a source of “unnecessary confusion” (p.315) before claiming that social inhibition had received numerous empirical backing both in laboratory and field settings, more recent reports (e.g., Fischer et al., 2011; Liebst et al., 2018, Liebst et al., 2019) show that the bystander effect does not always occur, particularly in dangerous emergencies.

Different methods of studying the bystander effect have been found to interact with the results of such studies, such as experimental studies yielding stronger results on bystander effect than quasi-experimental studies and laboratory studies yielding stronger results than field studies (Fischer et al., 2011). This could be due to the fact that experimental studies are subject to more stringent control of potential sources of error variance than quasi-experimental studies (Ibid.). Similarly, laboratory studies often have better control of potential cofounders than quasi-experimental studies (Ibid.). These observations suggest unsettling issues with research methods, calling for an expansion of methods. The fact that it has been found that bystander effect is less likely in dangerous or serious emergencies (Fischer et al., 2011; Liebst et al., 2018, Liebst et al., 2019) demands more analysis of serious emergencies, which can be argued to be more likely in real-life incidents that in experiments. Ethical considerations would normally limit the level of dangerousness violence in any studies. Thus, the current article aims to draw on real-life incidents involving dangerous emergencies to examine the reactions of the bystanders based on the genders of the perpetrators and victims.

Nevertheless, studies of real-life events involving dangerous emergencies captured on CCTV cameras involving both males and females as perpetrators, victims and bystanders have also been conducted (e.g., Liebst et al., 2018, Liebst et al., 2019). However, these studies were based on statistical analysis, thus, could not detail how the events started, the settings, the genders of the bystanders and the exact ways in which they reacted during the incidents. These qualitative details are considered important to be able to judge the reactions of bystanders more accurately. The current, article, therefore, employs an exploratory approach in order to show a clearer picture of how gender interacts during such serious emergencies. This will be done through qualitative analysis of real-life violent incidents captured on video cameras in three different countries. Further details will be provided in the “Methods” section of this article.

The current article is considered beneficial to policymakers and practitioners in the field. It will help them to understand how the influences of the gender of the perpetrators of violent incidents in public spaces and that of their victims may motivate intervention by bystanders or its lack. Understanding this is important to finding solutions for such incidents.

As previously noted, the bystander intervention has generated many academic literature since it was pioneered by Darley and Latané (1968). In fact, by the time studies on bystander intervention was reviewed by Latané and Nida (1981), over 50 studies were found to have been conducted on the subject. A more recent review by Fischer et al. (2011) involved 53 articles on the subject on the basis of over 7700 participants based on studies conducted between 1960 and 2010. The space of a decade between Fischer et al.'s (2011) review and the current article also meant that more studies have been conducted on the subject. Although different researchers on bystander intervention have focused on different issues of interest, the general focus of the studies on the subject is summarised by Fischer et al. (2011) below:

Participants work either alone or in the presence of one or more other participant(s) (passive bystanders) on an allegedly important task (e.g., filling out questionnaires, waiting for the experimenter). They suddenly witness a staged emergency (e.g., the experimenter becomes injured, a perpetrator offends someone, a thief steals something). Their responses to these emergencies are recorded, typically in terms of their probability of intervening and the time it takes them to do so. Results in the multiple-bystander condition are then compared with results in the single-bystander condition (Fischer et al., 2011, p.518, brackets in the original).

Darley and Latané (1968) investigated the bystander effect on 13 male and 59 female psychology students at the New York University. They hypothesized that “the more bystanders to an emergency, the less likely, or the more slowly, any one bystander will intervene to provide aid” (p.378). This hypothesis was upheld when Darley and Latané found that individual participants were less likely to help a victim having a fit if they believed that other people were equally present. Eighty-five percent of the participants who considered themselves as the only persons who knew about the victim's situation reported it compared to only 31% who thought there were 4 other bystanders. The findings suggest that the bystander's failure to help could be due to the presence of other observers, rather than indifference to the victim. The study found that the higher the number of bystanders, the less help the victim would receive. Further, lone bystanders were too fast to help (roughly 45 s) whilst multiple bystanders took much longer to intervene.

A few years after Darley and Latané's (1968) study, Latané and Darley (1970) proposed three psychological processes that would determine whether a bystander would help in an emergency or not. These are audience inhibition, social influence and diffusion of responsibility. Audience inhibition is when the bystander faces the risk of embarrassment if the situation ends up not being an emergency. Therefore, the person might choose not to intervene to avoid being judged negatively by others. The more bystanders during an emergency, the higher the chances of inhibition. Social influence occurs when the bystander relies on other bystanders to define the situation assuming the situation is ambiguous. That is, the bystander decides whether to intervene or not depending on how others make sense of the situation – other people's failure to act could be interpreted by the bystander as indicating a lack of emergency and a normal way of responding to the situation. Diffusion of responsibility refers to the psychological cost associated with a bystander's lack of intervention in an emergency. The presence of other bystanders means that the bystander could see him/herself as not the only person responsible for intervention. It is a way of shifting blame to others or consoling oneself for the lack of intervention. Fischer et al. (2011) have faulted these distinguished processes, arguing that they are close events though they are not fully identical as assumed by Latané and Darley (1970). Fischer et al. (2011, p.518) suggest that the discrepancy might be responsible for the “process ambiguity” trailing bystander research. Their meta-analysis shows ambiguity with respect to the above psychological processes.

The findings of Bickman's (1972) study of bystander intervention on 90 female undergraduates support those of Darley and Latané (1968) above in addition to new findings also made by the former. Bickman found that his participants responded faster to a victim (“a confederate who acted as if she might have been hurt”) when they perceived a bystander as unable (the “Not Able” condition) to help than when they believed that the bystander was able (the “Able” condition) to help. It could be that the participants as bystanders relied on others to define the situation (social influence) or simply saw themselves as not responsible for an intervention when others were there too (diffusion of responsibility).

Latané and Nida's (1981) meta-analysis shows that social inhibition was a powerful factor when it comes to a failure or a delay to act when it was expected that a person would normally act in some ways whether the event was as an emergency (e.g., accident, electric shock, etc.) or non-emergency (e.g., checking the door when someone knocked or pressed an intercom, etc.). They found that both males and females were affected by the group size phenomenon (i.e., social inhibition). However, there are counter-examples to the inhibitory bystander effect whereby “the presence of bystanders can facilitate acts of moral courage” (Fischer et al., 2011, p.518).

Many studies support Darley and Latané's (1968) original study on bystander intervention as well as the psychological processes (Latané & Darley, 1970) connected to it as determining how the bystander would act. Indeed, Latané and Nida's (1981) review of bystander intervention research demonstrates the level of empirical supports received by this psychological phenomenon. Their review shows that the propositions in the original studies were consistently supported irrespective of the places and conditions under which they were conducted with a few exceptions. They found that there was more helping alone in 48 of the 56 studies they compared. The review found that 50% of individuals who witnessed emergencies when alone responded compared to only 20% of those who were exposed to it in the presence of others. This is supported by more recent reports (e.g., Brody & Vangelisti, 2015; Fischer et al., 2011; Liebst et al., 2019; Obermaier et al., 2016).

Whilst Fischer et al.'s (2011) meta-analysis lends support to the diffusion of responsibility reported in classic studies, the work also uncovered some trends that are inconsistent with the classic studies. For example, they found less bystander inhibition in dangerous situations than in non-dangerous situations. This supports an empirical study by Fischer et al. (2006), which found that whilst more help was offered in emergencies with low potential danger by lone individuals than those in the midst of other bystanders, both lone individuals and those surrounded by other bystanders intervened in high potentially dangerous emergencies. Obermaier et al. (2016) studied university students on their intervention on cyberbullying and found that the students' willingness to intervene was boosted by “very severe cyberbullying” as they viewed this as an emergency and showed a greater sense of responsibility. Piliavin et al. (1969) could not establish diffusion of responsibility in their study of undergraduates' willingness or unwillingness to help a collapsed victim. Perhaps the more recent explanations about the absence of diffusion of responsibility in dangerous emergencies (Fischer et al., 2006, Fischer et al., 2011) could explain the outcome of this classic study.

The above findings on bystanders' reactions to serious emergencies contradict both the claims of lack of intervention in the Genovese incident and the classic study (i.e., Darley & Latané, 1968) that was informed by this incident. Nevertheless, with the truth of what actually occurred in the Genovese incident (i.e., the non-existence of the so-called 38 witnesses), the references to the incident as evidence of bystander apathy or diffusion of responsibility could no longer be considered valid. This may also explain why the results of the above reports could not support the claim. However, Fischer et al. (2011) have referred to the murder of Dominik Brunner at a German train station by two teenagers in 2009 as several witnesses failed to intervene to stress how such lack of intervention still persists in recent times. Nevertheless, they acknowledge the fact that the existence of the 38 unresponsive bystanders had been questioned but still concluded that “that there is a negative effect of bystanders on the rate of helping interventions, even if there were no 38 witnesses in the Kitty Genovese murder” (Ibid., p.534). Whilst this was the case with their general findings, the opposite was the case with dangerous emergencies in the same study. Fischer et al., 2006, Fischer et al., 2011 explain their findings that bystander intervention is more likely in dangerous situations than in non-dangerous situations because dangerous situations are perceived as desperate situations demanding definite assistance, which also increases arousal in the bystander.

Fischer et al. (2006) draw on the cost-reward model to explain the likelihood of intervention in dangerous emergencies, suggesting that the increased arousal that the bystander would experience in severe emergencies due to the victim's distress, which can be reduced by intervention means that the bystander is more likely to intervene in dangerous emergencies than in non-dangerous ones. Bickman (1972) also found that his participants responded faster when they perceived the situation as an emergency than when they did not. Similarly, adolescent students who intervened in bullying were those who were able to interpret it as an emergency that needed to be addressed (Jenkins & Nickerson, 2017). Girls were more likely than boys to interpret incidents as an emergency (Ibid.). Although some bullying incidents could be dangerous, they are often not so. Thus, considering the finding, it could be that the perception of an incident as an emergency was enough to motivate intervention as opposed to the dangerousness of the incident.

Whilst dangerous emergencies may attract more intervention, it can be argued that the nature of the danger may determine the willingness to intervene or not to intervene. Although it has been found that interveners in dangerous emergencies generally face low risk of victimization (Liebst et al., 2018), some dangerous emergencies are more frightening than others, which means that intervention may be less likely if not unlikely. For example, a certain bystander who is likely to intervene in a dangerous emergency may fail to do so in a dangerous emergency that appears very frightening, such as a case involving a perpetrator holding a gun with his/her finger on a trigger or a perpetrator wielding a large knife with a threatening appearance. Experimental studies would not use dangerous emergencies of this nature for both health/safety and ethical reasons, thus, it is not too surprising that these studies consistently found more intervention in dangerous emergencies than in non-dangerous ones. Nevertheless, real-life incidents captured on CCTV cameras have equally been studied (e.g., Liebst et al., 2018, Liebst et al., 2019) but details of individual incidents could not be sufficiently provided for a more adequate assessment of the level of threat to bystanders' safety posed by the perpetrators.

Additionally, other factors beyond a high level of dangerousness may determine the likelihood or unlikelihood of intervention in such emergencies, such as possessing a brave personality or its lack, access or lack of access to a personal weapon and personal perception of strength. Brewster and Tucker (2016) found that participants with relevant self-defense training reported that they would intervene in the incidents presented to them in a vignette. Students who intervened in the bullying of other students reported greater knowledge of intervention skills (Jenkins & Nickerson, 2017). In addition to self-defense skills, other personal characteristics, such as height and weight were also linked to bystander intervention (Brewster & Tucker, 2016). Although the little David was so brave to face the giant Goliath as reported in the bible, such bravery is not so common in society. Thus, bystanders as rational beings would normally weigh their chances of success before making a decision to intervene or to stay away, particularly in very dangerous emergencies that may be perceived as a “life and death” situation. Even in the absence of a weapon by a perpetrator, the violent appearance of a certain perpetrator may be enough to scare a bystander who may avoid intervention for personal safety. This may be particularly so when the bystander perceives him/herself as possessing less physical strength than the perpetrator. In such situations, bystanders should not be labeled as “apathetic” because it is natural to be scared, especially when potential harm is perceived as brutal. The difference between dangerous and non-dangerous emergencies with respect to intervention or its lack is explained by Fischer et al. (2011) who found that bystanders' assessment of the physical cost of intervention moderate the decision to intervene or not to intervene since this may serve as a window for them to see the level of danger faced by the victim. The process is explained below:

If participants recognize that they might be injured in case of intervention, they realize that they are indeed confronted with a dangerous situation (both for themselves as well as the victim). This is a more direct experience than assessing the potential cost to the victim from a third-person perspective. In other words, experienced arousal seems to have its origin in the assessment of own risks and not in the assessment of risks for a potential victim (Fischer et al., 2011, p.534, bracket in the original).

Fischer et al. (2011) found that additional bystanders in a dangerous emergency also increased the chances of offering help because the bystander offering help could look upon other bystanders as a back-up resource should the occasion demands this (e.g., when a perpetrator attacks the intervener). Further, some kinds of dangerous emergencies are better resolved in pairs, further suggesting that additional bystanders could encourage intervention. Horowitz's (1971) study found a positive association between group size and the likelihood of intervention among members of a service group upholding the norm of helping others. Calling for a backup team is very common among law enforcement officers due to the idea of strength in numbers. Relying on group size is more likely when the potential intervener feels confident that other bystanders are capable hands or more likely to come to his/her aid than when the intervener is unsure or not confident (Fischer et al., 2011). The role of self-defense skills in bystander intervention (e.g., Brewster & Tucker, 2016; Jenkins & Nickerson, 2017) has been previously discussed.

The role of gender in bystander intervention has been repeatedly investigated. Borofsky et al. (1971) investigated interference or its lack in assaults among male and female undergraduate students. This involved a man beating up a fellow man, a woman beating up a fellow woman, a man beating up a woman and a woman beating up a man. They found that male students in their study could not interfere when a man was beating up a woman even though the men interfered when a man was beating up a fellow man, when a woman was beating up a fellow woman and when a woman was beating up a man. Generally, more males interfered in the fights compared to females. However, when a man was beating up a woman, none of the 6 men interfered. Borofsky et al. (1971, p.317) suggested that the men who failed to intervene when a man was injuring a woman as they did when a woman was injuring a fellow woman might have done so “because they were deriving some kind of vicarious sexual and/or hostile gratification from seeing a man injure a woman”. An alternative explanation for this will be offered in the latter paragraphs.

Fischer et al.'s (2011) meta-analysis found that gender moderates intervention behavior (less bystander effect in males than in females). This is supported by a self-report survey on bystander behavior by Brewster and Tucker (2016) whereby more males stated that they would intervene in emergencies. Further, the higher males in a group, the less bystander effect in an emergency (Fischer et al., 2011). However, Benson et al.'s (1976) study did not find any significant differences between males and females in offering and receiving help from each other. This study, however, was not based on an emergency but a willingness to return a lost item. Nevertheless, Austin (1979, p.2119) found an “extremely high frequency of intervention” by both male and female bystanders when the victim faced a greater risk of harm. Female bystanders intervened both when the risk to the victim was low and high whilst their male counterparts mostly intervened when the risk was high. The gender of the perpetrator did not make any difference to the bystanders. More female victims received help. More female bystanders helped female victims than male victims; more male bystanders also helped more female victims than male victims; more female bystanders also helped male victims than male bystanders helped male victims.

Relationship between bystanders and victims has also generated research investigations. Shotland and Straw (1976) staged assaults to investigate people's reactions to those assaults and found that when the attacker and victim were perceived as strangers to each other, people were much more likely to intervene (65%) than when they were perceived as a couple (19%). Another aspect of the bystander-victim relationship was studied by Liebst et al., 2018, Liebst et al., 2019 who analyzed CCTV footage of violent incidents and found that bystanders were likely to intervene when they were related to the victim. The contributions of the bystander's social relations with the victim on the willingness to intervene showed greater magnitude than the number of other bystanders (Liebst et al., 2019). This supports Benson et al.'s (1976) finding on the interaction of race in helping behavior with Whites being more likely to help fellow Whites than Blacks. This study, however, was not based on an emergency but a willingness to return a lost item.

The influences of alcohol on bystander behavior have also been studied (eVan Bommel et al., 2016). The study found that those who consumed alcohol intervened much faster than sober participants in the presence of other bystanders. This finding challenges the classic study that bystanders who eventually intervened in the presence others took too long to respond (Darley & Latané, 1968). eVan Bommel et al. (2016) explain their finding on the basis that the disinhibition effect of alcohol means that the drunk participants were more likely to see the benefits of helping but not the potential risks, such as confusion of responsibility or embarrassment in the presence of others, which are the usual psychological processes that influence helping in the presence of passive bystanders (e.g., Latané & Darley, 1970).

Hortensius and de Gelder (2014) have studied the neural mechanisms of the bystander effect in young adults and found that neural responses at action preparation level were influenced by group size; decreased activity was observed in the left precentral and postcentral gyri and left medial frontal gyrus whereas an increase in activity was observed in areas related to visual perception and attention. More recently, Hortensius and de Gelder (2018, p.253) have suggested that the willingness to intervene or not to intervene in an emergency may also be determined by the bystander's personality – so-called “personality- and situation-dependent processes”. Hortensius et al. (2016) found that action preparation in people disposed to personal distress was reduced by the presence of bystanders during an emergency.

As previously noted, the bystander effect was not the only debate that emerged following the Genovese incident, rather the changing discourses of gender, race, sexuality and pathology in the American society in the mid-1960s (e.g., Gallo, 2014) in addition to alienation, anomie, moral decay, dehumanization resulting from urbanization and existential despair among others (Darley & Latané, 1968). Violence against women and the hesitation to assist a female victim, especially when the woman is viewed as intimate to the man, was stressed (e.g., Cherry, 1995). This concern is understandable considering the increased level of gender-based violence (GBV) most often perpetrated by men against women across histories and cultures.

GBV is defined as “a form of violence targeting a person based on the gender of an individual” (Mittal & Singh, 2020, p.2). Both men and women are affected by GBV and homicide resulting from it in some cases. However, women constitute significantly higher number of victims whilst men make up the higher number of the perpetrators as we shall see below. This may partly explain why GBV is commonly referred to as “violence against women” in many academic and organizational literatures. The World Health Organisation (WHO, 2017) defines violence against women as “any act of gender-based violence that results in, or is likely to result in, physical, sexual, or mental harm or suffering to women, including threats of such acts, coercion or arbitrary deprivation of liberty, whether occurring in public or in private life”.

Globally, it is estimated that 1 in 3 or 35% of women have experienced this nature of violence physically and/or sexually in their lifetime from their intimate or non-intimate partners (WHO, 2017). Whilst GBV can be experienced at any time and season, the rate is generally higher during conflicts and situations of instability (United Nations Human Rights, n.d.). Again, women, including girls are also more vulnerable in such situations where they could be subjected to sexual violence, including sexual violence as a tactic of war (Ibid.). Fifty-nine percent of Burundian female refugees in a Tanzanian camp were raped by fellow (male) refugees in the same camp in 1996 (Ward & March, 2006). Sixty-six percent of 400 internally displaced persons (mostly women) in refugee camps in three states in north-eastern Nigeria (where Boko Haram terrorists are active) were sexually abused by camp officials taking care of them in 2016 and among the perpetrators were government functionaries, soldiers, police officers, vigilante men and camp leaders (Human Rights Watch, 2016). Disasters around the world, including hurricanes, tsunamis and bushfires have all been found to have increased physical and psychological abuses of which women were the predominant victims (see Campbell, 2020 for discussion).

Similar situation also applies to children. For example, homeless girls in different African cities were found to be victims of rape and sexual violence by fellow homeless boys, including other boys and men from outside their street camps (Agazue, 2021; Chimdessa & Cheire, 2018; Mandalazi et al., 2013). Women are predominant victims of sexual violence of the quarantine paradox associated with pandemics across histories (Mittal & Singh, 2020). The present COVID-19 pandemic has also proven to be one of them (e.g., Campbell, 2020; Mittal & Singh, 2020).

Whilst women are predominant victims of violence in most circumstances, intimate partners are most often the perpetrators of this type of violence than strangers (WHO, 2017). Globally, male intimate partners are responsible for about 38% of homicide against women (Ibid.). Men are known to kill women, including children at a greater rate than women kill men in the home (Serran & Firestone, 2004). Brookman's (2005) study of murdered women in England and Wales (1995–2000) found that 57% of them died in the hands of their husbands, boyfriends or lovers, either current or former ones. Less than 8% of the murders were attributed to the women's friends, acquaintances or other members of the family. The yearly rate of spousal homicide in Canada (1979–1998) was 3 husbands and 10 wives per million couples (Serran & Firestone, 2004). The Crime Survey for England and Wales estimated that 7.7% of women (1.3 million) compared to 4.4% of men (716,000) were victims of any type of abuse in the country (Office for National Statistics, ONS, 2017). Domestic violence by intimate partners, however, may go unreported (Huecker & Smock, 2020). Women's experiences are made worse by their greater vulnerability to sexual assault at the rates of 3.2% of women and only 0.7% of men (ONS, 2017).

Violence by intimate partners sometimes escalates to homicide (e.g., Eriksson & Mazerolle, 2013; Serran & Firestone, 2004). Different patterns are notable between men and women as perpetrators and victims. Intimate partner violence (IPV) and intimate partner homicide (IPH) by women, for example, is most often committed in self-defense (Belknap & Melton, 2003; Belknap et al., 2012; Kethineni, 2001; O'Keefe, 1997; Roberts, 1996; Serran & Firestone, 2004). Serran and Firestone's (2004) review suggests that the use of violence by women against their intimate male partners is most often reactive. Homicide resulting from such incident stands as an extension of IPV (Eriksson & Mazerolle, 2013; Serran & Firestone, 2004). This is because whilst IPV frequently occurs across cultures, only a fraction of these incidents result in homicide, suggesting that IPH is not the norm with battered women, rather it is an exception (Eriksson & Mazerolle, 2013).

Belknap et al.'s (2012) analysis of 117 cases of IPH involving heterosexual couples in Denver Metro (United States) (1991–2009) found that the male victims were up to three times more likely to have been previously arrested for domestic violence compared to female victims. A similar trend applies to convictions with male victims being up to 7 times more likely to have a conviction for domestic violence than female victims (Ibid.). Male victims were also more likely than female victims to have had a restraining order arrest before (Ibid.). This suggests that the female perpetrators of the homicide might have been provoked by their male intimate partners who had a history of violent behavior. However, it has been suggested that the processing of these men by the criminal justice authorities could be a source of strain to the men, which might have increased anger and frustration, particularly when access to their female partners was restricted (Eriksson & Mazerolle, 2013).

Nevertheless, studies around the world consistently found evidence of battery by men and self-defense by women who sometimes killed the former even when the former were never processed by criminal justice authorities. Roberts (1996) found that women who committed homicide against their male partners did so following death threats on them or brutal assaults that occurred multiple times. Similar findings were also made by Campbell (1992). O'Keefe (1997) compared battered women convicted of serious assaults or murders of their batterers with battered women convicted of other offenses and found that those who assaulted and/or killed their partners were older, longer in their relationships and with a longer history of violent victimization in the relationships. This group of women also experienced severe battering and sexual assaults more frequently as well as injuries compared to the control group. The battered women were less likely to have a history of violence against their partners, less likely to have a criminal record or served time prior to these crimes and also more likely to have perceived their lives as in danger.

Wilbanks' (1983) study of homicide rates in Dade Country, United States found that the homicides committed by women were mainly on domestic settings. Pratt and Deosaransigh's (1997) findings whilst examining gender differences and the victim-offender relationships in homicides committed in Contra Costa, United States (1982–1993) suggest that females' involvements in homicide cases were restricted mainly to domestic violence whilst the settings for males' homicide were versatile. Kethineni's (2001) study of the causes of homicides committed by female murderers in India found that most of them were carried out as a way of ending ongoing abuse against them. Melton and Belknap's (2003) study of 2670 IPV cases in the United States also found that female perpetrators of violence were more likely to be defending themselves or fighting back their oppressors. Female perpetrators of lethal violence against men are victims of violence from men who resorted to homicide as a way of protecting themselves from death in the hands of their male partners, including their children (Serran & Firestone, 2004).

Lethal violence by women may be connected to their subordination in a patriarchal society (Serran & Firestone, 2004). Female victims of GBV were found to be entrapped in their relationships and believed that leaving the relationships was difficult or impossible (Serran & Firestone, 2004). Some of them had sought outside intervention but failed before committing homicides (Serran & Firestone, 2004). Roberts (1996) found that homicidal battered women in his study had attempted to escape the relationship by several means, such as through drug or alcohol abuse and suicide, which all failed.

Nevertheless, reports on the rates of aggression/violence by each gender are mixed. Harned's (2001) survey of 874 graduate and undergraduate students about their experiences of violence in dating relationships found similar rates of physical violence across genders although the patterns differed. Archer's (2000) report suggests that females are even more physically aggressive than males. His analysis shows that females are more likely than males to resort to physical aggression and also able to do this more frequently. Although Archer (2000) found that females resorted to physical aggression more easily and more frequently, males were found to be more likely than females to inflict an injury during conflict. However, Melton and Belknap (2003) found no significant differences regarding any sorts of injuries sustained by both genders. The major differences they found were more significant scratches caused by females and more significant bruises caused by males. They also found that men were more likely than women to make threats as well as using violent actions, such as grabbing, shoving, dragging and strangling.

Archer (2002) has attributed the conflicting results obtained by researchers on male-female patterns of aggression to the differing methodologies, samples and contexts. Further, gender differences in the reporting of intimate partner abuse may also affect the statistics on the subject (Belknap & Melton, 2003). Whilst many female victims of intimate partner abuse may minimize the level of abuse against them, their male counterparts who are the perpetrators may also underreport the abuse they mete out or simply minimize them (Ibid.). Female victims of intimate partner abuse may underreport abuse against them for several reasons ranging from shame or embarrassment to the fear of retaliation or reprisals, not trusting the interviewer and the idea that domestic abuse is a personal matter (Ibid.).

Although women are more victims of GBV as demonstrated above, in some cases, women are equally capable of resorting to the patterns of violence, including patterns of homicide that are traditional to men in dealing with their intimate partners.

… when we appear on television and radio discussing the “problem” of wife abuse, we frequently receive telephone calls from men who explain that they are victims of “husband abuse” and ask for equal time from researchers (Straus et al., 2006, p.12).

It has been observed that violence against men by their female intimate partners is also a problem, which has not attracted much academic literature. Incident 3 and Incident 4 analyzed in this current article were staged as a social experiment by the DareLondon for the Mankind Initiative with an aim to raise awareness of domestic violence against men. Hence, the video ended with the following text: 40% of domestic violence is suffered by men (Metro News, 2014; capital letters removed by me). Although the incidents were staged, the bystander reactions were not staged. This is to say that the incidents were suitable for analysis as done in this current article.

Reports of females killing their lovers and spouses with several weapons and in various circumstances, for instance, had made regular news headlines in the British press during the past decades (e.g., Birch, 1993). Referring to a report on the homicide data in the United States, Straus et al. (2006) note that many husbands had been slain by their wives and that kitchen was the most lethal room for such husbands (Straus et al., 2006). These were not always committed in self-defense. Whilst self-defense is often presented as a motivating factor for such violence as previously discussed, Belknap and Melton (2003) acknowledge that girls and women also exist who are not simply responding to violence from their male intimate partners in the form of self-defense, rather they initiate aggression without any provocations from their male partners.

Different patterns of violence are employed by males and females in IPV. Females are more likely to throw an object, hit their targets with an object and bite them as well as using a weapon (Melton & Belknap, 2003). Archer's (2002) analysis of 800 research articles on aggression by heterosexual partners found that women were more likely than men to kick, bite, slap, punch, hit with an object or throw an object whilst men were more likely than women to strangle, choke or beat up. Female perpetrators of IPH mostly “do so with weapons that they need to acquire, men do it with their hands or weapons that are immediately available” (Gavin, 2014). Nevertheless, males' actions were more serious and instilled more fear in their victims than the actions of their female counterparts (Melton & Belknap, 2003). It can be argued that this is due to men having stronger muscles that could lead to their actions provoking greater fear in their victims.