While it is commonly accepted knowledge that violence begets violence, many workshop speakers emphasized that epidemiological research methods can reveal the ways in which violence spreads, both from one act of violence to many and as a spillover from one type of violence to others. Institute of Medicine (IOM) Board on Global Health Director Patrick Kelley noted that in epidemiology, when trying to understand an infectious disease, the methodology begins with a description of the distribution of cases in person, place, and time. Therefore, an epidemiological survey of the contagion of violence should begin with what different types of violence exist, who is infected, and where and when the violence spreads.
Such a methodology is not new to violence research and prevention. Speaker Valerie Maholmes from the Eunice Kennedy Shriver National Institute of Child Health & Human Development (NICHD) pointed out that, in 1993, a National Institutes of Health panel recommended that research funding priorities in the area of violence should place an emphasis on the context in which violence occurs, and, 10 years later, NICHD led an initiative calling for research on the epidemiology of children exposed to violence. The data presented by many of the workshop speakers highlighted the epidemiological approaches that have been applied to research on and interventions to prevent multiple types of violence.
Speaker and planning committee member Gary Slutkin of the University of Illinois at Chicago defined infectious disease transmission as occurring when an individual or population is exposed to the particular disease and has an increased likelihood of developing the disease. An individual who is inflicted with a disease exhibits some form of symptoms, which vary depending on the disease. Dr. Slutkin suggested that a symptom of violence can be physically injuring another person; speaker Madelyn Gould of Columbia University added that self-directed injury also can be a symptom of violence. Many workshop speakers noted that violence can be transmitted through either direct victimization or merely through witnessing violence.
The incubation period from when the exposure occurs and until disease symptoms develop can vary. As Forum co-chair Mark Rosenberg of the Task Force for Global Health stated, “it can be a long time between something first affecting a person and when it shows up, so [for example] within a family, children [who are] exposed at a very young age may have its impact much later.” To highlight the similarity between varying incubation periods of violence and other infectious diseases, Dr. Slutkin made a comparison between a young child’s exposure to tuberculosis and child abuse. In cases of tuberculosis, reactivation of the disease can occur when the child is in his or her teens or twenties, just as someone exposed to child abuse may become a perpetrator of dating violence or intimate partner violence during adolescence or later in life.
Several workshop speakers pointed to research showing that violence manifests and spreads as different syndromes—collective, interpersonal, and self-directed—and transmission can result in an infection of the same type of violence to which an individual was exposed or as a different syndrome.
Transmission Within Types of Violence
Dr. Slutkin cited evidence that exposure to community violence can lead to perpetration of community violence (Kelly, 2010). The 2011 London riots are an example of how community violence can quickly spread. Furthermore, large-scale political violence can spread to additional perpetration of political violence, as were the cases in World War II and the mass killings in Rwanda in the 1990s, which Dr. Slutkin cited as examples.
Like the spread of acts of collective violence, evidence shows that exposure to interpersonal violence leads to additional acts of interpersonal violence. Speaker and planning committee member Charlotte Watts of the London School of Hygiene and Tropical Medicine noted that there is
evidence of the relationship between intimate partner violence and other types of interpersonal violence. She cited the relationship between early exposures to child sexual abuse, violent households, and harsh punishment as a child, and a woman being more vulnerable later in life to experiencing violence (Abramsky et al., 2011). Furthermore, she pointed to evidence that similar early exposure to violence for men is linked to increased likelihood of perpetrating violence. This early exposure to violence can be the child being directly violated or witnessing violence in the home. Additionally, childhood exposure to interpersonal violence in the home can lead to the child’s perpetration of interpersonal violence against peers later in life, through bullying and dating violence (Crooks, 2011).
The contagion of self-directed violence also has been shown to exist; Dr. Gould noted that the evidence base on the impact of media reporting on suicide, suicide clusters, and adolescent exposure to a suicidal peer has shown an increase in cases of suicidal behavior, both directly and indirectly (Gould, 1990).
Transmission Among Types of Violence
Violence spreads not only as one act of violence to many, but as one act of violence to acts of other types of violence. Many speakers cited evidence of this spread between types of violence. Dr. Watts noted that there is evidence that suicidal behavior can manifest from exposure to other forms of violence; women’s experience with intimate partner violence is linked to increased suicidality (Devries et al., 2011). Similarly, exposure to collective violence can lead to increased rates of intimate partner violence and other forms of interpersonal violence. Speaker Eric Dubow of Bowling Green State University presented evidence that links exposure to ethnopolitical violence and multiple forms of interpersonal violence. He cited studies that support ethnopolitical violence as a higher level stressor that increases other forms of violence at other ecological levels, such as violence within the community, within the schools, and within the family (Dubow et al., 2010; Cummings et al., 2010, 2011). In addition to spread through ethnopolitical violence, exposure to community violence also can lead to an increase in family violence (Mullins et al., 2004).
Dr. Slutkin commented that the manifestation of family violence resulting from exposure to ethnopolitical violence is particularly interesting for the disease model because there is no rational explanation why exposure to violence from an enemy would lead to perpetration against family members. He suggested that this type of transmission shows that violence spreads not for logical reasons, but because it is a communicable disease. He compared the manifestation of different syndromes of violence to the
Understanding the relationship between multiple forms of violence is important for detecting risk factors for the manifestation of future transmissions of violence, and the contagion model can be used to illuminate such pathways. Dr. Gould provided an example that highlighted the importance of such research, including a study that examined multiple forms of violence. Such research, rather than that which is singularly focused on one type of violence, can avoid missing unexpected links among the multiple forms of violence. For example, suicide clusters are primarily a male phenomenon; however, one exception has been among African American girls in gang-related situations where they have been coerced into gang membership and sexual behaviors. Their exposure to collective and inter-personal violence has led to an association with a contagion of suicide within the group.
SUSCEPTIBILITY AND CONTRIBUTING FACTORS
Dr. Slutkin noted that, like other infectious diseases, not everyone who is exposed to violence exhibits symptoms and many individuals can act as carriers without serving as a vector. Physical symptoms of violence are inflicted on those individuals who are susceptible to the disease. Many speakers discussed contributing factors that affect an individual’s or a population’s susceptibility to violence. Many of these factors apply to contagion within and across multiple types of violence.
Many workshop speakers commented that the contagion of violence is dependent on norms associated with violence. A disconnection from or erosion of positive social norms makes individuals and communities more susceptible to the contagion of violence. In citing an example of youth violence in New York City, speaker Jeffrey Fagan of Columbia Law School noted the disconnection between the social norms of the police and the youth as a contributing factor in the contagion of violence. He suggested that there is an extraordinary detachment of youth from the social norms that the police are trying to enforce, which creates a cynicism about the legal system. The higher levels of cynicism about the legal system lead to detachment from the moral and social norms of the law and result in higher rates of violence in those areas. Dr. Dubow noted that there is evidence that violence resulting from conflicts with out-groups is also generalized toward in-group members in society, showing a gradual, consistent, and continuous process of erosion of basic social norms regarding violence in society.
Many workshop speakers noted that although deteriorating social norms can increase susceptibility to violence, changing social norms can be a tool for interrupting the contagion. Dr. Gould noted that recent suicide preventive interventions are focusing on changing peer norms in schools. A program for high school students called Sources of Strength is focused on encouraging students to go to a trusted adult if the student is concerned a peer may be at risk for suicide. The program works by changing norms through emphasizing the importance of help-seeking behavior. Dr. Watts cited violence prevention intervention models and evaluations from Brazil and from South Africa that show that active engagement of men and boys to redefine masculinity can reduce the perpetration of intimate partner violence.
Dr. Fagan noted that the contagion of violence is primarily a social network phenomenon, and increased network density increases the risk of violence transmission. He cited the social network density within public housing communities as an example of such a phenomenon. However, the increased risk is not a factor merely of the density, but the transmission of norms and cultural software that is amplified and reinforced through the network structure. Within insular social networks where violence and danger are learned norms, there is little opportunity to introduce a different kind of social norms model that could teach risk regulation behavior and reduce violence transmission. Dr. Fagan showed a map of incidents of gun-related violence in New York City. The mapping demonstrated the formation of co-offending networks that coalesce around individuals who originally had no or minimal connections, but over time became tighter and tighter social networks.
Dr. Gould also presented the evidence on suicidal behavior based on exposure to a suicide within a peer network. She noted that there has not been that much research in this area, but the majority of the 16 studies that have been done have found a significant association between being exposed to a suicidal peer and the subsequent suicide attempt with odds ratio from 2.8-11.0.
The dose-response effect, that is, the role of increased and repeated exposure to violence, was brought up by several speakers. Dr. Dubow commented that the more ethnopolitical violence to which children are exposed, the greater the occurrences of community, school, and family violence, and individual aggressive behavior. Dr. Gould cited that more than 50 studies
on non-fictional stories of suicide reported in the media have consistently shown that there is a dose-response effect; the more coverage and the more dramatic the coverage, the greater the increase in suicide rates. The reverse also has been shown; suicide rates go down following a decrease in the number of media reports on suicide (Motto, 1970; Hagihara et al., 2007). Dr. Watts suggested there is a dose-response relationship in the contagion of intimate partner violence as well; if both the man and the woman come into the relation with histories of violence, the risk of violence occurring increases.
Forum member and planning committee chair Rowell Huesmann of the University of Michigan stated that evidence has clearly shown that media violence promotes the contagion of violence significantly and substantially. Dr. Gould highlighted the role of the media on suicide clusters. She cited that the most consistent finding is related to the dose-response effect, which is that there are significant increases in suicides when the frequency of media reporting on suicides increases. In addition to increases based on the number of reports, there is a greater likelihood of an increase in suicide when the headlines of the stories are dramatic and when the coverage is on the front page. Dr. Gould pointed to evidence that interventions targeting media coverage have been shown to decrease suicide contagion. She cited an example of media guidelines in Vienna focused on suicides on the subway system in which there was a 75 percent decrease after the guideline implementation. Despite the role of media on transmissions of suicide, she cautioned that media reporting on suicide alone does not lead to suicide contagion; the host, audience, and observer’s preexisting susceptibility all play a role as well.
Dr. Gould noted that while there is a body of evidence on the relationship between traditional media reporting and suicide contagion, the effects of the Internet have not been well studied. She suggested that trying to determine the effects of the Internet on suicide contagion is challenging because the speed at which the communication is shared is faster than anything seen before or even envisioned. Dr. Fagan suggested there is a paradox when it comes to the role of the media and community violence. From one perspective, the more time a youth spends on the Internet, the less time he or she is out in the community engaging in violent behavior. However, youth are exposed to violent content through the Internet. That raises many other questions about, for example, what the dose-response curves are and what personal characteristics are mediating factors. Dr. Watts also commented on the paradox of violence and the Internet. As an example, she stated that individuals who have leanings toward pedophilia
may be in scattered physical locations, but the Internet provides an opportunity to link up with like-minded people and to reinforce and condone those behaviors and maybe lead to action. But she also acknowledged that the Internet has provided extensive opportunities in terms of promoting alternatives and providing youth different forms of relationships and ways to have relationships.
Several workshop speakers suggested that age can play a role in the contagion of violence. Dr. Dubow cited that the youngest children within his studies have been the most impressionable in terms of the exposure to violence. Additionally, evidence shows that exposure to ethnopolitical violence adversely affects a child’s emotional security toward his or her community, which in turn leads to more externalizing behaviors such as aggression and attention disorders (Cummings et al., 2010, 2011). Dr. Gould noted that suicide clusters occur primarily among teenagers and young adults in the United States. She commented that one of the hypotheses for the youth factor is that neurocognitive functioning in adolescence is not fully developed. Youth decision making and impulsivity might be one reason why young people may be more susceptible to transmissions through media reporting and other peer and social networks.
Dr. Rosenberg noted that one of the increasing interests in global health and disease prevention is social and economic determinants, possibly even more so than physiological determinants of health. He suggested this is an area that holds great potential for contributing to the contagion of violence model. Dr. Slutkin commented that violence itself is a social and economic determinant of the other health issues and, arguably, could be the dominant social and economic determinant of health outcomes. Dr. Fagan added that the socioeconomic determinants that often are risk factors of violence are also risk factors for other adverse health outcomes.
IMPLICATIONS FOR INTERVENTION
The evidence supporting the contagion of violence within and across types of violence has implications for designing interventions to interrupt the contagion. Many speakers commented that, like other infectious diseases, a reduction in the spread of violence requires interventions that reduce susceptibility and devise new norms. Several speakers also noted that interventions designed to prevent the spread of one type of violence
Interventions can be multidirectional. Dr. Watts cited an example of an intervention in Côte d’Ivoire that was focused on preventing intimate partner violence by working with men to redefine constructs of masculinity. In follow-up surveys, the data collected suggested that some men involved in the intervention program chose not to become involved in current ethnopolitical violence because of the experience they had during the intervention program. However, some multidirectional consequences can be negative. Dr. Fagan told an anecdote of a policing program in New York that involves stopping individuals to search for illegal guns, increasing the number of young women carrying guns because they are not stopped and searched as often as men.
Some speakers suggested that interventions should focus on changing social norms. Dr. Watts suggested that changing social norms around the construct of masculinity has been shown to prevent the contagion of family violence. She also noted that intervention programs often focus on the woman who had been a victim of intimate partner violence, but do not address the children within the household. She suggested that interventions targeted for the entire household are key to interrupting the contagion of violence. Dr. Fagan suggested that retooling the relationship between the police and gun offenders could help interrupt community-level violence. Unregulated punishment can exacerbate susceptibility to violence and increase the network density of people who share police victimization experiences. Dr. Gould commented that, to interrupt suicide contagion, social norms regarding talking openly about suicide risks need to change. There is a myth that because suicide is contagious, you cannot ask about suicide. However, you can assess for suicidal ideation without making a person think that he or she should commit suicide.
Dr. Dubow commented on the importance of interventions focused on protective factors. Most interventions to prevent ethnopolitical violence are trauma-focused. However, the evidence is showing the importance of protective factors and such interventions can be implemented in school and community settings. He also noted the importance of enhancing the protective factor of the family (specifically, the family is protective against exposure to violence on children), by bolstering the family itself. This could be through providing mental health services to families during times of ethnopolitical conflict, or in the case of reintegration post-conflict (such in the case of child soldiers), by providing extra-familial activities such as work that reduce stress on the family structure itself.
Speaker Carl Bell of the Community Mental Health Council commented that one of the challenges with public health interventions to interrupt epidemics is that the epidemics often are cyclical. He gave the example
of a syphilis epidemic in gay men in Chicago: “We put signs on the buses. The epidemic went away. The signs came down. The epidemic came back.” He suggested that three things are needed to stop an epidemic: an evidence base, an implementation system, and political will.
Key Messages Raised by Individual Speakers
• Violence is contagious both within and across types of violence (Dubow, Fagan, Gould, Huesmann, Slutkin, Watts).
• Social norms contribute to the contagion of violence and norms change has the potential to interrupt it (Fagan, Gould, Slutkin, Watts).
• Media can both facilitate and prevent the contagion of violence; however, the role of the Internet in the contagion process is not well understood (Fagan, Gould, Huesmann, Watts).
• Dose-response effect applies across types of violence (Dubow, Gould, Watts).
• Understanding the contagion process can inform the development of violence prevention interventions as well as illuminate potential unintended consequences that affect other types of violence (Bell, Fagan, Gould, Watts).
Abramsky, T., C. Watts, C. Garcia-Moreno, K. Devries, L. Kiss, M. Ellsberg, H. Jansen, and L. Heise 2011. What factors are associated with recent intimate partner violence? Findings from the WHO multi-country study on women’s health and domestic violence. BMC Public Health 11:109.
Crooks, C. V. 2011. The science of interrupting/preventing the cycle of violence. Presented at IOM Workshop on Preventing Violence Against Women and Children, Washington, DC, January 27.
Cummings, E. M., E. Cairns, M. C. Goeke-Morey, C. E. Merrilees, A. C. Schermerhorn, and P. Shirlow. 2010. Political violence and child adjustment in Northern Ireland: Testing pathways in a social-ecological model including single- and two-parent families. Developmental Psychology 46:827-841.
Cummings, E. M., E. Cairns, M. C. Goecke-Morey, C. E. Merrilees, A. C. Schermerhorn, and P. Shirlow. 2011. Longitudinal pathways between political violence and child adjustment: The role of emotional security about the community in Northern Ireland. Journal of Abnormal Child Psychology 39:213-224.
Devries, K., C. Watts, M. Yoshihama, L. Kiss, L. B. Schraiber, N. Deyessa, L. Heise, J. Durand, J. Mbwambo, H. Jansen, Y. Berhane, M. Ellsberg, C. Garcia-Moreno. 2011. Violence against women is strongly associated with suicide attempts: Evidence from the WHO multi-country study on women’s health and domestic violence against women. Social Science Medicine 73(1):79-86.
Dubow, E. F., L. R. Boxer, J. Ginges, S. Gvirsman, L. R. Huesmann, S. Landau, and K. Shikaki. 2010. Exposure to conflict and violence across contexts: Relations to adjustment among Palestinian children. Journal of Clinical Child and Adolescent Psychology 39:103-116.
Gould, M. S. 1990. Suicide clusters and media exposure. In Suicide over the life cycle: Risk factors, assessment, and treatment of suicidal patients, edited by S. Blumenthal and D. Kupfer. Washington, DC: American Psychiatric Association. Pp. 517-532.
Hagihara, A., T. Abe, and K. Tarumi. 2007. Media suicide-reports: Internet use and the occurrence of suicides between 1987 and 2005 in Japan. BMC Public Health 11(7):321.
Kelly, S. 2010. Exposure to gang violence in the community: An integrated review of the literature. Journal of Child and Adolescent Psychiatric Nursing 23:61-73.
Motto, J. A. 1970. Newspaper influence on suicide. Archives of General Psychiatry 23:143-148.
Mullins, C. W., B. A. Jacobs, and R. Wright. 2004. Gender, streetlife and criminal retaliation. Criminology 42:911-940.