by Sheyla A. Delgado, Kwan-Lamar Blount-Hill, Marissa Mandala, and Jeffrey A. Butts
Research & Evaluation Center
May 15, 2015
This report was made possible with funds from the New York City Council and the Robert Wood Johnson Foundation of Princeton, NJ. The authors of this report are grateful to all of the Cure Violence organizations involved in this study for their cooperation and support. In addition, the authors are very grateful to Dr. Richard Curtis for his guidance in the methodological design and conduct of this study, and also to Evan Misshula for his statistical expertise and advice. The study would not have been possible without Kathleen A. Tomberg and Rhoda Ramdeen at the Research & Evaluation Center who provided essential editorial and administrative support. Data-collection for the study was directed by Sheyla Delgado with the essential help and management of Efram T. Thompson, Justice Banks, Laila Alsabahi, Yeireline Rodriguez, Clarence Neil Colon, Baron Plaza, Anjelica Camacho, and Leonardo Dominguez.
Violent crime in New York City declined sharply during the previous two decades, but some neighborhoods remain highly vulnerable to gun violence. In 2011, the City Council’s Task Force to Combat Gun Violence recommended the implementation of a new “Crisis Management System” (CMS) in five New York City neighborhoods. The CMS approach includes strategies from the Chicago-based Cure Violence model along with other social and legal services.
The Research and Evaluation Center at John Jay College began assessing the implementation and effects of these efforts in 2013. One element in the project involves in-person surveys with young men (ages 18-30) in many of the neighborhoods implementing the strategy. The study operates under the brand name, “NYC-Cure.” This report contains survey results from the first four neighborhoods to be involved in the NYC-Cure study.
The survey instrument measures personal attitudes towards violence and experiences with violence, as well as each respondent’s awareness of local violence prevention efforts. Additional surveys are being conducted in these and other neighborhoods around New York City in an effort to detect any changes over a three-year period. The study relies on Respondent-Driven Sampling (RDS) methods to recruit survey respondents.
- According to surveys conducted from March through June of 2014, Cure Violence programs have established a strong presence in New York City neighborhoods. The majority of young males in each neighborhood surveyed for this study recognized the educational materials (e.g. flyers, pamphlets, etc.) used by the organizations to promote their services. When asked if they recognized any staff from the programs (using unlabeled photographs), the majority of the survey respondents recognized at least one staff member.
- Gun violence in these neighborhoods remains a real concern. When respondents were asked about their exposure to guns and gun violence, the majority reported hearing gunfire in their neighborhood at least once in the past 12-months and almost one-quarter (23%) heard gunshots more than 10 times.
- Violent victimizations are common in these neighborhoods. Almost one in five survey respondents reported being stabbed at some time in the past, and almost 40 percent reported that they had been the target of gunfire in the past.
- Contact with law enforcement was also common. Nearly 80 percent of all survey respondents reported that they had been “stopped, questioned, and frisked” in their neighborhoods at least once within the past year.
In 2013, John Jay College’s Research & Evaluation Center, in collaboration with the Robert Wood Johnson Foundation and the City of New York, began an evaluation of the Cure Violence model (formerly known as Chicago CeaseFire), a public health approach to gun violence reduction.
Cure Violence (CV) relies on the efforts of community-based “outreach workers” and “violence interrupters” to identify individuals vulnerable to gun violence, and to use their personal relationships and social supports to dissuade them from engaging in violence (Butts et al. 2015). The New York City Council, and later the Mayor’s Office in New York, provided support for Cure Violence-based programs supplemented with other wrap-around services. Together, these efforts are known as New York’s Crisis Management System (CMS). The wrap-around services of CMS focus on education, employment, mental health, and legal assistance for community residents (Delgado et al. 2013).
Cure Violence is designed to have effects beyond direct program participants. The developers of the model suggest that if these strategies are implemented with high levels of fidelity, they should begin to denormalize violence in the entire community. With this in mind, the John Jay research team set out to measure changes in attitudes towards violence of young adult men living in the targeted neighborhoods, rather than focusing on program participants only.
John Jay researchers conducted nearly 800 surveys with men between the ages of eighteen and thirty residing in the four treatment neighborhoods implementing the City’s violence prevention system. These neighborhoods include Central Harlem, East New York (Brooklyn), South Jamaica (Queens), and the South Bronx.
Initial survey data collected in 2014 serves as a baseline for measuring changes in violence-related factors in the four study neighborhoods. The evaluation study includes plans for two additional data-collection periods. Surveys will be conducted in 2015 and 2016 to determine whether violent norms appear to change over time and whether that change is more pronounced in neighborhoods where CV-based programs operate.
Based on the first round of neighborhood surveys, the study findings suggest that Cure Violence programs are widely recognized in their neighborhoods and residents in those neighborhoods are at relatively high risk for gun violence:
The Crisis Management System
In 2011, the New York City Council created the Task Force to Combat Gun Violence in response to an increasing number of shooting incidents occurring in the city. After reviewing New York Police Department shooting data, the Task Force recommended the implementation of the Crisis Management System (CMS) in at least one neighborhood in each borough.
The CMS was designed as a comprehensive response to violence which includes the Cure Violence Model as a core component. The other components of the CSM included mental health, job readiness, and legal advocacy services for participants and their families, as well as, conflict mediation, legal education, and community health services for the community at-large.
The Cure Violence Model was developed by Dr. Gary Slutkin, a physician and epidemiologist who spent over a decade treating health epidemics in Africa (Cure Violence 2015). The model is premised on the view that violence can be treated using similar methods employed to combat and control viral diseases and other pathogens affecting a population. It posits that violence, like all human behavior, is acquired through everyday interactions between individuals. Based on these assumptions, the model encompasses three key ideas:
In addition to the Cure Violence model, the New York City Council funded wrap-around services as part of the CMS. The Legal Aid Society provides legal services to all individuals involved in the initiative, including both community residents and program staff. Other organizations in each of the boroughs offer school conflict mediation and job readiness workshops. Community health and mental-wellness services are available to all Cure Violence participants and their family members. Taken together, these efforts are hypothesized to combine with Cure Violence in an overall reduction of violent incidents in each treatment neighborhood.
In 2014, the John Jay research team began conducting surveys with residents in four of the New York City neighborhoods where the Crisis Management System has been implemented for at least one year. Operating under the name NYC-Cure, researchers surveyed samples of young male residents between the ages of 18 and 30, the demographic most at risk for violent offending and victimization (Federal Bureau of Investigation 2013). The survey instrument was designed to measure each respondent’s attitude towards violence, as well as other factors that could influence the endorsement of violent behavior.
NYC-Cure Survey Instrument
The NYC-Cure survey was developed by the John Jay research team to measure willingness to use violence, this study’s outcome of interest, and other factors that have been shown to moderate violent behavior, such as having strong social ties. In addition, the instrument included items to measure each person’s exposure to the Cure Violence program staff in the four treatment neighborhoods.
The John Jay research team borrowed and adapted items from survey instruments used in prior evaluation studies of the Cure Violence model, including the Survey of Attitudes about Guns and Violence (SAGAS) and the Save Our Streets (S.O.S.) Community Survey. SAGAS was used as part of a John Hopkins University study of Baltimore’s Safe Streets program (Webster et al. 2012), and the S.O.S. Community Survey was used by researchers at the Center for Court Innovation as part of a study of a Cure Violence program in Crown Heights (Brooklyn, New York).
To measure willingness to use violence, the research team constructed 17 scenarios where participants selected from five hypothetical reactions. These reactions ranged from “ignore” to “use a weapon.” The scenarios included events involving competition over intimate partners, retaliation, debts or stolen property, challenges to social identity or status, disrespect, protection of others, and territorial disputes.
All surveys were administered via tablet computers. This allowed the research team to access data quickly for results and inspection. In addition, this permitted each survey respondent to engage with the survey questionnaire privately, which helped maintain the confidentiality of their responses.
The study sample was identified using “respondent-driven sampling” (RDS), a method of gathering survey data quickly and efficiently from traditionally hard-to-reach populations. The method relies on field contacts and direct referrals from respondents, bolstered by the visible and constant presence of the research team.
Data collection typically required eight to ten continuous working days in each survey location. The consistency of subject recruitment was critical in maintaining the chain of referrals for each study site. Survey sessions began around 3:00 p.m. and usually concluded by 6:00 or 7:00 p.m. Seeing the research team at the same time and location for several days was important in prompting survey respondents to encourage their friends to participate.
Prior to the start of recruitment and data-collection, the research team scouted the areas in which each of the Cure Violence programs operate, called the “catchment” area, to find an appropriate location conducive to the continual administration of dozens of interviews. Each location was chosen based on considerations of safety for the research team and survey respondents, as well as the opportunity for private interaction between the respondents and survey staff.
The first day of data-collection – the “seed day” – began with the recruitment of one person – the “seed participant.” In each catchment area, the seed participant was someone who fit the sample criteria perfectly, who was willing to participate in the survey, and who was willing to assist the study team in recruiting three other subjects who fit the criteria and who resided in the catchment area. Each of those subjects was then asked to recruit another three subjects, etc.
The average time required to participate in the study was 20 to 30 minutes. The process began with a research field supervisor screening each subject to ensure that he met the study’s sample criteria (i.e., males between the ages of 18 and 30 years who lived in the catchment area). Eligible residency was determined using maps and a list of addresses within the catchment areas. After screening, each subject was introduced to a survey team member who explained the study and obtained verbal consent before initiating the administration of the survey.
At the conclusion of the survey, respondents were paid $30 cash and given three numbered coupons. Each respondent was then encouraged to refer up to three friends who fit the study’s criteria in exchange for an additional $10 incentive payment for each new recruit who successfully completed the survey. Using this strategy, the study was able to generate survey samples of nearly 200 subjects in each neighborhood. For this initial phase of data-collection, a total of 771 young men participated in the NYC-Cure survey across four study sites.
Description of the Sample
Forty-five percent of the survey participants were between 18 and 20 years of age and 32 percent were between ages 21 and 24. Over three-fourths (78%) of survey respondents reported having been stopped, questioned and frisked in the past 12 months from the time they were surveyed (see Table 1).
More than half the survey respondents (55%) reported being unemployed and a quarter reported working only part-time. Two in ten reported not having graduated from high school, while 57 percent had a high school diploma or GED. Seventeen percent had some college education and one percent had earned a bachelor’s degree. Several other aspects of the subjects’ lives were explored using the NYC-Cure survey instrument, including their sleeping habits, time spent in the neighborhood, and participation in non-“street” activities, such as school, work, and engagement in civic programs.
More than two of every five (43%) respondents had been “shot at” or stabbed in their lifetimes. Roughly a third of all respondents reported having been the target of gunfire before (Figure 1), and one in nine (11%) had been both shot and stabbed at some point in their past.
The survey used several measures of social ties and social cohesion adapted from items reported by the Project on Human Development in Chicago Neighborhoods: Community Survey, 1994-1995, as well as two original items constructed by the research team. On average, participants reported daily communication with six neighbors. Half the respondents reported never having a conflict with a neighbor and nearly three-fourths (73%) recounted helping a neighbor within the last week.
Confidence in Formal Institutions
Previous research suggests that trust in formal institutions, especially police, may be critical in maintaining general compliance with legal norms (Tyler 2004). The NYC-Cure survey included items to measure respondents’ trust in several types of formal institutions, including police, fire department, emergency medical services (EMS), general “public leaders,” and social programs. In the study sample, more than half of survey respondents (53%) reported low levels of trust that the police department would “help” in times of distress (Table 2).
In contrast, more than two-thirds (69%) reported that the fire department would be helpful in the event of violence. Respondents had less confidence in public officials. Only 25 percent trusted that their local public leaders would be of assistance in the aftermath of a violent event, while more than half (51%) did not believe that they could count on their public officials to be helpful.
Study participants were asked about their general awareness of neighborhood gun violence (Figure 2). More than 40 percent of survey respondents reported they had personally seen at least one gun in their neighborhood in the past year, and nine percent had seen more than 10 guns in the same period. Almost one quarter of respondents (24%) reported hearing more than 10 gunshots in the past year, while 62 percent had heard of someone in the neighborhood who had been threatened with a gun during the same time period.
Perceptions of Safety
The research team used a four-item construct to measure perceptions of safety in the home and the neighborhood. Despite the high levels of neighborhood violence reported, survey participants felt high levels of safety within their communities; the vast majority (89%) of the sample felt “very safe” at home and in their neighborhood at all times of the day. Only about fifteen percent of the sample did not feel safe at home or in their neighborhoods.
This study’s main outcome of interest is to measure changes in violent norms in areas with Cure Violence programs over a three-year period. The following results give an overview description of the relationships between willingness to use violence and factors known to play a role in mediating violent tendencies. Changes in propensity towards violence during follow-up survey administrations will be used to determine whether violent norms are increasing, decreasing, or staying the same.
Self and Peer Violence Index
The research team created a violence index as a composite score from all seventeen hypothetical scenarios. Each possible response was assigned a value from one to five, ranked in order of severity. An “ignore” response received a value of one while “use a weapon” received a valued of five (Figure 3).
A peer violence index was created using the same coding scheme to measure each respondent’s estimate of how the person who referred them to the study might respond to the same hypothetical scenarios. This allowed the research team to rank all hypothetical scenarios by the severity of responses from both participants and referrers.
In all hypothetical scenarios, survey participants believed that their peers would respond more violently than they would themselves in particular situations. This is consistent with existing studies that have noted tensions between peer pressure (Prinstein, Meade and Cohen 2003; Romer et al. 1994) and the influence of social desirability on survey responses (Brown, Clasen and Eicher 1986; Chung and Monroe 2003).
The difference between participant and peer responses was greater in the least provocative scenarios (Figure 4). Survey respondents ranked the scenarios involving the need for protection (e.g., being physically attacked at a party or witnessing a friend being physically attacked) as warranting the most violent reactions, and they reported no difference between their own reactions and the likely reactions of their peers to such provocations. The five provocations eliciting the most severe responses, as predicted by respondents both for themselves and their peers, all involved an element of past or present physical threat.
After the scenarios presenting physical threats, respondents ranked the most serious provocations as those involving property disputes, disrespect, and competition over intimate partners, in that order. The scenarios garnering the least violent response also exhibited the greatest difference between self-reported violence scores and those predicted of peers. This might be due to the influence of social desirability – the notion that survey respondents tailor their answers to the expectations of survey administrators. Or, it may be an accurate reflection of how violent behavior is communicated through peer pressure and social expectations. In other words, people are more likely to choose a violent response to a conflict if they believe that their peers would do so as well, and young men from vulnerable communities may over-estimate the extent to which violence is expected in any given situation. Anecdotal evidence from this study suggests that this is so, as Cure Violence program staff indicate that petty conflicts often result in the need for intervention (Delgado et al. 2013).
Age, Education and Employment Status
The study results support prior research showing that the likelihood of engaging in high-risk behaviors decreases significantly with age and maturity (Blokland and Nieuwbeerta 2010; Farrington 2003). Older respondents reported lower tendencies toward violence. When sites are examined individually, however, this finding proved significant for only two of the sites: the South Bronx and Harlem.
Despite strong support for the mitigating impact of education on crime in criminological literature, this study found no such effect. In other words, those with higher education levels did not seem to report a lower willingness to use violence. However, employment did appear to mitigate resort to violence; those respondents with more stable employment (i.e. Full-Time) were less likely to endorse the use violence in response to one or more hypothetical scenarios.
The study tested the relationship between personal victimization and willingness to use violence. Survey respondents who reported being stabbed demonstrated a significantly greater propensity towards violence than those who reported never having been stabbed. Survey respondents who reported having been shot in the past had the highest violence index scores, even when they had not been stabbed before. These findings support the idea that young men who report having directly experienced violence have a higher likelihood of becoming perpetrators of violence (Figure 5). As the level of previous victimization increases, willingness to use violence increases as well.
Trust in Formal Institutions
Recent research studies suggests that belief in the legitimacy of public institutions, such as the police and governing officials, is linked to an individuals’ compliance with the law and desistance from anti-social behaviors, such as the violent use of guns (Tyler 2004). This notion is supported by the findings in this study. Overall, survey respondents who reported greater distrust in formal institutions also reported a greater willingness to use violence when compared to those who had more trust in formal institutions. To determine which institutions contributed most to the finding mentioned above, the research team analyzed each institution individually. Respondents who reported less confidence in public leaders (e.g. City Council members) expressed greater willingness to use violence. The same was true with regard to those who had low confidence in the police, as this relationship was even stronger. In other words, lack of confidence in the police was strongly associated with greater willingness to use violence.
Neighborhood Social Ties
The research team tested the relationship between neighborhood social ties and willingness to use violence. Amongst this sample of young men in the four treatment neighborhoods, items used to measure neighborhood social ties were not correlated with survey respondents’ willingness to use violence. On the other hand, having engaged in a dispute with a neighbor in the past was significantly associated with increased willingness to use violence. However, it is important to note that these findings were not uniform across all neighborhoods. Respondents in the South Bronx and South Jamaica were found to be significantly more likely to engage in violence if they reported having had a conflict with a neighbor in the past, while this association was insignificant in both East New York and Harlem.
Perceptions of Safety
The research team found some unexpected results when examining the relationships between perceptions of safety and violent behavior. Almost 9 out of 10 survey respondents reported feeling safe in their neighborhoods and homes at all times of the day. The association between perceptions of safety and willingness to use violence was positive. As a respondent’s level of safety increased, so did willingness to use violence. To further explore this finding, the research team tested the association between the number of guns seen in the past twelve months and safety. Surprisingly, there was a modest (.30) and positive correlation. In other words, the more guns respondents reported seeing in the past year, the safer they felt. Prior studies have shown that the need for protection is a major motivating factor for carrying a weapon (Wilkinson 2001). Individuals living in environments with high levels of gun violence should be expected to have a greater likelihood of being armed than those living in crime free neighborhoods. The study will continue to explore this finding.
Exposure to Gun Violence
Exposure to gun violence was significantly associated with willingness to use violence, with the exception of survey respondents in East New York. In general, survey respondents who reported higher exposure to guns over the past year were more willing to resort to violence (with some variation). The number of guns seen by respondents in the previous year was a significant predictor of their willingness to use violence, but there was not a uniform increase with each additional gun seen. Willingness to use violence also generally rose with the frequency of gunshots heard in the neighborhood.
The research team tested the relationship between social media threats and willingness to use violence. Across all sites, the number of threats observed by survey respondents on social media platforms was associated with willingness to use violence (Figure 6). Social media interations are known to play a role in youth violence (Patton et al. 2014).
Exposure to Programming
Finally, the research team looked at respondents’ exposure to the Cure Violence program. Cure Violence staff work hard to establish the program’s visibility in the community by holding public events, posting public education materials in prominent locations, and being out in the neighborhood meeting with residents as often as possible.
In general, the percentage of respondents reporting at least some familiarity with Cure Violence public education materials was high (80%). Survey respondents in South Jamaica reported the lowest recognition of public education materials (58%), but this could be due to the relocation of that program’s catchment area boundaries in 2014. Programs in East New York and the South Bronx had the highest percentage of recognition (93% and 90%, respectively) (Figure 7).
Survey participants were shown unlabeled photographs of Cure Violence staff members as an additional measure of program familiarity (Figure 8). Across all sites, most (59%) respondents recognized at least one staff member. East New York, Brooklyn achieved the greatest (79%) staff recognition, while just over half the respondents in other sites recognized a staff member.
Challenges and Limitations
The RDS approach is an accepted recruitment technique for sampling high-risk and hard-to-reach populations, and it is capable of generating reliable population estimates (Heckathorn 2011). It is not a probability sample, however, and there is always the possibility that surveyed individuals may not be representative of all neighborhood residents. The study was careful to ensure that each respondent fit the selection criteria, that each was surveyed only once, and that the RDS sampling structure was maintained. In several instances, however, the research team had to recruit more than one seed (first person recruited) in a single site. This required the research team to restart the recruitment process, which may affect the integrity of the study sample.
This report presents preliminary data from the NYC-Cure study at John Jay College. The findings present a baseline against which subsequent data collections will be used to estimate programmatic impact and effectiveness. No conclusions may be drawn yet about the effects of New York City’s efforts to shape violence-related norms and attitudes using the strategies of the Cure Violence program. This analysis, however, provides a starting point from which evaluators may be able to compare changes in city neighborhoods with and without the presence of violence reduction programs.
Blokland, Arjan. J., and Paul Nieuwbeerta (2010). Life course criminology. International Handbook of Criminology, 51–94.
Brown, Bradford. B., Donna R. Clasen, and Sue A. Eicher (1986). Perceptions of peer pressure, peer conformity dispositions, and self-reported behavior among adolescents. Developmental Psychology, 224 (4), 521-530.
Butts, Jeffrey A., Caterina Gouvis Roman, Lindsay Bostwick, and Jeremy Porter (2015). Cure violence: A public health model to reduce gun violence. Annual Review of Public Health 36: 39-53.
Chung, Jane., and Gary S. Monroe (2003). Exploring social desirability bias. Journal of Business Ethics, 44 (4), 291–302.
Delgado, Sheyla., Laura Negredo , Popy Begum, Michelle A. Cubellis, Alana Henninger, Doug Evans, Kathleen Tomberg, and Jeffrey A. Butts (2013). Assessment of gun violence reduction strategies in New York City. New York, NY: Research & Evaluation Center, John Jay College of Criminal Justice, City University of New York.
Farrington, David. P. (2003). Developmental and life course criminology: Key theoretical and empirical issues – The 2002 Sutherland award address. Criminology, 41 (2), 221–225.
Heckathorn D.D (2011). Comment: Snowball versus respondent-driven sampling. Sociological Methodology 41: 355-366.
Patton, Desmond U., Jun S. Hong, Megan Ranney, Sadiq Patel, Caitlin Kelley, Rob Eschmann, Tyreasa Washington (2014). Social media as a vector for youth violence: A review of the literature. Computers in Human Behavior, 35, 548-553.
Prinstein, Mitchell. J., Christina S. Meade, and Geoffrey L. Cohen (2003). Adolescent oral sex, peer popularity, and perceptions of best friends’ sexual behavior. Journal of Pediatric Psychology, 28 (4), 243–249.
Romer, Daniel., Maureen Black, Izabel Ricardo, Susan Feigelman, Linda Kaljee, Jennifer Galbraith, Rodney Nesbit, Robert. C. Hornik, and Bonita Stanton (1994). Social influences on the sexual behavior of youth at risk for HIV exposure. American Journal of Public Health, 84 (6), 977–985.
Tyler, Tom. (2004). Enhancing police legitimacy. The ANNALS of the American Academy of Political and Social Science, 593, 84-99.
Webster, Daniel.W., Jennifer M. Whitehill, Jon S. Vernick, and Elizabeth M. Parker (2012). Evaluation of Baltimore’s Safe Streets program: Effects on attitudes, participants’ experiences, and gun violence. Baltimore, MD: Johns Hopkins Center for the Prevention of Youth Violence.
Wilkinson, Deanna. L., and Jeffrey Fagan (2001). What we know about gun use among adolescents. Clinical Child and Family Psychology Review, 4 (2), 109-132.