Organizational and coalition strategies for youth violence prevention: A longitudinal mixed-methods study

Abstract

This longitudinal study identifies espoused change orientations and actual youth violence prevention (YVP) practices over five years by 99 public and nonprofit organizations in one city. Annual key informant interviews provided both qualitative and quantitative data, including organizational collaborative network data. Data were also obtained on participation in a citywide YVP coalition, juvenile arrests and court referrals. On average, organizations both in and outside the coalition adopted a problem-focused as often as a strengths-based change orientation, and were only marginally more oriented toward empowering community members than professionals and changing communities than individual youth. Most surprisingly, YVP coalition members adopted more of a tertiary (reactive/rehabilitative) than primary prevention orientation compared to nonmembers. The number of different YVP strategies implemented increased over five years from mainly positive youth development and education interventions to those strategies plus mentoring, youth activities, events and programs, and counseling youth. Network analysis reveals dense initial collaboration with no critical gatekeepers and coalition participants more central to the city-wide organizational network. Coalition participation and total network collaboration declined in Years 3–5. Youth violence arrests and court referrals also declined. The coalition was marginally involved in successful community-collaborative, school-based interventions and other strategies adopted, and it disbanded a year after federal funding ended. Despite, or possibly due to, both national and local government participation, the coalition missed opportunities to engage in collective advocacy for local YVP policy changes. Coalitions should help nonprofit and public organizations develop more effective change orientations and implement commensurate strategies at the community level.

INTRODUCTION

Two main justifications for the importance of community-driven violence reduction programs have been identified: (1) the preferability of relying on civil society interventions rather than the fear and punishment of criminal justice institutions and (2) the reality that community-based partnerships-- including coalitions of local voluntary and nonprofit organizations, government agencies, and public and private funders—are already how most services are provided in the United States, rather than by public employees and agencies alone (Skogan, 2011). Yet community coalitions-- due to complexities of agreement on activities, interorganizational collaboration, and a host of other challenges-- are also hard to implement, sustain, and evaluate (Skogan, 2011; Sridharan & Gillespie, 2004). Those challenges, particularly a lack of research on all relevant efforts in a whole city and how they are integrated (or not) and change over time, were the driving force of the present research.

We present a five-year study of youth violence prevention (YVP) efforts by 99 public and nonprofit organizations in one southern United States city, describing and exploring the types of YVP strategies and intervention orientations, the links between YVP efforts, organization types, and coalition structure and participation, and the changes in these variables over time. Coalition participation can take various forms, but generally involves collaboration by member organizations and their staff and/or volunteers on joint funding, planning, recruitment, organizing, intervention, public education, and/or advocacy activities (Bess, 2015). The study has several key aims: (1) to identify the types of local public and nonprofit organizations engaged in YVP and which of those participate actively in a city-wide coalition of YVP organizations; (2) to identify the different intervention orientations espoused and compare those to the types of YVP strategies actually used by organizations both in and outside the coalition (which may shed more light on the causal assumptions and implicit theories of change at work among the diverse kinds of organizations engaged in YVP); (3) to determine whether organizations participating in the coalition have different intervention orientations and YVP strategies than do organizations outside the coalition; (4) to see whether the strategies and orientations adopted change over time; (5) to track trends in coalition participation over time; (6) to further illuminate goals 1, 4, and 5, we conduct a social network analysis of inter-organizational YVP collaboration over five years among all organizations, in and outside the coalition; (7) we conclude by tracking city-wide juvenile violent crime arrests and court referrals over time from two years before to two years after the coalition was active. This is the first study to include virtually all organizations engaged in YVP in an entire city and to analyze their activity and collaboration over five years.

REVIEW OF LITERATURE

Youth Violence Prevention Strategies

Known risk factors for youth violence at the individual, family, school, and community levels argue for prevention targeted at each level rather than relying heavily on reactive, after-the-fact strategies of interrupting retaliatory violence or rehabilitating offenders (Williams et al. 2007). Prevention programs and policies should address root causes of, and environmental risk factors for, violence. They should consider multiple spheres of influence such as the family and peer environments, institutional and community factors, and employ participatory methods and social mobilization (Farrington et al. 2017), each of which was a feature of the present study and particularly the community coalition at its center.

Juvenile justice programs using threat of punishment to deter youth violence are generally ineffective (Klenowski et al., 2010). In contrast, the interventions adopted by the YVP coalition in this study emphasized positive youth development programs, which aim to strengthen social, emotional, behavioral, cognitive, and moral competencies and self-efficacy, provide opportunities plus clear, shared behavioral expectations through program structure and consistency, increase healthy relationships with peers, younger children and adults, and attend to important social and environmental factors that affect the successful completion of developmental tasks. Research on the effect of positive youth development programs on youth violence is still undertheorized and inconclusive, however (Bonell et al., 2016).

A related, popular YVP strategy, mentoring programs, can prevent or reduce at-risk youth delinquency and associated outcomes (Tolan et al., 2014). Effects of mentoring vary across programs, however, and reports from interventions persistently lack details about program features and outcomes (Tolan et al., 2014). Thus our understanding of mentoring’s mechanisms and benefits remains far from complete. Further, the level of primary prevention in many mentoring programs is questionable, as often youth must be first identified as at-risk, if not already-in-trouble, before entering some programs.

School-based social-emotional learning interventions were among the most common in the coalition we examined and represent yet another popular universal YVP strategy and can reduce violence, aggression, bullying, and substance use (Taylor et al., 2017). Still, the mechanisms of effects--especially mediating, moderating and program factors--remain understudied (Waschbusch et al., 2019). Multi-component, multi-level interventions are more promising, but are hindered by their complexity in implementation (Waschbusch et al., 2019). Thus, the etiological complexity of youth violence leads to difficulty in identifying and implementing strategies that are truly preventive and maintain lasting effects. This study identified and explored eight YVP strategies used by the organizations in our sample.

Organizational Coalitions for Youth Violence Prevention

Due to the limited success by individual organizations in preventing youth violence, many communities turned to collaborative partnerships and organizational coalitions (Fagan et al. 2008; Griffith et al., 2008), which have been popular in other areas of public health promotion, especially substance abuse prevention (Johnson et al., 2017). However, not all coalitions are equally effective; prior collaboration and participation of local leaders may be more important than organizational attributes to the success of coalitions (Bess et al., 2012). For example, Griffith et al. (2008) examined YVP organizations and coalitions and found that a combination of empowering internal organizational structures and interactions between coalition members led to critical community mobilization that resulted in a variety of positive outcomes. Theoretically, such effects may result from social capital network interactions between individual actions and social structures, such as organizational or institutional transformations (Lin, 2001). Specifically, this extra-organizational change was due to networking and collaborative partnerships allowing key stakeholders to increase access to resources, political influence, community engagement, and information dissemination. Similarly, Hays et al. (2000) found that organizational and structural features of coalitions-- including intersectoral and other forms of diversity, strength of participation, and collaboration between members—matter in prevention system planning, implementation and effectiveness. Thus, while coalitions are viable options for YVP, the structure of coalitions and member diversity and relationships greatly determine their overall effectiveness (Bess et al., 2012; Bess, 2015; Griffith et al., 2008).

A promising strategy, especially for coalitions, is to go beyond mere service provision to engage in advocacy for policy change (Schmid et al., 2008). At the organizational level, Pentz (2000) found that organizations that involve community leaders in decision-making and network with other community leaders and organizations were more likely to engage in advocacy. Schmid et al (2008) showed that greater access to resources including a large volunteer base and budget enabled political involvement in nonprofits. In coalitions, Hays et al. (2000) found that political advocacy positively related to member diversity and the number of sectors of the community represented in the coalition. Surprisingly, they also found that collaboration among coalition members negatively correlated with political action (Hays et al., 2000). However, Griffith et al. (2008) contradicted these results, and showed that increased networking in YVP coalitions actually led to increased influence in political arenas, suggesting the need for more research. YVP coalitions should perhaps look to certain community coalitions for youth HIV prevention which created sustainable structural changes affecting health and juvenile justice policies, practices and programs (Chutuape et al., 2010). Clearly, more research is needed to fully understand the relationship between coalitions and their involvement in advocacy.

Shifting the Paradigm toward Strengths, Prevention, Empowerment and Community Change

The traditional medical and social work paradigms in health and human services have been criticized as too exclusively deficit- or pathology-focused, passive-reactive, professional/expert-driven, disempowering, and victim-blaming. In contrast, a “SPEC” intervention orientation promotes individual and community strengths, prevention of problems before they become intractable, and empowerment through participation in organized, community-level change efforts (Bess et al., 2009; Evans et al., 2011).

A strengths orientation emphasizes the individual and community affirmation, resilience, coping skills and resources, and ability to thrive in challenging situations rather than labeling people or neighborhoods as “dangerous” or pathological, which can become a self-fulfilling prophecy (Maton et al., 2004). Social, cultural, and physical assets are also critical to recognize, develop, and map at the community level (McKnight & Kretzmann, 2005).

Prevention aims to reduce both the incidence and prevalence of suffering, whether universal (community-wide) or selective (e.g., targeting critical developmental “milestones” or life transition points) primary prevention, or indicated or secondary prevention targeting at-risk youth or first-offenders (Kingston et al., 2016). Prevention should link to the other SPEC elements by focusing on developing youths’ strengths and empowerment through engagement in changing community conditions.

Empowerment is a multi-level (individual, organizational, community) process--involving mutual respect, critical reflection, and democratic participation--by which people gain control over their lives and a critical understanding of their environment (Perkins, 2010). Empowering YVP interventions aim to give youth greater “voice and choice” in family decisions, their schools, and the groups and institutions that affect them and so a greater sense of sociopolitical control (Christens & Peterson, 2012).

Community change is about addressing root causes (of youth violence or whatever the problem) in the historical, social, economic, or environmental context, instead of labeling, blaming or changing individuals. It is about creating new systems and structures that remove barriers to services and supports, and promoting policies that enhance community wellness and safety (such as reducing illegal gun access and increasing opportunities; Evans et al., 2011).

SPEC-oriented strategies are generally more cost-effective than traditional medical/human service models, which rely on expensive individualized care by professionals supported by large bureaucracies that passively respond only after violence or other problems have occurred (Evans et al., 2011). It is less clear how many organizations have been able to implement clearly identifiable SPEC-oriented strategies (Bess et al., 2009).

CURRENT STUDY

In sum, clear and consistent positive effects of most YVP efforts remain elusive. Important and promising elements of preventive, multi-level efforts are a variety of YVP strategies, including political advocacy, a SPECs orientation, and local coalition building. Many communities and the United States Centers for Disease Control and Prevention (USCDCP) have responded by promoting local YVP coalitions with varying success depending on coalition structure and membership composition. Information is also scarce on both the stability or evolution of YVP strategies and coalition participation over time and whether or how coalitions influence organizations’ strategy choices, with particular regard to advocacy and policy change. Our study examines what types of organizations engage in YVP and fully participate in coalitions, the range of YVP strategies used and how those compare with values of individual and community strengths, prevention, empowerment, and community change, and whether type of organization is related to actual strategies adopted.

Research Questions

This study thoroughly explores YVP efforts in one Southern US city, focusing on types of organizations, YVP coalition participation, YVP strategies (with special attention to collective advocacy, considered a strength of coalitions), YVP orientations, and tracing the links among these variables across five years. This study addresses eight questions in the context of the universe of public and private, nonprofit YVP organizations in one city, both participating and not participating in a local coalition: (1) What types of organizations were engaged in YVP and what types participated in the coalition? (2) What kinds of YVP strategies were used, to what extent, and what kinds of SPEC orientations to YVP were espoused and to what extent? (3) How did coalition participants and nonparticipants compare in both the YVP strategies used and SPEC orientations espoused? (4) How did YVP strategies used and stated SPEC orientations change over five years? (5) How did coalition participation change over time? (6) Were coalition participants more central than nonparticipants to the city’s interorganizational collaborative YVP network and did that change over time? (7) Do city-wide juvenile violent crime arrests and court referrals over time suggest anything about the possible impact of the coalition?

METHODS

Setting and Sample

Our goal was to include the complete universe of public and private, nonprofit organizations addressing youth violence in one U.S. metropolitan area, including 30 organizations participating for one-to-five years in a USCDCP-funded, city-wide YVP coalition and nonparticipating organizations. To identify relevant organizations for this study, each year starting in 2006, we consulted with a variety of key informants, including local researchers, coalition leaders, and other school and agency representatives to select a full range of organizations engaged in YVP. Participating organizations included individual public middle and high schools, city government agencies, nonprofit youth membership organizations, religious congregations, private funding agencies, neighborhood associations, and a wide variety of human service and advocacy organizations serving children and youth. Additionally, we collected sign-in sheets from organizational meetings at both the school and city-wide levels in order to ensure that we included all appropriate participants in the study. The result was an initial list of 115 potential organizations. We then verified that each was still active and focused substantially on YVP, reducing the sampling frame to 109 in Year 1 (2007), 107 in Year 2, 103 in Year3, and 99 in Years 4 and 5. We asked each organization to identify the leader or staff to interview with the most knowledge of organization’s work on YVP and its collaborations with other organizations. The response rates were 61% Year 1, 66% Year 2, 67% Year 3, 74% Year 4, 68% Year 5.

Procedures and Measures

In each year of data collection, we conducted a three-part survey with representatives from each of the participating organizations. The survey was administered by members of the research team in face-to-face interviews with participants. Using a semi-structured format, the first part of the questionnaire extracted a detailed, open-ended description of the YVP activities and resources of each organization. The central question tapping into YVP strategies and the extent of their adoption was: “Can you tell me, in as much detail as possible, how your organization is working on youth violence in the community and what you are currently doing?” Probing questions were also planned to elicit more details about activities. After the responses were elaborated, a chart was shown to the respondents with descriptions of eight possible YVP strategies to check if more strategies would be recognized by the interviewees. Responses were content analyzed and coded by members of the research team as falling into one or more of eight types of YVP strategies (see Table 2); the categories were based on guidelines published by CDC (source) and were formally defined for all researchers. Analytic memos were written by coding members of the research team to create an audit trail and to enable discussions with other team members on more ambiguous strategies. This protocol was consistently applied throughout the five years (points) of data collection.

Part Two of the survey (Years 1, 3, 5 only) contained eight items measuring the orientation of organizations’ stated current actual (not ideal) YVP practices, two of which relate to each of the four intervention orientation dimensions (Strengths: e.g., “We focus mainly on reinforcing people’s successful strategies (as opposed to trying to change problematic behaviors)”, Primary prevention: e.g., “The timing of our interventions is usually before problems develop (rather than after problems develop)”, Empowerment: e.g., “Community members seeking help are usually active participants who exercise voice and choice”, Community conditions change: e.g., “We define problems primarily in terms of community sources such as poverty and lack of health care (rather than individual and interpersonal sources)”). All responses were recoded to a five-point scale where 5 is “mostly SPEC-oriented”, 3 is “equal”, and 1 is mostly the opposite (i.e., deficit, reactive, expert-driven, and individual-change-focused).

The last part of the survey contained the organizational network questions. Respondents were shown the complete list of organizations throughout the city that participate in YVP work and asked to identify which organizations they collaborated with on youth violent prevention work over the previous twelve months for five collaboration types: information sharing, program/service delivery, resource sharing, training/education, advocacy/policy. Data were analyzed in UCINET (Borgatti, Everette, & Freeman, 2002). Based on the results, a sixth composite variable for YVP collaboration was generated to indicate the presence or absence of any relation between two organizations. To assess change in the level of network connectedness, or in this case YVP collaborative activity among network organizations, density measure was calculated for each year. To assess the relationship between organizational network position and coalition participation we used normed degree network centrality. In this case, degree centrality measures each organization’s level of involvement in the YVP network based on the number of connections it has with other actors (Wasserman & Faust, 1994). Two-tailed t-tests were then conducted in UCINET to examine group differences in degree centrality between coalition and non-coalition members.

Coalition Participation.

Organizational participation in the coalition was coded based on attendance sheets at official NCCYS events, including monthly coalition meetings, executive committee meetings, strategic planning meetings, and workgroup meetings. Those organizations that had representatives attending at least two coalition quarterly general meetings or monthly work group or committee meetings in the past 12 months were considered coalition participants, or members, for that year of the study.¹

RESULTS

Types of YVP Organizations

Once all local public and non-profit organizations engaged in YVP were identified, they were grouped into organization types according to their main focus of activity (e.g. education, health, human services), sector (municipal administration, community-based secular non-profits, religious), or primary target population (immigrants, youth). Some organizations could be placed into multiple categories, but were placed in the type that best reflected their primary mission and work. Eight major categories were used for analysis (Table 1). Youth organizations represent private-non-profit and public organizations focusing primarily on positive youth development, some run by youth. These organizations engage in a broad array of activities from mentoring and counseling to providing opportunities for out-of-school activities. Human service organizations include charity, family counseling, and other more professionalized youth service organizations such as court counselors and advocates. Community-based organizations are neighborhood associations, family resource centers, or other service nonprofits embedded in residential catchment areas and run by local citizen groups with few if any paid staff. Educational institutions include public middle and high schools and a few informal learning institutions such as the public library and non-profits focused on skill-building. The municipal administration category consists mainly of the city agencies that do not provide a specific service but manage YVP resources or decisions. Health organizations are all public or non-profits that focus on health and mental health services and prevention. Immigrant organizations are nonprofit agencies serving local immigrant communities with a variety of programs.

Table 1.

Types of Organizations engaged in Youth Violence Prevention in a Coalition and Overall

	Coalition Participants					Total Respondents
Organizational Type:	Year 1 (%)	Y2 (%)	Y3 (%)	Y4 (%)	Y5 (%)	Year 1 (%)	Y2 (%)	Y3 (%)	Y4 (%)	Y5 (%)
Youth-focused	4 (17.4)a	5 (19.2)	2 (14.3)	2 (13.3)	4 (30.8)	16 (24.2)	16 (22.5)	11 (15.9)	14 (19.7)	14 (22.6)
Human Services	8 (34.8)	6 (23.1)	4 (28.6)	2 (13.3)	2 (15.4)	15 (22.7)	13 (18.3)	18 (26.1)	16 (22.5)	15 (24.2)
Community/Neighborhood	3 (13.0)	2 (7.7)	0 (0)	3 (20.0)	2 (15.4)	10 (15.2)	8 (11.3)	8 (11.6)	10 (14.1)	8 (12.9)
Education-Schools & Admin.	3 (13.0)	3 (11.5)	5 (35.7)	5 (33.3)	4 (30.8)	10 (15.2)	13 (18.3)	15 (21.7)	14 (19.7)	12 (19.4)
Municipal Administration	1 (4.3)	1 (3.8)	1 (7.1)	1 (6.7)	0 (0)	2 (3.0)	3 (4.2)	2 (2.9)	3 (4.2)	4 (6.5)
Health/Mental Health	1 (4.3)	2 (7.7)	2 (14.3)	2 (13.3)	1 (7.7)	7 (10.6)	8 (11.3)	6 (8.7)	6 (8.5)	4 (6.5)
Immigrant	0 (0)	1 (3.8)	0 (0)	0 (0)	0 (0)	4 (6.1)	4 (5.6)	4 (5.8)	4 (5.6)	3 (4.8)
Religious congregations	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)	2 (3.0)	6 (8.5)	5 (7.2)	3 (4.2)	2 (3.2)
Total	23	26	14	15	13	66	71	69	71	62

^aPercentages are among the category participants/total respondents. For example, 4 youth organizations in year 1 (first row, first column) represent 17.4% of the coalition participants who completed the survey for that year.

Coalition participation was more likely for certain categories of organization. Human service, youth, education and community-based organizations dominated the coalition. The composition suggests that the impetus for the coalition came from professionalized youth services and government agencies. The absence of immigrant organizations and churches is noteworthy. The composition of the coalition reveals its inability, at least in its inception, to draw support from and include civil, grassroots and youth-driven organizations.

YVP Strategies Adopted

As shown in Table 2, the vast majority of organizations in our sample were engaged over Years 1, 3, and 5 of the study in various individual and group youth behavior, knowledge and skills development-focused prevention/promotion interventions. These included positive youth development training of resiliency skills and pro-social behaviors, mentoring to provide positive role models, educating youth and/or their families about the dangers of gangs, drugs, alcohol and violence, and counseling at-risk youth or those affected by violence. The number of organizations engaged in each of those strategies increased substantially over the five years. A second major category of intervention-- providing positive, adult-supervised activities and environments for youth—also increased from under a third to two thirds of the organizations. A third type of intervention— sponsoring YVP programs or organizing events (e.g., picnic, rally, summit conference, etc.) increased slightly from less than a third in Year 1 to half doing one or both of those in Year 5, with the greatest increase (from 3% to 30%) in sponsoring programs. In contrast, the final category of YVP intervention strategies – engaging in organizational level advocacy either to influence some youth violence-relevant government policy (e.g. advocating for a change in schools, policing, etc.) or even to advocate on behalf of particular youths in school, justice or other systems—was not adopted by most organizations in our sample and, perhaps surprisingly, remained very small throughout the coalition work.

Table 2.

Approaches and Orientations to Youth Violence Prevention by n and % of Coalition Participants, Nonparticipants and Total Organizations: Years 1, 3, and 5

	Participants						Nonparticipants						Total surveyed
	Y1 (n=20)		Y3 (n=14)		Y5 (n=13)		Y1 (n=46)		Y3 (n=55)		Y5 (n=49)		Y1 (n=66)		Y3 (n=69)		Y5 (n=62)
YVP Approaches:	n	%a	n	%	n	%	n	%	n	%	n	%	n	%	n	%	n	%
Youth development work	7	35.0	6	42.9	9	69.2	22	47.8	31	56.4	36	73.5	29	43.9	37	53.6	45	72.6
Mentoring	0	0	8	57.1	7	53.9	4	8.7	26	47.3	26	53.1	4	6.1	34	49.3	33	53.2
Education re violence	4	20.0	5	35.7	10	76.9	23	50.0	29	52.7	33	67.4	27	40.9	34	49.3	43	69.4
Counseling	0	0	3	21.4	7	53.9	6	13.0	24	43.6	21	42.9	6	9.1	27	39.1	28	45.2
Providing activities	3	15.0	7	50.0	9	69.2	15	32.6	26	47.3	31	63.3	18	27.3	33	47.8	40	64.5
Sponsoring programs	1	5.0	3	21.4	4	30.8	1	2.2	4	7.3	14	28.6	2	3.03	7	10.1	18	29.0
Organizing events	8	40.0	6	42.9	8	61.5	7	15.2	19	34.6	16	32.7	15	22.7	25	36.2	24	38.7
Advocacy	0	0	2	14.3	1	7.6	3	6.5	3	5.5	3	6.1	3	4.6	5	7.3	4	6.5
SPEC Orientations average scores (1–5)	Participants						Nonparticipants						Total surveyed
	Y1		Y3		Y5		Y1		Y3		Y5		Y1		Y3		Y5
Strengths (SD)	3.05 (.69)		2.96 (.78)		2.96 (1.11)		2.71 (.77)		2.98 (.75)		2.95 (.94)		2.81 (.76)		2.98 (.75)		2.95 (.97)
Primary prevention (SD)	3.28 (.72)		3.04 (1.09)		2.69 (1.05)		3.55 (.93)		3.25 (1.05)		3.59 (1.01)		3.47 (.87)		3.21 (1.05)		3.39 (1.08)
Empowerment (SD)	3.50 (.74)		3.65 (.59)		3.54 (.78)		3.51 (1.07)		3.57 (.87)		3.59 (.83)		3.50 (.97)		3.58 (.82)		3.58 (.81)
Community conditions (SD)	3.53 (.82)		3.46 (.88)		3.19 (1.16)		3.16 (1.01)		2.94 (.98)		3.26 (1.09)		3.27 (.97)		3.05 (.98)		3.25 (1.10)

^aPercentage as a proportion of that year’s participants in the coalition (minimum of 2 attendances/year), nonparticipants and total respondents.

Youth development work was one of the most common strategies by coalition participants and non-participants. The proportion of organizations adopting youth development strategies increased from under half to three-fourths from Year 1 to Year 5. There are diverse manifestations of this strategy. A middle school administrator described it as follows:

“We individually talked to (the youths) about their expectations, asked them what they wanted to do in life, where they want to go. Talk about character education, and how what you do today can follow you the rest of your life, and…think about the university you might possibly want to attend, and if you don’t want to go to college, that job that you really want to get-- your character can cause you to not make it. So we talk about pride, respect, morals, this is their school, what do they want…”

A community based service organization had a more empowering approach:

“Our youth are getting more opportunities to do their own community organizing. This summer, their whole camp experience, we just sent them on a whole-day camp…about being a citizen in your own backyard, using the resources…available to you, learning to use…(the) Bus System so that you can get to things that seem to be important to you and to get out of the neighborhood when you feel you need a break.”

Mentoring was a strategy applied in Year 1 by no coalition members and few other organizations. In later years, more organizations adopted this strategy, often by delegating mentoring to specialized partner organizations. A youth mentoring organization representative described their approach:

“Our main business is matching one on one with an adult, responsible volunteer who forms a relationship that allows them to move into teenage years and beyond with some more confidence and competence and caring…So in terms of decreasing violence, each of those elements of our volunteers’ activities obviously impacts the child’s ability to resist violence and stay away from criminal activity and to stay away from bullying and stay away from feeling isolated. They have an adult they can depend on.”

Counseling as a YVP strategy also increased over time. Like mentoring, many organizations reported that they used the services of other organizations, in this case for professional counseling. These partnerships around mentoring and counseling are typical examples of organizational collaboration for YVP, which will be explored more below.

Another YVP strategy that increased in application was education regarding the consequences of violence. The specific practices vary from bringing speakers to school classrooms to talk about bullying, to more immersive, “Scared Straight”-style juvenile awareness programs like this one:

“We take the girls to the women’s prison, we take the boys to (the state prison) and first we show a documentary, for the girls, of the women’s prison, and the one for the boys’ prison… Common sense will tell you that once these individuals see and hear what prison life is all about, then only a fool would continue down that road.”

The strategy of providing positive activities for youth has two distinctive approaches. Some of our participants were providing a safer, more respectful and esteem-building environment for ordinary youth activities through supervision; others created alternatives for how youth spend their time. Here are two examples from a school and a youth organization:

“Our teachers when the classes change, they walk their children to the next class if they’re leaving the immediate area…We have worked at improving the climate by the way the adults talk to adults or the way adults talk to students here…And what is has done is opens them up to someone saying a kind word to them.”

“We do the video production program. We…do arts. We have a new…jewelry apprenticeship program…I want to run it on a co-op where the kids can make money off of it and learn how to manage money.”

Sponsorship of YVP programs became much more prevalent after Year 1. This change is mainly due to the changing concept of sponsoring for participants. From a narrow meaning of “funding programs” and “administering public grant funds,” the understanding of sponsorship changed to a broader support for programs–organizational, logistic, promotional.

Organizing YVP events was also a strategy involving active collaboration. These events ranged from political advocacy to more typical community-building events. Informants from a large human service nonprofit, a small youth organization, and a church, respectively, said:

“During the Mayor’s Campaign, we held a quality of life debate here, and all the candidates came.”

“Things like public conversations with law makers and decision makers and legislators. We do events with law enforcement where they do…public outreach.”

“We have a Friends and Family Day. It takes place here each year during the summer. It’s a weekend where we have games, activities, food on the grounds, and when this handout said youth prevention, that is strictly geared toward the family on that day.”

Advocacy for policy change was far less common than other strategies. We suspect that the number of organizations engaged in advocacy may be underestimated because in year 3 we included an additional follow-up question that would have brought the total from five to twelve; however, for consistency of measurement, we did not include responses to the new question in Table 2. Another observation is that some organizations (not included in the count) reported advocacy on behalf of individual students instead of policy change. Contrast the following example of political advocacy with one of individual advocacy after that:

“We are actually meeting with state legislators next week to discuss our position with (pre-kindergarten) and urging them to help support ‘pre-k’ in the Governor’s budget. We have a government relations volunteer committee that has helped us craft a public policy position statement that our board has adopted, not only on ‘pre-k’ but on three or four other issues that are important for us.”

“Once a child is ready for step down,…we are active in our advocacy work for the child. There were things like sending children to a site, a group home…So we send these kids in shackles and handcuffs and orange jump suits, and we objected to this. We said, you know, you don’t have to treat them like that.”

Change in YVP Strategies over Time and Types of Organizations Using Them

Although not included in Table 1 or 2 due to space constraints, in this section we summarize the types of organizations engaged in each YVP strategy.

Youth development work.

In Year 1, human service and youth-focused organizations were the types most engaged with youth development work. Schools, community-based and immigrant organizations also focused on youth development. In Year 3, human service and youth-focused organizations were joined by churches and education organizations in doing youth development interventions, while in Year 5 the leading organizations expanded to human service, youth-focused, schools, community-based, and health organizations.

Mentoring.

Mentoring was not a common strategy in Year 1 (as seen in Table 2) and no organization type stood out as a provider. However, by Year 3, the number of organizations engaged in formal or informal mentoring had increased sharply with youth-focused and education organizations frequent providers, and community-based and human service organizations a secondary prominent group. In Year 5, the proportion of organizations involved in mentoring continued to expand slightly, including the majority of community-based, human-service, youth-focused, church and school organizations.

Education regarding violence.

This was a common strategy of YVP from Year 1. The most frequent providers were by far schools, then immigrant, health, community-based, and youth-focused organizations. In Year 3, human-service organizations joined the list of frequent providers, while most immigrant organizations no longer used the strategy. The same kinds of organizations were observed in Year 5, with an even higher proportion of organizations engaging in education about youth violence in every group except immigrant organizations (see Discussion), and with government agencies joining the effort.

Counseling.

Very similarly to mentoring, counseling was uncommon in Year 1 but expanded dramatically by Year 3. Community-based, human-service, youth-focused, educational and health/mental health institutions were all leading providers of counseling in Year 3. In Year 5, those same types were still likely to use counseling as a strategy.

Youth activities.

As expected, the main providers of youth activities were youth-focused organizations, schools, and human-service organizations in the first year. In Year 3, the number of organizations almost doubled, and community-based organizations became frequent providers as well. In Year 5, the core of providers was expanded with churches.

Sponsoring programs.

This remained a less-common strategy throughout the five years of the study, but did increase over time. In Year 5, some youth-focused and community-based organizations and municipal agencies reported this kind of engagement as well.

Organizing events.

In Year 1, community-based and human-service organizations were the most frequent YVP event organizers, although neither group had more than half of its members using this strategy. In Year 3, churches, education and community-based organizations were the most frequent, followed by youth-focused and human-service organizations. In Year 5, the list shrank to mainly the community-based, human-service, and municipal organizations.

Advocacy.

This was not a common strategy throughout the study, and the small numbers do not allow for distinctions or tracking of trends. Across the five years, youth organizations were most active in advocacy, followed by human service and health organizations.

Espoused Intervention Orientations: SPECs

The bottom of Table 2 presents intervention orientation means on the Strengths, Prevention, Empowerment, Community Conditions Change (SPECs) scales in Years 1, 3, and 5 for participants in the coalition, non-participants, and the total sample. Repeated measures ANOVA tests were conducted for each orientation only for the total sample, because participants and non-participants changed over time. T-tests were conducted for each year for each orientation between participant and non-participant groups. The overall level of SPECs orientations espoused by coalition participants and nonparticipants was only moderate. On the 1–5 scale, they ranged from a low of 2.69 (participants’ prevention orientation at Year 5) to a high of 3.59 (nonparticipants’ prevention and empowerment orientations at Year 5). We did not find the expected skew toward espousing SPEC principles, which would have limited any possible increase in them over time. Despite that, there were few statistically significant changes or differences in SPECs orientations. It was expected that SPECs scores might increase, particularly among coalition participants and for the prevention scale. However, the changes in SPECs scores across years were not significant: F(2, 68) = .012 for strengths orientation; F(2, 64) = .206 for empowerment; F(2, 68) = 1.133 for community change. One noteworthy exception was primary prevention (F(2, 68)=4.833, p=.011)), which in fact decreased significantly in the total sample from Year 1 to Year 3, where pairwise comparisons using Bonferroni adjustment showed a significant difference between years 1 and 3 (p=.024). Unexpectedly, the primary prevention orientation among participants was also significantly lower than for non-participants in year 5 (t(57)=2.801, p=.007). All other comparisons of SPECs means between participants and non-participants showed no significant differences.

While the other contrasts were not statistically significant, they were still suggestive (see Table 2). In Year 1, participants in the coalition had slightly higher scores on community condition and strengths orientation, while non-participants had a higher primary prevention orientation. In Year 3, the community conditions orientation of the participants group remained higher than for non-participants, while the preventive orientation of nonparticipants remained slightly higher (although both decreased), and the other two SPEC elements were similar between the two groups. In Year 5, all SPEC elements had comparable levels between participants and non-participants, except the above statistically significant difference in primary prevention orientation, which surprisingly started lower among coalition participants and decreased over time. We can also see that strengths orientation scored lowest throughout the study compared to the other elements, for the two groups and the total sample.

Overall Trends in the Coalition

As shown in Table 1, participation in the coalition declined over time. Starting with 23 active participating organizations in the first year of the coalition, it expanded to 26 in Year 2, and then declined to just 14 organizations in Year 3. Active coalition participation leveled off for the remainder of our study with a core group of 13–15 organizations.

Different types of organizations focusing on YVP joined the coalition. In Year 1, human service organizations, education organizations, and youth organizations provided 16 of 23 participating organizations. Their majority is indicative of the impetus for the coalition coming from professionalized youth services. The small number or absence of civil organizations, such as community-based, immigrant organizations, and churches prevented a broader-based, more grassroots community coalition. Coalition staff made a substantial effort to widely publicize both coalition and individual organizations’ events, and Year 2 did see some additional youth organizations, an immigrant organization and a church joined the coalition. But numbers declined again in Years 3–5. Table 1 shows how different types of organizations (among respondents) stopped participating actively in the coalition, most remarkably human service organizations. Educational and youth organizations remained the largest core group of participants, probably because several schools each had a youth services coordinator paid by the same grant that supported the coalition. Community, and human service organizations also continued to participate at about two or three apiece.

Changes in the Organizational and Coalition Network over Time

Over the five years, little change occurred in composite YVP network density representing the level of collaboration among organizations. Network density initially increased in the Year 1 (.19) and 2 (.26), decreasing slightly in Year 3 (.23), 4 (.19). By year five (.13) the level of reported network ties fell well below the level of Year 1. Network density across collaboration types (information sharing, program/service delivery, resource sharing, training/education, advocacy/policy; see Table 3) reveals a similar pattern across the five years. Overall, network participants report the highest number of ties for information sharing and program/service delivery and the lowest for advocacy/policy.

Table 3.

Network Density by Year and Collaboration Type

	Y1	Y2	Y3	Y4	Y5
Advocacy & Policy	0.02	0.04	0.05	0.03	0.02
Information Sharing	0.09	0.13	0.14	0.14	0.06
Program & Service Delivery	0.06	0.12	0.14	0.09	0.06
Resource Sharing	0.07	0.09	0.10	0.08	0.04
Training & Education	0.05	0.07	0.09	0.07	0.03
Composite YVP Collaboration	0.19	0.26	0.23	0.19	0.13

Similar to patterns of network density, mean centrality for Composite YVP for both coalition and non-coalition groups increased initially and then decreased to their lowest levels in Year 5. This pattern holds true across collaboration types as well (see Table 4). Coalition participants, however, were on average consistently more central to the whole network than were nonparticipants. Figure 1, which presents network maps for Years 1, 3, and 5 for Composite YVP collaboration, provides a pictorial representation showing coalition participants (represented by the white circles) occupying positions toward the map’s center compared to non-coalition participants’ (black circles). T-tests performed in UCINET (Borgatti et al., 2002) indicate significant between group differences in Composite YVP for Years 1, 2, 3 and 5. Between group differences were also found across collaboration types, reaching levels of significance in four of the five collaboration types in Years 1, 3 and 5, and in two in Year 2. No significant differences were found in Year 4. This may reflect that the coalition was most active in its first two years and then some core participants active in the YVP arena stopped attending coalition meetings thereby becoming noncoalition participants. Note that in Figure 1 immigrant-serving organizations were peripheral to the network despite their communities being disproportionately affected by youth violence. (See Bess (2015) for more detailed longitudinal network analyses.)

Table 4.

Difference in Mean Degree Centrality Between Participants and Nonparticipants by Collaboration Type and Year

	Year	Participant	Nonparticipant	Difference	Two-tailed T-Test
YVP Composite	Y1	46.5	25.3	21.3	0.001
	Y2	49.6	32.7	16.9	0.001
	Y3	46.5	33.0	13.5	0.01
	Y4	34.6	29.8	4.8	ns
	Y5	37.3	16.4	20.9	0.001
Advocacy &Policy	Y1	5.6	3.5	2.1	ns
	Y2	1.2	0.8	0.4	ns
	Y3	15.3	7.5	7.7	0.01
	Y4	7.3	5.6	1.7	ns
	Y5	2.0	0.7	1.3	0.1
Information Sharing	Y1	25.2	13.0	12.2	0.001
	Y2	33.5	16.0	17.5	0.001
	Y3	35.7	20.4	15.3	0.01
	Y4	25.7	25.1	0.7	ns
	Y5	22.2	7.5	14.7	0.001
Program & Service Delivery	Y1	14.1	9.0	5.1	0.05
	Y2	2.7	2.0	0.7	ns
	Y3	27.7	21.4	6.3	ns
	Y4	16.2	15.0	1.2	ns
	Y5	4.2	2.1	2.1	0.01
Resource Sharing	Y1	19.5	10.6	9.0	0.001
	Y2	2.2	1.6	0.6	ns
	Y3	23.9	14.2	9.7	0.01
	Y4	12.2	14.0	−1.8	ns
	Y5	12.1	5.4	6.6	0.001
Training & Education	Y1	8.132	4.96	3.2	0.05
	Y2	1.8	1.1	0.6	ns
	Y3	24.3	14.2	10.1	0.01
	Y4	16.4	12.4	3.9	ns
	Y5	12.9	4.4	8.5	ns

Figure 1.

Organizational Collaboration: White = Coalition Participant; Black = Non-Participant

Trends in City-wide Juvenile Crimes and Arrests

Table 5 provides city-wide juvenile arrests and court referrals from before the coalition was organized in 2007 through its main years of operation from 2008 through 2011 until after it was unfunded at the start of 2012 but continued information sharing and disbanded around 2013 (with minor population data from https://datacenter.kidscount.org). Juvenile arrests, arrests for violent and weapons offenses, homicide arrests, and case referrals to juvenile court for violent crimes all decreased substantially over the years the coalition operated, although the decrease for violent offenses was clearest as the coalition was ending. Not shown in Table 5, reports of both total juvenile crimes reported and firearm-related juvenile crimes also decreased during the life of the coalition. Limitations of the data are discussed below.

Table 5.

City-wide Juvenile Arrests and Referrals to Juvenile Court Before, During and After YVP Coalition: 2006–2013 (Juvenile Population)

	2006 (130,399)	Coalition organized: 2007 (134,728)	Coalition operated: 2008 (135,764)	2009 (136,853)	2010 (136,391)	2011 (140,365)	Coalition unfunded: 2012 (143,189)	Coalition disbanded: 2013 (144,155)
Total juvenile arrests:	15,997 (12.27%)	13,378 (9.93%)	12,667 (9.33%)	12,095 (8.84%)	10,690 (7.84%)	9,549 (6.80%)	9,473 (6.62%)	8,918 (6.19%)
Total not counting curfew or runaway:	11,524 (8.84%)	11,133 (8.26%)	10,661 (7.85%)	10,265 (7.50%)	9,082 (6.66%)	8,370 (5.96%)	8,080 (5.64%)	7,491 (5.20%)
Arrests for violent offenses:	1,524 (1.17%)	1,657 (1.23%)	1,532 (1.13%)	1,504 (1.10%)	1,408 (1.03%)	1,454 (1.04%)	1,348 (0.94%)	1,287 (0.89%)
Weapon-related arrests:	640 (0.49%)	620 (0.46%)	525 (0.39%)	438 (0.32%)	481 (0.35%)	533 (0.38%)	393 (0.27%)	410 (0.28%)
Homicide arrests:	38 (.029%)	22 (.016%)	23 (.017%)	29 (.021%)	24 (.018%)	16 (.011%)	4 (.003%)	11 (.008%)
Juvenile court referrals for violent crimes:	1,832 (1.40%)	1,575 (1.17%)	1,228 (0.90%)	711 (0.52%)	649 (0.48%)	618 (0.44%)	577 (0.40%)	424 (0.29%)

Note: percents are total arrests and referrals divided by juvenile population, but % of individual juvenile offenders would be lower based on recidivism or multiple-crime arrests.

Discussion

Summary of results

Our findings show an overall increase in the number of different YVP strategies implemented in the city throughout the five years of the study. By Year 5, youth development work, education, providing activities and mentoring were most prevalent, while sponsoring programs and advocacy were the two strategies adopted by the fewest organizations although still almost one third of the sample used those. This expansion of strategies was not specific to participants in the coalition, which saw a marked decrease in active participation starting by Year 3, despite the overall increase in YVP strategies used by both participants and nonparticipants.

Examination of the data shows that participation in the coalition made little difference for most of the strategies adopted or for espoused orientation of YVP work, except for coalition participants’ unexpectedly weak and declining espoused prevention orientation. Intervention orientations toward strengths, primary prevention, empowerment, and addressing community conditions did not increase; in fact, prevention orientation decreased significantly in the total sample from Year 1 to Year 3 and was lower among coalition participants than non-participants in year 5. In contrast to their espoused orientation, by Year 5, participants engaged heavily in education on youth violence, positive youth development activities and events, mentoring, counseling, and advocacy on behalf of youth in schools and the justice system.

In sum, although the number of YVP strategies used increased and coalition members remained more central to the citywide network, several other expected coalition outcomes did not occur: coalition participation and network collaboration decreased after the first year or two; the coalition avoided any advocacy for policy change; and organizations’ SPEC orientation toward change did not increase. The apparent disconnect may be due to coalition members’ concern for youth already experiencing problems and thus a preference for secondary and tertiary prevention over primary prevention. It may also relate to failures of the coalition to actively engage local leaders beyond the promise and excitement felt in the first year (Bess et al., 2012). Another limitation of the coalition and many individual organizations may have been insufficient empowering internal organizational structures which could have led to the wider community mobilization that was needed (Griffith et al. 2008; Hays et al. 2000). Organizations did not receive additional funding for participating in the coalition—that lack of incentives and insufficient resources and structural changes are consistent with the limitations of superficial social capital networks (Lin, 2001).

The clearer patterns related to organization type. Some strategies were more likely to be used by certain types of organizations (e.g. sponsoring programs by government agencies, youth development work by youth-focused organizations). Human service, education, and youth-focused organizations used a wide variety of YVP strategies, especially after the first year. These types of organizations also formed the consistent core of the coalition throughout the five years of the study.

What the present network analysis shows is that the overall level of collaboration was fairly dispersed and stable over time, with no single nor even just a few critical gatekeepers. At the start of the coalition (Year 1), most of its participants were core in the city-wide network, but by Years 3 and 5, some of those “key players” had stopped participating. Connections between participation in the coalition, YVP strategies adopted, and organizational collaboration are somewhat speculative, but we do know that active coalition participation declined, as is unfortunately common (Skogan, 2011; Sridharan & Gillespie, 2004), but particularly so among organizations that were core in terms of network collaboration. Yet diversification of YVP strategies increased.

Implications for youth violence prevention policy and organizational and coalition practice

Coalitions of schools, human services and other public and private nonprofit and voluntary organizations have been organized throughout the U.S., often with Federal support, to address a variety of public health issues, including substance abuse and, in the present study, youth violence (Bess, 2015; Hays et al. 2000). Even the most ambitious and comprehensive coalitions and individual organizations (e.g., Hernández-Cordero et al. 2011) primarily engage in direct social services, public education and information sharing, and school and community-based prevention programs that focus on group activities and individual responsibility (Perkins et al. 2007) and there is little rigorous evidence they significantly reduce substance abuse, violence or other crimes even in the targeted communities, let alone in the local population as a whole. In the city in this study, organizations both in and outside the coalition increased the number of different YVP strategies adopted over the five years of the study; there is also evidence that juvenile arrests, arrests for violent and weapons offenses and homicide arrests all decreased substantially over the years the coalition operated. Although arrests are not the most accurate measure of crime, rates of reported juvenile crimes, firearm-related crimes, and juvenile court referrals for violent crimes also decreased steadily each year of the coalition and for at least two years after it was unfunded. Furthermore, the fact that police were making fewer juvenile arrests and court referrals over time is itself an important goal.

The one strategy that did not increase consistently over time was advocacy, which did increase from 2007 to 2009, but then decreased. This might have been due to a perception of insufficient power to influence policy, or a misunderstanding and apprehension about restrictions on advocacy by 501c3 organizations, or to a more individualistic, instead of structural, understanding of the causes and remedies of youth violence. This occurred despite the coalition holding a strategic planning workshop in early 2010 aimed in part at addressing what kinds of advocacy nonprofit organizations can legally do (yet advocacy still dropped off slightly even among coalition members).

We know that the power and relational resource capacity of the YVP Coalition increased slightly during the five years that the coalition operated, while the capacity of the broader YVP network of which it was a part actually decreased (Bess, 2015). Although it is difficult to discern as all types of YVP strategies increased over time, regardless of coalition participation, our data suggest that the coalition may have helped increase the use of certain YVP strategies that were emphasized by the coalition, such as mentoring, public youth and adult education regarding violence, and the identification of safe places and provision of activities for youth.

The greatest surprise in the SPEC scales measuring espoused intervention orientation (Evans et al. 2011) was that organizations participating in the coalition started with a less primary preventive orientation than nonparticipants and that orientation decreased over the five years of the study and ended significantly lower than organizations not participating in the coalition. Although the researchers thought of the coalition as focusing on prevention, its organizers made an explicit decision to call it a community coalition for “youth safety” and not for “violence prevention” because they wanted it to be more focused on youth and community strengths rather than problems. Ironically, strengths were the weakest of the four SPEC orientations for both coalition participants and nonparticipants, although prevention ended even lower than strengths for participants in Year 5. Thus, there is an apparent disconnect between espoused intervention orientation and actual YVP strategies employed. Coalition participants claimed that their organizations emphasized empowerment and changing community conditions (which usually tend toward collective, structural interventions) more than prevention or strengths. But our qualitative analysis suggests their actual intervention strategies were geared more toward building individual strengths and opportunities and “indicated” or secondary prevention (with high-risk youth) rather than primary prevention. Some may have helped foster youth self-efficacy and self-esteem, if not true multi-level empowerment. But few even aimed to address root causes or conditions at the community level. There are many possible reasons for this, including funding that targets more conservative individual casework or indicated prevention rather than community change as well as organizational staff knowledge, training and professional identity constraints (Bess et al. 2009; Nation et al. 2011; Perkins et al. 2007). Whatever the reasons, the fact that SPEC orientations espoused are not clearly reflected in strategies actually implemented suggests a need for more research on that gap, funding for SPEC interventions, staff hiring or retraining, and systematic monitoring of implementation (Fagan et al. 2008).

Although coalition activity and collaboration declined after Year 2, the fact that YVP strategies diversified over time may be an encouraging outcome. Unfortunately, advocacy remained much less common than other strategies, but coalition participants were slightly more likely to engage in some form of advocacy (more on behalf of youth having difficulties with the school or justice systems than for policy change) than were nonparticipants. With assistance of policy experts, coalitions may be able to help organizations with little knowledge or experience in advocacy to overcome those limitations. Coalitions are particularly natural venues for advocacy given the clout and voice that come with the potentially large numbers of organizations and individuals they can mobilize.

The role of immigrant organizations in the complete network of institutions engaged in YVP is noteworthy. Although few in number, when asked about the causes of youth violence, representatives of immigrant organizations were much more likely than other types to cite structural causes related to racial discrimination in criminal and juvenile justice, education, and other systems and economic disadvantage. Yet unfortunately they generally did not participate in the coalition, did not engage in advocacy, and were all on the periphery of the network across the five years. A majority of them engaged in education about youth violence in Year 1, but they were the only type of organization that decreased its use of that strategy over time. This represents an important missed opportunity as greater participation by immigrant organizations in advocacy and public education about systemic causes of violence and in the coalition would not only have helped them become more central to the network and gain a more prominent voice in the wider community, but it would have helped broaden the coalition, added an important perspective to its agenda, and contributed more to reducing youth violence in the city (Hays et al., 2000). Immigrant and grassroots organizations more peripheral to the network should have been more actively recruited by the coalition.

Strengths and Limitations of the Study

We had the rare opportunity to study and participate in a collaborative and diverse effort aimed at an important social issue from its very inception throughout five years of its development. We collected longitudinal data annually measuring a host of process and outcome variables. The sample included diverse types of organizations, from schools and government agencies to community-based, immigrant, and youth-member organizations and from churches to professional health and human services. A diverse set of mixed methods were applied for data analysis: quantitative statistical methods (including a new brief SPECs survey scale of intervention orientation toward strengths, prevention, empowerment, and changing community conditions), qualitative coding and interpretation, and network analyses. The most unusual aspects of the study were the inclusion of virtually all relevant local YVP organizations, both coalition participants and nonparticipants, and our ability to track changes in the coalition and organizational strategies over five years.

The study also had limitations in its scope and consistency. The number of organizations in the sample was too small to perform some quantitative analyses, especially those applying eight or more groupings of organizations. While attrition of respondents across each wave of data collection was no more than typical, annual changes were sufficient to make longitudinal panel analyses difficult if not impossible. Furthermore, diminishing enthusiasm by some respondents for the coalition (with which our project was associated and manifested in decreased attendance), leads us to question the quality of some data from the last year of collection. Indeed, the level of detail and the consistency of responses diminished with time and significant amounts of missing data were observed on some measures in our survey. Disappointing also was the refusal of some of the coalition members, most importantly government agencies, to provide timely crime and health-related data that could help evaluate the impact of the coalition on the outcomes of interest and its purpose for existence. We reported annual juvenile arrest and court data before, during, and two years after the coalition was unfunded, but those are not enough years to provide an interpretable interrupted time series analysis. The decreasing juvenile crime rate is not necessarily attributable to the coalition. Finally, the myriad contextual specificities of both the city and the YVP coalition make it impossible to know how generalizable our findings are to other metropolitan areas or regions where the violence patterns, demographic, social, and policy-making structures, and organizational matrix may be different in important ways.

Conclusions

Most organizations addressing youth violence do so through individually-focused prevention/promotion (such as positive youth development programs that provide structured, supervised group activities), despite mixed evidence of effectiveness, because it is how administrators, staff, and volunteers have wanted to intervene and it is what has been funded. This contrasts with the theory of change implicit, if not always recognized, in the local organizational coalitions formed throughout the United States in recent decades to prevent substance abuse, violence, and other public health problems which assumes that increased contact and coordination through face-to-face meetings and information sharing will lead to greater voice, clout and impact, especially through campaigns to address structural causes of such problems.

Recently, the health department in the study city organized a new YVP coalition. Although many of the same agencies from the original coalition are again involved, there are also new players and a more explicit public health and primary prevention focus was adopted. We hope that this time espoused YVP intervention orientations will lead to commensurate actual practices.

Biographies

Douglas D. Perkins is Professor of Human and Organizational Development at Vanderbilt University and was Founding Director of the graduate program in Community Research and Action. His research focuses on participation and empowerment in grassroots organizations to improve public policy making in the context of urban neighborhood change, youth violence, crime, fear and disorder; and on the global development of applied community-focused studies in education, psychology, economic and social development, public health, urban planning, sociology, public administration/policy studies, applied cultural anthropology, Liberation Theology, and interdisciplinary community research and action.

Nikolay L. Mihaylov is an assistant professor of sociology and organizational psychology at the Medical University Varna, Bulgaria. His recent research is focused on school communities and climate, social-emotional learning, and community organizing. He received his PhD in Community Research and Action from Vanderbilt University, USA.

Kimberly D. Bess is Associate Professor of the Practice of Human and Organizational Development at Vanderbilt University. Her research focuses on community and organizational learning, intervention, and change, particularly in the human service and educational sectors. She has also studied youth and adult development and uses a mix of research methods, including qualitative case studies, surveys and in-depth interviews, narrative and content analysis, organizational and social network analysis, and community-based participatory research.