How do we navigate public safety while utilizing community assets that can have a measurable impact on youth and families in Salt Lake County?
How do we hold youth accountable, promote public safety, and intervene at the right windows to reduce contact with the system?
Three parts. One question. Each section builds toward an answer. Click any section to jump ahead, or scroll to explore the full narrative.
What drives youth violence, where it concentrates, and why these neighborhoods produce repeat contact across generations.
How structural conditions produce the behaviors the system responds to, and where the interception points are.
What works, what's already working locally, and how to deploy interventions where the data shows the greatest return.
What drives youth violence, where it concentrates, and why these neighborhoods produce repeat contact
If we want to reduce gang involvement, we have to understand what gangs are replacing — and offer something better.
Gang involvement is a rational response to constrained choice sets. When prosocial pathways to belonging, identity, safety, and economic participation are absent, gangs become the enticing alternative. Effective prevention must restructure the incentive landscape by replacing the returns to gang involvement with higher-return prosocial alternatives.
Click any card to expand · Toggle to see matched interventions
The policy implication: Suppression-only strategies are limited in effect because they remove the supply without addressing the demand. Effective prevention must restructure the incentive landscape by offering higher returns to prosocial participation than gang involvement currently provides. The care desert maps identify exactly where those neighborhoods are.
These six incentive domains map onto the risk and protective factor framework identified across the gang membership literature. The Good Lives Model extends this by reframing intervention around building "a life worth living" rather than suppressing risk. Mallion & Wood (2020) →
The gang incentive structure is one manifestation of a broader pattern. The CDC identifies the community conditions that produce it.
The CDC identifies risk and protective factors at four levels. Most interventions target individuals. Our spatial analysis targets the level with the largest population impact: community conditions.
CDC Youth Violence: Risk and Protective Factors →
| Domain | Risk Factors | Protective Factors |
|---|---|---|
| Individual | Offence supportive cognitions, negative life experiences, low self-esteem, internalizing/externalizing behaviors, impulsivity, lack of prosocial activities, PTSD, anxiety, substance misuse, low empathy, high callous-unemotional traits, poor interpersonal skills, anger rumination | Effective coping strategies, high emotional competence, emotion regulation skills, resilient temperament, future orientation, impulse control, low ADHD symptomology, high self-esteem, intolerant attitude towards antisocial behavior |
| Peers | Negative peer influence, association with delinquent peer group, victim or perpetrator of bullying, alienation from prosocial peers, strong emotional connection to delinquent peers, peers' substance misuse | Interaction with prosocial peer groups, strong social skills, low peer delinquency, prosocial bonding |
| Family | Poor parental supervision and monitoring, lack of attachment to parents, family involvement in street gangs, delinquent siblings, hostile family environment, parental substance misuse, inconsistent discipline, low familial socioeconomic status, single-parent households, childhood maltreatment | Strong parental monitoring, control and supervision, parental warmth, cohesiveness within the family, positive parental attachment, stable family structure, low levels of parent-child conflict |
| School | Poor academic attainment, lack of commitment to education, lack of aspirations, unsafe school environment, suspension/exclusion, truancy, inconsistent discipline, victimization at school, inadequate teaching, negative relationships with staff | Positive child-teacher relationships, clear familial expectations regarding schooling, personal commitment to education, positive role models, fair treatment from teachers, safe environment, connectedness, regular school participation, academic achievement |
| Community | Disorganized neighborhood, high rates of crime, exposure to street gangs and violence, availability of firearms, poverty, lack of community resources, experiencing unsafe environments | Opportunities for prosocial involvement, positive community role models, perceived neighborhood safety, low economic deprivation |
Adapted from Mallion & Wood (2020), Table 1. Sources include: Home Office (2015), Lenzi et al. (2018), Mallion and Wood (2018), Melde et al. (2011), Merrin et al. (2015), O'Brien et al. (2013), Raby and Jones (2016), Smith et al. (2019). Good Lives Model: Street gang intervention →
Why we focus on community: Risk and protective factors exist at every level. But individual, peer, and family factors require case-by-case assessment. Community factors can be spatially measured, mapped, and targeted at the population level. Our maps measure exactly where community protective factors are absent — the domain where place-based investment has the widest reach.
Juvenile court referral hotspots and adult correctional risk converge in the same places. Zoom into Salt Lake County. Drag the slider to compare.
Scroll down to explore each layer and see where Salt Lake County's protective infrastructure is strongest and where critical gaps remain.
Let's explore these neighborhoods layer by layer.
Eight layers of spatial evidence · Each narrative card changes the map
This map shows youth enrichment reach: how many organizations a child in each neighborhood can actually get to. Green = strong care ecology. Red = care desert.
tracts confirmed as care deserts by all four estimation methods
Sharkey's IV design across 264 cities: every 10 additional community nonprofits per 100K → a 9% reduction in murder. Cumulative long-term effects of 12% reduction in homicide, 10% in violent crime, and 7% in property crime. The previous map showed you where they're missing.
Sharkey et al. (2017). American Sociological Review.
The care desert tracts from the previous view are the same tracts with the highest suspension rates. Middle schools average 9.9%: nearly 1 in 10 students excluded. The overlap is not coincidental.
Middle school avg suspension rate · CRDC 2020-21
Schools don't just exclude students; they refer them directly to law enforcement. The same tracts with the highest suspension rates also have the highest school-based LE referral rates.
Referrals to law enforcement · CRDC 2020-21
Children with neurodevelopmental disorders are at 5× higher risk for mental, emotional, and behavioral problems. Early screenings for ADHD, autism, and learning disabilities are important early prevention windows.
The pattern repeats across generations. Juvenile incarceration decreases high school completion and increases adult incarceration. Concentrated disadvantage effects on recidivism are significant for juveniles but not adults. The neighborhood environments matter most during adolescence!
Doyle (2007) · Jacobs et al. (2020) · Wu et al. (2024)
Serious offenses (violent, weapon, or firearm) make up less than 10% of all juvenile episodes, but their geographic concentration mirrors the same high-risk tracts we've seen across every other layer.
Our SLCO spatial study found that a greater concentration of female-headed households is associated with increased juvenile court referrals. How can we mitigate juvenile justice contact by supporting the households and the communities they live in?
"The most powerful determinant of where you end up in life is the zip code where you grew up."
Raj Chetty, Harvard University · Opportunity Insights
This map shows the fraction of children born 1978-1983 who grew up in each Salt Lake County neighborhood and were incarcerated on April 1, 2010 — when they were 27 to 32 years old.
These are the outcomes of the previous generation. The darkest red tracts — concentrated on the west side — are the same neighborhoods our care desert maps identify today. The community conditions that produced those outcomes have not changed.
Earlier intervention = larger returns. Chetty's MTO experiment proved that moving before age 13 changed adult outcomes. Moving as an adolescent had slightly negative impacts. The interception window is widest in early childhood.
The Opportunity Atlas tracks 20.5 million children from the 1978-1983 birth cohorts using de-identified Census and IRS data. Incarceration = being in a correctional facility on April 1, 2010 (Census day). The MTO experiment (1994-98) randomly assigned 4,600 families housing vouchers to move from high-poverty public housing.
Child Opportunity Index 3.0 · diversitydatakids.org
The Child Opportunity Index confirms the same pattern from the education side. The west-side tracts that show the highest incarceration rates on the Opportunity Atlas also score "Very Low" on:
Incarceration outcomes and educational outcomes are shaped by the same neighborhood conditions. The children in the darkest tracts on the incarceration map are also in the lowest-opportunity tracts on the education map.
Every layer of evidence we've examined converges on the same census tracts. The spatial data, the causal research, the education outcomes, and the household burden all point to the same places.
These neighborhoods had limited access to protective factors for the previous generation. This still holds true for the current generation.
Place is not a backdrop. It is the mechanism.
How structural conditions produce the behaviors the system responds to, and where the interception points are
skip to solutions →Understanding the problem is necessary. But how do we balance accountability with prevention?
We acknowledge that accountability, including retribution, deterrence, and incarceration, is part of the response. But the majority of justice-involved youth will reenter the community. The question is: what type of response produces the best outcomes for public safety?
The system's resources should concentrate on high-risk youth. The community's resources should concentrate on prevention in the neighborhoods producing the most contact.
Vincent, Skeem & Weber (2024). Youth Reoffending: Prevalence and Predictive Risk Factors in Two States. UMass Chan / CSG Justice Center · Full PDF →
ACEs interact with community conditions in a dose-response pattern — each additional adverse experience increases the odds of justice involvement. But the pathway is not inevitable. We're not talking about throwing lifelines to youth already in the river. We're talking about building the fences upstream. Click each stage to see where protective factors can redirect the trajectory. See the structural conditions beneath this pathway →
Click any stage to see details and the interception point
Poverty, housing instability, food insecurity, community violence, parental mental health and substance use, and incarceration in the family. These upstream drivers produce the chronic stress that becomes ACEs. Scarcity narrows cognitive bandwidth, reducing working memory, planning, and impulse control.
Restorative justice can reduce suspensions and improve relationships, but it cannot eliminate poverty, stabilize housing, or change neighborhood conditions. Lasting change requires upstream supports (housing, income, healthcare access) combined with midstream interventions (counselors, youth programs, credible messengers).
The strongest results come from layered approaches: economic stability + family support + school climate + mental health services + diversion.
Justice-involved youth are 12× more likely to have experienced at least one adverse or traumatic experience vs non-system peers. Each additional ACE increases the odds of justice contact by 1.68×. More ACEs = younger age of first offense.
Parents with 4+ ACEs have children with 3.22× higher odds of conviction before age 26. The cycle is intergenerational — and geographic. The same neighborhoods produce ACEs across generations.
Early childhood trauma alters two critical brain regions: the prefrontal cortex (impulse control, decision-making, planning) and the amygdala (threat detection, emotional regulation, fear response). The brain adapts to a threatening environment by remaining in a heightened state of alertness, which can look like "behavioral problems" in a classroom setting.
Interception point: 🩺 Pediatric screening can identify trauma-related developmental delays before school entry — routing children toward treatment instead of discipline.
Children with neurodevelopmental disorders are at 5× higher risk for mental, emotional, and behavioral problems (AAP 2025). ADHD (3.1× justice involvement), undiagnosed ASD (misidentified as defiance), and dyslexia (academic failure → disengagement) all present as "behavioral problems" when unscreened.
Interception point: 🩺 Expanding screening beyond anxiety/depression/SI to include ADHD, autism, and learning disabilities.
When a child with undiagnosed ADHD acts out, the counselor is unavailable, and the officer is right there. The behavioral incident becomes a law enforcement contact.
Interception point: 🏫 School counselors route incidents toward support instead of enforcement. Schools with counselors and no LE have the lowest absenteeism.
43% of court-referred youth are low risk — 9 in 10 do not reoffend (Vincent, Skeem & Weber, 2024). Juvenile incarceration causes large decreases in high school completion and large increases in adult incarceration (Doyle 2007). By this point, every upstream window was missed.
Interception point: 🏘️ Community organizations provide structured alternatives as the last upstream window before system involvement. Every 10 orgs per 100K reduces murder by 9%.
Mental health alone doesn't reduce reoffending; youth need risk-reduction services at sufficient dosage.
The programs that succeed are the ones that address these four needs directly. Click each to learn more.
Often what youth struggle with most is having their basic needs met in their homes: stable housing, food security, a safe place to sleep, adults who are present and not in crisis. When these foundations are missing, no school program or youth organization can fully compensate. A child who didn't eat dinner, who slept on a couch in a crowded apartment, who watched a parent get arrested — that child arrives at school in survival mode. Our upstream structural conditions analysis measures exactly where these deficits concentrate, and our care desert maps show where the services that could stabilize families are absent.
Youth need continuity of care from a trusted adult who is there before system contact, not after. A probation officer arrives at the final stage and acts as a social worker when the system has already intervened. But a mentor, a teacher, a community member who knows the youth and stays through transitions — that is the protective factor research consistently identifies as the most important. This is what Friends of the Children provides for 12+ years: a paid, professional mentor starting at age 4. The Friend doesn't disappear when funding cycles end or when the youth moves schools. They stay. That consistency is what makes the difference between a program and a relationship.
Resources exist, but youth don't find them on their own. Someone in their life needs to recognize what the youth needs, know what resources are available, and make that connection through a supportive, trusting relationship. This is the bridge function that family liaisons at Promise South Salt Lake, Friends at Friends of the Children, and credible messengers in Cure Violence all perform. The resource itself matters less than the relationship that delivers it. A flyer on a bulletin board doesn't change behavior. A person who says "I know what you need, and I'm going to take you there" does.
Many households have already been affected by the justice system. Parents who were incarcerated, siblings who were removed, families who were surveilled. These families carry institutional distrust that makes them unlikely to seek services voluntarily, opt into screenings, or engage with programs perceived as connected to enforcement. This distrust is not irrational — it's learned from experience. Effective interventions must be embedded in community organizations that families already trust, not delivered through the institutions they have learned to avoid. This is why culturally specific organizations matter, why community research walks matter, and why the "who delivers it" matters as much as "what is delivered."
The through-line: Every effective program we highlight in this presentation, from Promise South Salt Lake to Friends of the Children to Homeboy Industries, succeeds because it meets these four needs: it addresses basic stability, provides a consistent trusted person, connects youth to resources through relationships rather than referrals, and earns the trust of families who have reason not to trust systems. The care desert maps identify where all four needs are most acute.
Framework alignment: These needs align with the Utah State Board of Education's Protective Factor Framework — Concrete Supports, Social Connections, Resilience, and Cognitive/Social/Emotional Competence. When programs like PSSL provide structured programming, they activate multiple protective factors simultaneously. Protective factors don't stand alone; building one strengthens the others.
The further downstream we wait, the narrower the window and the higher the cost. Our spatial analysis measures each window at the neighborhood level.
Early investment in care and screening catches developmental issues before they become school behavioral problems and justice contact.
AAP (2025): Neurodevelopmental disorders → 5× higher behavioral risk. Early screening redirects to treatment.
Perry Preschool (RCT, 40-yr): 36% fewer arrests by age 40. $16 return/dollar, 88% from crime savings.
Nurse-Family Partnership (3 RCTs): 56% fewer arrests at age 15. 90K violent crimes were prevented with nurse home visits during pregnancy and the first two years.
Chicago Child-Parent Centers: 29% reduction in juvenile arrests. $7 return per dollar with preschool + parent engagement in high-poverty neighborhoods.
Heckman (2010): Every dollar spent on early childhood intervention returns $7-16 in reduced crime, health costs, and increased productivity. Returns decline with age: earliest investment = highest return.
Head Start REDI Preschool (RCT): Preschool social-emotional supports → sustained fewer behavior problems in high school.
Counselors and restorative practices route behavioral incidents toward support rather than enforcement. Every child passes through this window.
Carrell & Hoekstra (2014): Lower counselor ratios causally reduce disciplinary problems and improve achievement (regression discontinuity, Florida).
RAND PERC RCT (Augustine 2018): 16% suspension drop across 44 Pittsburgh schools through restorative practices.
Restorative Justice meta-analysis (394K students): School interventions significantly reduce exclusion (p < 0.001). More effective for suspensions than expulsions.
Make-it-Right RCT (Shem-Tov 2024): RJ conferencing → reoffense rates ~⅓ lower than traditional court processing.
School teletherapy (2025, preprint): Reduced disciplinary actions and improved attendance in socioeconomically vulnerable youth.
Youth organizations, mentoring, and crisis intervention. The last upstream window before justice system contact.
Sharkey et al. (2017, ASR): IV design, 264 cities, 20 yrs. +10 orgs/100K → 9% murder. Cumulative long-term effects of 12% reduction in homicide, 10% in violent crime, and 7% in property crime.
Sharkey (2023, Vital City): NYC precinct-level replication confirmed causal relationship at local level.
Lavecchia et al. (2024): Pathways to Education in Toronto substantially reduced youth crime. Cross-national confirmation.
Branas et al. (2018, PNAS): Greening vacant lots → 29% gun violence reduction (cluster RCT, Philadelphia).
Urban Institute (2022): "Community quarterback" model — community org at center of violence reduction is most effective.
These windows don't fail independently. The neighborhoods where all three are weakest are the same neighborhoods producing the highest juvenile court referral rates, which are the same tracts you just saw on the map.
"Internet banging" — public gang taunting on platforms — normalizes violence and creates status hierarchies. Recruitment now extends beyond neighborhood boundaries. Rivalries are performed publicly.
Digital disputes escalate to physical harm. The victim-offender overlap is spatially concentrated — it occurs most in neighborhoods with the weakest institutional supports, where trauma goes unaddressed and cycles of retaliation are unchecked.
Youth with strong prosocial connections, adult mentorship, and structured activities are substantially more resilient to online recruitment. The neighborhoods with the fewest protective factors are where online radicalization poses the greatest risk. Social media accelerates risk, but it doesn't create it. The underlying vulnerability is a community-level condition that our care ecology maps measure.
The U.S. Surgeon General's 2023 Advisory found that social isolation carries health risks equivalent to smoking 15 cigarettes per day and is twice as harmful as obesity. The care desert tracts we've identified are also social isolation deserts.
Risk factors for gang joining: meta-analysis → · 2025 youth gang involvement study →
What works, what's already working locally, and how to deploy interventions where the data shows the greatest return
In Salt Lake County, these programs need to reach the communities where the need is greatest — and where they are currently most absent.
The national evidence base is compelling, but Salt Lake County has its own proof of concept. Afterschool and summer programs are functioning as both preventive interventions and diversion mechanisms, engaging youth in prosocial activities during the hours when juvenile crime peaks.
25% of all juvenile crime on school days occurs between 2 and 6 PM, the hours immediately after school when parents are often unavailable. Afterschool programs fill this gap with structured, supervised environments that build protective factors and redirect behavior.
Fight Crime: Invest in Kids / Council for a Strong America →
🗺️ Why PSSL matters for our framework: Promise South Salt Lake is a working example of what the care desert maps identify as missing. It fills the 2-6 PM gap with structured programming, builds protective factors through SEL and mentoring, provides diversion pathways that keep youth out of the justice system, and connects families to upstream services. The 71% decline in juvenile arrests and the drop in gang involvement risk from 25.6% to 7.2% demonstrate that place-based, community-wide afterschool systems produce the outcomes our spatial analysis predicts. The question is where to replicate this model next, and the care desert maps answer that question.
The first and only long-term professional mentoring program in the country. A paid, professional mentor called a "Friend" commits for 12+ years, from kindergarten through high school graduation. Each Friend works with 8-10 youth, spending 14-16 intentional hours per month with each child. This is not volunteer mentoring. It's a full-time job.
The model was founded on research showing that the single most important factor in overcoming childhood adversity is a long-term, nurturing relationship with a consistent and caring adult. By the time Friends of the Children meets them, children in the program have experienced an average of four ACEs.
A probation officer arrives at the final stage. A Friend arrives at age 4. The Friend recognizes what the youth needs, makes the connection to resources, and stays through the entire developmental pathway. This is the stable, trusted relationship that can intercept the ACEs-to-justice trajectory before it begins.
The maps identify where care deserts exist. But data alone doesn't produce change. We're taking the spatial analysis directly into the neighborhoods it describes, walking the streets with residents, service providers, and local leaders to ground-truth the data and co-produce intervention strategies.
Modeled after the Newark Public Safety Collaborative
The DICE model, developed by the Newark Public Safety Collaborative with BJA support, partners university researchers with community-based organizations to translate spatial crime data into resident-led safety strategies. Instead of top-down enforcement, communities become co-producers of public safety.
We're adapting this model for Utah. Our community research walks bring the care desert maps, school discipline data, and neighborhood risk profiles directly into the tracts they describe. Residents identify what the data misses, validate what it captures, and shape the intervention priorities.
Salt Lake County & Weber County · Care desert tracts
Walking the care desert tracts with community partners to identify service gaps the maps highlight, validate findings with residents who experience these conditions daily, and connect spatial evidence to on-the-ground intervention design. In partnership with local coalitions, county government, CCJJ is engaging educators, families, and community stakeholders through focus groups, participatory mapping, and community research walks to collect resident-informed insights on community safety and youth well-being.
The principle: Spatial analysis identifies the neighborhoods. Community engagement ensures the interventions reflect what residents actually need. Data without community voice produces policy for people. Data with community voice produces policy with people.
The maps show where. The evidence shows why. Now: what to do about it.
Five steps from spatial analysis to targeted protective factor investment
The interception window framework is an allocation tool. It translates spatial analysis into targeted action in five steps. Click each step to expand.
Los Angeles · Father Gregory Boyle, S.J.
"Nothing stops a bullet like a job."
Paid work-therapy through social enterprises. Mental health, tattoo removal, case management. Five identity-replacing services: jobs, new identity, family, sobriety, and plans for a future.
Click any program to see why it works and the evidence base
Each window maps to a set of interventions with rigorous evidence. Earlier windows produce higher returns. Click any intervention to see the evidence and how it connects to our spatial analysis.
The highest return on investment in the entire pipeline. Every dollar spent on early childhood intervention returns $7-16 in reduced crime, health costs, and increased productivity. Returns decline with age (Heckman 2010).
Maps to: Pediatric Provider Access Map
40-year follow-up. 36% fewer arrests by age 40. $16 return per dollar, 88% from crime savings.
The gold standard of early childhood intervention evidence. Random assignment of 123 African American children (ages 3-4) to high-quality preschool vs. control. At age 40: 36% fewer arrests, higher earnings, more home ownership. The $16 return per dollar is driven primarily by reduced criminal justice costs. ROI: $16.14 per $1 invested (Schweinhart et al., 2005)
56% fewer arrests at age 15. 90K violent crimes prevented nationally. Home visiting for first-time mothers.
Three independent RCTs (Elmira, Memphis, Denver). Nurse home visits during pregnancy and first two years. Reduces child abuse, improves maternal health, and produces lasting effects on criminal behavior. Utah has NFP slots but they don't reach the care desert tracts where refugee families resettle. ROI: $5.70 per $1 invested (WSIPP)
29% juvenile arrest reduction. $7 return per dollar. Preschool + parent engagement in high-poverty neighborhoods.
Longitudinal study of 1,539 children in Chicago. Combines half-day preschool with parent involvement and school-age follow-up. The parent engagement component is critical for refugee families who may distrust institutions. ROI: $7.10 per $1 invested (Reynolds et al., 2011)
Neurodevelopmental disorders → 5× higher behavioral risk. Early screening is a critical prevention window.
AAP (2025) documents that children with neurodevelopmental disorders are at 5× higher risk for mental, emotional, and behavioral problems. Early screening redirects toward treatment instead of discipline. Current SLC screening is opt-in, English-only, and excludes ADHD, autism, and learning disabilities. 74% of SLC tracts have no local pediatrician to conduct these screenings. AAP (2025) →
The pattern across all three windows: Every intervention with rigorous evidence is place-deployable. The maps identify where. The evidence identifies what. The framework matches intervention to gap. Earlier windows produce higher returns, but every window matters because the neighborhoods where all three are weakest are the same neighborhoods producing the highest justice contact rates.
When these upstream conditions are unstable, they produce the ACEs, the undiagnosed needs, and the service gaps that the three windows are trying to intercept. Without addressing these root causes, midstream interventions alone cannot break the cycle.
CTC = Child Tax Credit · EITC = Earned Income Tax Credit · NFP = Nurse-Family Partnership · Pre-K = Pre-Kindergarten
Without upstream stability and midstream interception, behavioral incidents escalate through suspension, law enforcement referral, and court involvement. Incarceration does not reduce reoffending — 70-80% of incarcerated youth are re-arrested within 2-3 years.
Our spatial analysis measures both layers: the upstream conditions (poverty, housing, healthcare access, environmental burden) and the midstream gaps (service access, counselor ratios, nonprofit density). The care desert maps identify neighborhoods where both layers are failing simultaneously.
The fiscal case for upstream and midstream investment is overwhelming. The average annual cost of incarcerating a young person now exceeds what most communities spend on an entire year of upstream support for dozens of youth.
Primarily from reduced incarceration costs. WSIPP (2004) →
Community alternatives outperform confinement. Youth.gov →
Incarceration doesn't reduce reoffending. Annie E. Casey →
The key policy lesson: The states with the most success moved money upstream, from punishment to prevention, from incarceration to investment. No single intervention is a silver bullet. Systems change + upstream investment = lasting impact. Our spatial analysis identifies exactly where those investments produce the highest marginal returns.
All spatial evidence is publicly available. Hover, toggle, search by district or county.
Top 100 neighborhoods, primary factors
Coming SoonTract-level childhood outcomes — earnings, incarceration, mobility. Chetty et al.
Explore →Click each category to expand
Public safety and youth well-being are not competing goals. When we strengthen the neighborhoods where risk concentrates, we protect both the community and the young people growing up in it.
