Residential Racial Segregation and Youth Firearm Aggression: Neighborhood disadvantage and exposure to violence as mediators

Abstract

Youth interpersonal firearm violence (e.g., homicides) disproportionately affects Black youth and contributes to racial health disparities. Institutional racial discrimination – in particular, residential racial segregation - is a determinant of racial disparities in firearm violence. Residential racial segregation, which is enforced to limit racially minoritized members to undesirable residential areas (e.g., fewer educational and employment opportunities), is associated with the socioecological risk factors of youth firearm violence (e.g., exposure to violence [ETV], neighborhood disadvantage). The socioecological mechanisms underlying the link between, however, remains unclear. Therefore, we tested the mediating role of neighborhood disadvantage (mediator 1) and ETV (mediator 2) in the association between residential racial segregation and youth firearm violence. Participants consisted of 338 Black youth who used illicit drugs in the past year and sought care in an urban emergency department. Using serial mediation analysis, residential racial segregation was indirectly associated with youth firearm violence via neighborhood disadvantage and then exposure to violence. Identifying the downstream socioecological consequences of residential segregation can inform the development of firearm violence prevention programs that address the socioecological consequences of racism.

Residential racial segregation and youth firearm violence: Downstream socioecological consequences

Racial disparities in firearm violence

Firearm homicide is the second leading mechanism of death for all youth (ages 14–24), but akin to other public health issues in the US, firearm homicide disproportionately affects Black youth and is the leading mechanism of death.^1,2 Moreover, firearm homicides account for 87% of all firearm-related deaths for Black youth and 38.9% of firearm-related deaths for White youth.¹ The overrepresentation of Black youth experiencing firearm injury and homicides has been consistently documented.^2,3 Fewer researchers, however, have focused on identifying the root causes (e.g., institutional racial discrimination) of racial disparities in youth firearm violence.^4–6 This is a significant gap in firearm injury prevention research as race is a social construct (distinct from biology) and operates as a proxy for the enactment of racist ideology including racially discriminatory policies and practices. ⁷ Thus, we have a critical need to assess how institutional racial discrimination contributes to racial disparities in youth firearm violence.

Institutional racism and firearm violence

Institutional racial discrimination, which refers to “…racially discriminatory policies and practices embedded in social institutions such as the government, the economy, the education system, the healthcare system, religious institutions, the family, and the media,”⁸ is a significant contributor to racial health disparities, including youth firearm violence.^4,9–11 Residential segregation, in particular, has been identified as the fundamental driver of racial health disparities. Owing to the longstanding legacy of neighborhood racial segregation in the US (e.g., Redlining),^4,12 as well as the contemporary policies and practices the perpetuate racial segregation,^5,6 Black and other racially minoritized youth are more likely to reside in neighborhoods that have concentrated poverty and limited societal resources (e.g., access to mental health care, high-quality schools). Historical and contemporary indicators of residential racial segregation have been consistently associated with firearm violence (e.g., fatal, nonfatal firearm assault injuries). Despite these interesting findings, fewer researchers have examined the pathways through which residential racial segregation increases the likelihood of youth firearm violence. To our knowledge, only a single study has evaluated neighborhood-level socioeconomic indicators (i.e., poverty, low educational attainment) as mediators to the association between historical redlining and firearm violence.⁹ Thus, while residential racial segregation plays a critical role in exacerbating racial disparities in firearm violence, less is known about how residential racial segregation perpetuates these disparities.

Socioecological mechanisms from residential segregation to youth firearm violence

Researchers have theorized multiple mechanisms by which residential racial segregation reinforces racial health disparities, including youth firearm violence. Guided by the framework for the study of racism and health,¹⁰ residential racial segregation is a fundamental cause of racial health disparities. Researchers have posited that residential segregation limits access to socioeconomic opportunities (e.g., stable labor market), societal resources and support services (e.g., access to health care, affordable housing), and increases the likelihood of encountering racism-related stressors (e.g., over-policing).^10,13 These consequences, in separate studies, have been associated with firearm violence in racially minoritized communities.^9,14,15 More recently, Burrell and colleagues posited a theoretical model by which structural racism such as residential racial segregation can contribute to community violence among young African American males living in an urban area.¹⁶ These scholars postulated that economic disenfranchisement can invoke psychological consequences such as mistrust, compliance with violent conduct norms (e.g., retaliatory attitudes), cultural disorientation (e.g., feeling disliked because of one’s race), and willingness to participate in an underground economic (e.g., illegal drug market), which, in turn, increase community tension and community violence (including firearm violence).¹⁶ Thus, when taken together, researchers have postulated various downstream consequences by which residential segregation can ultimately contribute to racial disparities in youth firearm violence. Yet, only a single study has evaluated mechanisms from historical redlining to firearm violence.⁹

To address this gap, we will leverage serial mediation analysis to evaluate whether residential racial segregation influences firearm violence via neighborhood disadvantage (mediator 1) and then exposure to violence (mediator 2; see Figure 1) among Black youth. We hypothesize that residential racial segregation is positively associated with neighborhood disadvantage and that neighborhood disadvantage is associated with higher levels of exposure to violence, which, in turn, increases the risk for youth firearm violence. We also hypothesize that the influence of residential racial segregation on youth firearm violence will be mediated by neighborhood disadvantage and exposure to violence.

Figure 1.

Conceptual model bridging residential racial segregation with youth interpersonal firearm violence.

Method

Participants

Our analytic sample consisted of 349 Black youth (i.e., ages 14–24) from a prospective cohort study (i.e., Flint Youth Injury study [FYI]).¹⁷ Youth were recruited from an emergency department in Flint, Michigan, and those enrolled were followed at baseline and then in 6-month intervals for 24 months (i.e., 5 measurement periods). The objective of the FYI study was to assess the timing and pattern of health behaviors and outcomes among drug-using, urban youth who were admitted to the emergency department (ED) with and without an acute violent injury.

Violent crime rates in Flint are comparable to other de-industrialized cities (Federal Bureau of Investigation, 2019). Table 1 provides detailed information on the demographic characteristics of the study participants. Participant addresses were geo-coded to census tracts to assess of residential racial segregation and neighborhood disadvantage.

Table 1.

Descriptive Statistics

Predictors	M (SD) or N (%)
Male	203 (41.8%)
Age	19.91 (2.37)
Public Assistance	185 (54.7%)
Violent Injury	206 (59%)
Alcohol Use	1.18 (1.19)
Tobacco Use	3.34 (2.76)
Marijuana Use	2.97 (1.22)
Internalizing Symptoms	0.65 (0.72)
PTSD Diagnosis	0.10 (0.30)
Residential Segregation (2000)	0.39 (0.62)
Exposure to violence (wave 2)	0.78 (0.77)
Neighborhood Disadvantage (2010)	0.38 (0.14)
Neighborhood Disadvantage (2000)	0.15 (0.06)
Youth firearm violence	42 (12%)
Youth firearm violence (waves 3–5)	43 (12.35)

Note. M = mean, SD = standard deviation, N = sample size, and % = proportion of the sample who self-reported into the category.

Measures

Youth firearm violence.

Three-items were aggregated to measure youth firearm violence (i.e., “in the past 6 months, you pulled a gun on someone”, “you used a gun on someone”, and “you used a gun on him/her [partner]”) at waves 3 to 5 (i.e., 12- to 24-month follow up period). Participants who endorsed higher than “never” on any of the 3 firearm violence items at waves 3, 4, or 5, received a score of 1 (i.e., firearm violence during the 12- to 24-month follow-up period). A score of 0 indicates that the participant did not engage in firearm violence during the measurement period. This variable, therefore, reflects firearm violence at least once within 12 to 24 months of their baseline visit. Using this dichotomized firearm violence variable, 12.03% of participants reported firearm violence during the past 24 months. Lastly, baseline firearm violence was also calculated at wave 1 and included as a covariate in the mediation analyses.

Residential racial segregation.

We measured residential racial segregation using the Index of Concentrations at the Extremes (ICE_race), which quantifies the social polarization of non-Hispanic, White (White) and non-Hispanic, Black (Black) households within a census tract. To compute ICE_race, we first assessed the difference between the proportion of Black (non-Hispanic) households and White (non-Hispanic) households within the census tract. This difference was then divided by the total number of Black and White households. ICE_race ranges from −1 to 1, with lower, negative scores indicating a greater concentration of Black households relative to White households. Alternatively, positive scores indicate a greater concentration of White households relative to Black households in the census tract. A score of 0 can occur for two reasons. First, and more plausibly, 0 could reflect that the census tract has no Black households and only White households. Second, 0 may reflect an equal concentration of Black and White households.

We leveraged American Community Survey (ACS) estimates for 2000 to compute ICE_race scores. Our data, however, only contained 2010 census-tract information and census tracts from 2010 cannot be directly compared to census tracts from 2000 since census tracts are updated at each decennial census. Thus, the 2000 ACS estimates used census tracts from the 2000 census. To match the census tracts from 2000 to 2010, we used the census bureau’s tract-level relationship file.¹⁸ Specifically, we matched 2010 tracts to 2000 tracts if over 95% of the population in the 2010 tract resided in the 2000 tract’s geography. Of the 176 unique tracts in the FYI data, 172 census tracts could be matched using these narrow criteria.

Neighborhood disadvantage.

To evaluate neighborhood disadvantage, we used the ACS 2010 census tract level measure of the (1) percent of households with an income below the federal poverty line, (2) percent of unemployed residents between age 16 to 65, (3) percent of residents age 21 or older who did not graduate high school, (4) percent of household receiving food stamps or the supplemental nutrition assistance program, and (5) percent of housing units that have more people than rooms (i.e., overcrowding). A unidimensional factor analytic model was estimated to develop a composite measure of neighborhood disadvantage (ω = 0.89). Lastly, we also estimated a similar factor analytic model for neighborhood disadvantage in 2000 to include as covariate in the mediation a analysis.

Exposure to violence.

Exposure to violence was measured by aggregating the participants’ response across four measures at wave 2 (i.e., 6-month follow-up): (1) violent victimization with a weapon, (2) community violence, (3) partner aggression, and (4) non-partner aggression. Violent victimization with a weapon was a five-item measure from the National Longitudinal Study of Adolescent Health (ω = 0.83).¹⁹ Participants responded on a Likert-type scale of 0 (never) to 6 (20+ times) and sample items include “someone shot you” and “Someone cut or stabbed you.” Community violence was measured using 5-items from the Survey of Exposure to Community Violence (ω = 0.81).²⁰ Participants responded on a Likert-type scale of 0 (never) and 3 (many times), and sample items include “I have seen somebody get stabbed or shot” and “my house has been broken into.” Partner aggression was assessed with 12-items from the Conflict Tactics Scale (CTS-2) (ω = .92).²¹ Participants responded on a scale of 0 (never) to 6 (20+ times) and sample items include “he/she grabbed you” and “he/she choked you.”²¹ Lastly, non-partner aggression – that is, acts of violence that happened to the participant by a non-partner such as a friend, stranger, or neighbor – was assessed with 12-items from the CTS-2 (ω = .90).²¹ Participants responded on a scale of 0 (never) to 6 (20+ times) and sample items include “someone pushed or shoved you” and “someone slapped you.”²¹ We first standardized the participants’ average score across on each measure since the measures are on different scales. Next, we fit a unidimensional factor analytic model using the four indicators to develop a composite measure of exposure to violence.

Covariates.

All models controlled for sociodemographic factors including sex (female vs. male), age, public assistance status (no assistance vs. public assistance) at baseline (wave 1). We also controlled for youth who presented to the ED at baseline with or without a violent injury. Two mental health variables were assessed at baseline - post-traumatic stress disorder (PTSD) and symptoms of internalizing problems. Participants were administered the Mini International Neuropsychiatric Interview²² by a research assistant to determine whether participants met the diagnostic criteria PTSD. The Brief Symptom Inventory was administered to assess the participants’ internalizing symptoms in the past 6 months (i.e., depressive & anxiety symptoms).²³ Lastly, the frequency of cigarette use, marijuana use, and alcohol use in the past 6 months was assessed at baseline. The frequency of marijuana and cigarette use were measured on Likert scale of 0 (never) to 6 (every day/almost daily), while 0 (never) to 4 (4 or more times a week) was used for alcohol use.

Procedures

FYI was conducted in the region’s only Level-1 trauma center in Flint, Michigan.¹⁷ Participants included youth (ages 14–24) seeking ED treatment for an assault injury and reporting past 6-month drug use (AIG; N = 350). A proportionally sampled comparison group of youth presenting for other reasons who also reported past 6-month drug use were recruited and enrolled (CG; N = 250). Participants were recruited from 12/2009 to 9/2011, and trained RAs recruited participants everyday (excluding holidays). On Tuesday and Wednesdays, participants were recruited 21 hours a day (except from 5am-2am), whereas participants were recruited 24 hours a day on the other days. After obtaining the participants’ written consent (or assent from youth with parental consent if youth is younger than age 18), participants completed a screening survey. Participants were considered assault-injured if their injury was intentionally caused by another person. Drug use in the past 6 months was measured using the National Institute on Drug Abuse Alcohol, Smoking, and Substance Involvement Screening Test (i.e., NIDA-ASSIST). Exclusion criteria included ED presentation for sexual assault, suicidal ideation/attempt, child maltreatment, or a cognitive condition precluding consent (e.g., acute psychosis, alcohol intoxication). In addition, youth who arrived at the ED in active police custody or those not speaking English were excluded. Youth were included in the AIG cohort if they screened positive for an assault injury and past 6-month drug use were enrolled in the AIG. Youth in the CG cohort was recruited in parallel with the AIG cohort (to limit seasonal and temporal variation). Both cohorts were balanced across age groups (i.e., 14–17, 18–20, 21–24) and sex (male/female). Youth enrolled in the AIG and CG cohort completed a self-administered baseline survey in conjunction with an RA-administered structured interview. Follow-up assessments were conducted in-person at 6-, 12-, 18-, and 24-months post baseline. Participants received $1 for completing the screening survey, $20 for the baseline, and $35, $40, $40, and $50 for the 6-, 12-, 18-, and 24-month follow-ups. The FYI Study was approved by the University of Michigan (HUM00026787) and Hurley Medical Center IRBs (IRBNet ID: 177665) and an NIH certificate of confidentiality was obtained.

Analytic approach

As a precursor to evaluating our conceptual model (see Figure 1), we first fit unidimensional factor analytic models for neighborhood disadvantage and exposure to violence.²⁴ Unidimensional models were specified given the limited number of indicators per latent construct (e.g., 4 indicators for exposure to violence, 5 indicators for neighborhood disadvantage). Models were determined to it the data well if the: (1) chi-square test of absolute fit is not statistically significant, (2) root mean square error of approximation (RMSEA) ≤ .06, and (3) confirmatory factor index (CFI) ≥ .90.²⁵ Of note, we adjusted standard errors to account for non-independence across observations (i.e., participants nested in census tracts).²⁶ After fitting measurement models, we estimated a structural equation model (SEM) with serial mediation analysis to evaluate direct and indirect effects of residential racial segregation on youth firearm violence via neighborhood disadvantage and then exposure to violence (see Figure 1). Indirect effects were estimated using the products of coefficients method.²⁷ Bias-corrected bootstrapping was implemented to estimate the sampling distribution of the model coefficients. Thus, we report 95% bias-corrected confidence intervals (95% BCI) to determine the statistical significance of direct and indirect effects. Lagrange Multiplier (LM) tests were used to identify sources of model misfit and implement theory-driven model modifications to improve fit.²⁵ Given the binary nature of our dependent variable (i.e., youth firearm violence), weighted least square mean and variance adjustment estimator was used. As in the measurement models, cluster adjusted standard errors were used.²⁶ Statistical tests were conducted using Mplus (v 8.9).²⁸

Results

Descriptive statistics are presented in Table 1. The measurement model for neighborhood disadvantage fit the data well (i.e., χ2 (4) = 5.52, p = 0.24; RMSEA = .03, CFI = .99). Standardized factor loadings ranged from 0.51 (percent unemployment) to 1.01 (percent poverty). The measurement model for exposure to violence also fit the data well (i.e., χ2 (2) = 2.59, p = 0.27; RMSEA = .03, CFI = .99), with standardized factor loadings ranging from .30 (community violence) to .89 (non-partner aggression). Thus, the measurement models for neighborhood disadvantage and exposure to violence fit the data well (see Table 2).

Table 2.

Measurement models for mediators

Neighborhood Disadvantage (2010)	Λ	s.e.	p
% receives food stamps/SNAP	0.89	0.04	< .001
% unemployed (age ≥ 16)	0.51	0.08	< .001
% Did not graduate HS (age ≥ 25)	0.44	0.14	< .001
% Overcrowding at home	0.63	0.06	< .001
% Household income below poverty level	1.01	0.04	< .001
Exposure to Violence (Wave 2)	Λ	s.e.	p
Community violence	0.30	0.05	< .001
Non-partner aggression	0.89	0.06	< .001
Partner aggression	0.67	0.08	< .001
Violent victimization with a weapon	0.43	0.08	< .001

Note. Λ = factor loading, s.e. = standard error, and p = p-value. Model was estimated using weighted least square robust mean and variance adjustment estimator.

Next, we estimated structural models to evaluate whether residential racial segregation is directly and indirectly associated with youth firearm violence, with neighborhood disadvantage and exposure to violence as mediators (see Table 3). The structural model did not fit the data well (χ2 (227) = 227.56, p = 0.01; RMSEA = .04, CFI = .82). Per the model modification index, estimating residual correlations between neighborhood disadvantage (from 2000) and public assistance, partner and non-partner aggression, and violence victimization and non-partner aggression improved model fit (χ2 (233) = 259.15, p = 0.12; RMSEA = .02, CFI = .90). Post-hoc modifications are consistent with research, which has indicated that residents of disadvantaged neighborhoods are more likely to participate in public assistance programs (e.g., food stamps),²⁹ and that exposure to violence is not a singular event as most individuals exposed to violence experience several exposures of different kinds.^1,30 While the association between residential racial segregation and youth firearm violence was not significant (b = −0.04, 95% BCI [−0.22, 0.12]), residential racial segregation was linked with neighborhood disadvantage (b = −0.26; 95% BCI [−0.53, −0.04]), suggesting that a greater concentration of Black households (relative to White households) is associated with higher levels of neighborhood disadvantage. In addition, neighborhood disadvantage was positively associated with exposure to violence (b = 0.21; 95% BCI [0.01, 0.34]), and exposure to violence was positively associated with youth firearm violence (b = 0.46; 95% BCI [0.16, 0.80]). Lastly, residential racial segregation on youth firearm violence was indirectly associated through neighborhood disadvantage and then exposure to violence (b = −0.03; 95% BCI [−0.09, −0.01]). Other indirect pathways (i.e., exposure to violence or neighborhood disadvantage as independent mediators) were not statistically significant. Lastly, neighborhood disadvantage in 2000 was positively associated with neighborhood disadvantage in 2010 (b = 1.08; 95% BCI [0.93, 1.29]), and alcohol use at baseline was associated with exposure to violence (b = 0.26; 95% BCI [0.04, 0.45]). Other control variables were not significant in the structural model.

Table 3.

Structural equation model with serial mediation analysis

Youth firarm violence (waves 3 to 5)	β	95% bias-corrected bootstrapped confidence interval
Residential Segregation (2000)	−0.04	−0.22, 0.12
Exposure to Violence (wave 2)	0.46	0.16, 0.80
Neighborhood Disadvantage (2010)	0.003	−0.26, 0.20
Baseline youth firearm violence (wave 1)	0.11	−0.04, 0.28
Male	0.12	−0.05, 0.31
Public Assistance	0.01	−0.17, 0.22
Age	−0.05	−0.26, 0.22
Violent Injury	−0.03	−0.27, 0.14
Alcohol Use	−0.13	−0.41, 0.07
Marijuana Use	−0.03	−0.24, 0.17
Tobacco Use	0.14	−0.03, 0.29
Internalizing Symptoms	−0.09	−0.39, 0.13
PTSD	0.05	−0.18, 0.22
Exposure to Violence (wave 1)	0.08	−0.15, 0.28
Exposure to Violence (wave 2)	β	95% bias-corrected bootstrapped confidence interval
Residential Segregation (2000)	−0.04	−0.22, 0.12
Neighborhood Disadvantage (2010)	0.21	0.01, 0.34
Baseline youth firearm violence (wave 1)	0.07	−0.15, 0.30
Exposure to Violence (wave 1)	0.20	0.02, 0.37
Male	−0.03	−0.17, 0.11
Public Assistance	0.02	−0.14, 0.24
Age	−0.15	−0.32, 0.03
Violent Injury	−0.01	−0.14, 0.14
Alcohol Use	0.26	0.04, 0.45
Marijuana Use	0.14	−0.03, 0.33
Tobacco Use	0.11	−0.04, 0.27
Internalizing Symptoms	0.06	−0.13, 0.27
PTSD	0.14	−0.09, 0.39
Neighborhood Disadvantage (2010)	β	95% bias-corrected bootstrapped confidence interval
Residential Segregation (2000)	−0.26	−0.53, 0.04
Neighborhood Disadvantage (2000)	1.08	0.93, 1.29
Indirect Effects	β	95% bias-corrected bootstrapped confidence interval
Residential Segregation (2000) → Neighborhood Disadvantage (2010) → Youth firearm violence (waves 2–4)	−0.001	−0.06, 0.06
Residential Segregation (2000) → Exposure to Violence (wave 2) → Youth firearm violence (waves 3–5)	−0.02	−0.15, 0.06
Residential Segregation (2000) → Neighborhood Disadvantage (2010) → Exposure to Violence (wave 2) → Youth firearm violence (waves 3–5)	−0.03	−0.09, −0.01

Note. β = standardized coefficient.

Discussion

Residential racial segregation – as operationalized by ICE_race – was indirectly associated with future youth firearm violence by increasing neighborhood disadvantage and exposure to violence. As a result of the enduring effects of historical redlining and other segregation practices, neighborhoods with a higher concentration of Black households relative to White households are often associated with neighborhood disadvantage.⁹ Residential racial segregation leads to neighborhood disadvantage by restricting access to socioeconomic resources and opportunities, thereby perpetuating concentrated poverty (e.g., higher unemployment and poverty rate).^10,13 Our findings also reveal that higher neighborhood disadvantage is associated with higher levels of exposure to violence.^31–33 In line with social disorganization theory,^34–36 neighborhoods characterized by concentrated poverty often face a scarcity of resources for both formal and informal control mechanisms (e.g., caring adult neighbors). This scarcity, in turn, may cultivate a perception among residents that there are few repercussions for engaging in violence which may cultivate norms and values that normalize violence. Lastly, our that exposure to violence was associated with youth firearm violence is consistent with past research.^37,38 Retaliatory attitudes,^37,39 diminished future expectations,³⁸ and the desensitization to violence operate as potential pathways.⁴⁰

Of note, contrary to other studies, residential racial segregation was not directly associated with youth firearm violence.^6,11 To date, most researchers have evaluated the association between residential segregation and firearm violence for adults and youth using place-based, administrative data (e.g., arrest records).⁶ One might surmise that analyzing data at different levels of aggregation (e.g., administrative- and respondent-level data) can yield distinct results. For instance, place-based measures of firearm violence may be primarily influenced by other place-based, contextual factors such as residential racial segregation or neighborhood disadvantage.^5,9,11 Alternatively, the influence of place-based factors such as residential racial segregation on youth self-report of their firearm violence engagement may be mediated by respondent-level individual characteristics (e.g., exposure to violence). Our study provides a first look at how residential segregation may operate to influence a self-report measure of youth firearm violence.

Lastly, neighborhood disadvantage nor exposure to violence independently mediated the longitudinal association between residential racial segregation and youth firearm violence. The ecological consequences of residential racial segregation such as neighborhood disadvantage may increase the likelihood of youth firearm violence through its effects on prior exposure to violence. Our study highlights the importance of examining the interplay between ecological (e.g., disadvantage) and social (e.g., exposure to violence) factors to comprehensively understand how residential racial segregation plays a role in youth firearm violence. Identifying these mechanisms can inform youth firearm violence prevention programs that address the multi-level consequences of residential racial segregation to foster safer communities. For instance, interventions that address poverty (e.g., community economic development initiatives; affordable housing programs)⁴¹ and place-based violence interventions (e.g., vacant land stabilization), in tandem, may be more efficacious for disrupting the underlying mechanisms from institutional racism to youth firearm violence.⁴²

While our study contributes to our understanding of how residential racial segregation shapes firearm violence among youth, several limitations require attention. First, our study focused on adolescents living in one city with high crime rates. Additionally, our participants reported illicit drug use within the past year and sought medical care at an urban emergency department. It is, therefore, crucial to acknowledge that our findings may not generalize to non-clinical or non-substance-using youth residing in different geographic locations, such as rural areas. Nevertheless, our sample offers insights into the influence of residential racial segregation on youth firearm violence, specifically through the pathways of neighborhood disadvantage and exposure to violence. Our results, however, may be particularly relevant for youth in contexts characterized by high rates of violence which is often observed in economically challenged small cities across the U.S. Second, it is also important to consider that our study was based on data collected over ten years ago so may be time specific. Yet, researchers using more recent data have observed that residential racial segregation, neighborhood disadvantage, and exposure to violence remain significant risk factors for firearm violence.^5,9 Our study is important because we begin to assess the underlying mechanisms of how residential racial segregation may be associated with youth firearm violence. Lastly, while we assess longitudinal pathways from residential racial segregation, future research that employs quasi-experimental designs (e.g., propensity score matching) to support the ability to draw causal inferences would be useful. Nonetheless, our study was the first we know about that has examined the relationships between historical policies of racism and firearm violence using a longitudinal design.

Despite these limitations, our findings suggest that residential racial segregation is associated with downstream consequences (i.e., neighborhood disadvantage, exposure to violence) which, in turn, increase the likelihood of youth firearm violence. Our results suggest several approaches to prevention youth firearm violence prevention strategies. First, our results underscore the importance of implementing place-based policies aimed at mitigating residential racial segregation practices.⁴³ Examples of such policies include inclusionary zoning and the low-income housing tax credit, which have the potential to foster safer communities by addressing firearm violence.⁴³ Second, implementing place-based policies and programs that prioritize increasing socioeconomic opportunities within low-income communities of color may disrupt the link between residential racial segregation and youth firearm violence.¹⁵ Lastly, programs that reduce youth exposure to violence, such as those that facilitate positive adult and peer interactions (e.g., organized activities),³⁷ present valuable opportunities for preventing youth firearm violence within the context of residential segregation.