You are here

Article Text

Article menu
Original research
Race, structural racism and racial disparities in firearm homicide victimisation
  1. Kelsey M Conrick1,2,
  2. Avanti Adhia2,3,
  3. Alice Ellyson2,4,
  4. Miriam Joan Haviland2,
  5. Vivian H Lyons5,
  6. Brianna Mills2,6,
  7. Ali Rowhani-Rahbar6
  1. 1 School of Social Work, University of Washington, Seattle, Washington, USA
  2. 2 Firearm Injury & Policy Research Program, University of Washington, Seattle, Washington, USA
  3. 3 Department of Child, Family, and Population Health Nursing, School of Nursing, University of Washington, Seattle, Washington, USA
  4. 4 Center for Child Health, Behavior and Development, Seattle Children's Research Institute, Seattle, Washington, USA
  5. 5 Social Development Research Group, School of Social Work, Department of Psychiatry, University of Washington Allies in Healthier Systems for Health & Abundance in Youth, Seattle, Washington, USA
  6. 6 Department of Epidemiology, School of Public Health, University of Washington, Seattle, Washington, USA
  1. Correspondence to Kelsey M Conrick, School of Social Work, University of Washington, Seattle, WA 98105, USA; kmc621{at}uw.edu

Abstract

Objectives To identify an approach in measuring the association between structural racism and racial disparities in firearm homicide victimisation focusing on racism, rather than race.

Methods We examined associations of six measures of structural racism (Black/white disparity ratios in poverty, education, labour force participation, rental housing, single-parent households and index crime arrests) with state-level Black-white disparities in US age-adjusted firearm homicide victimisation rates 2010–2019. We regressed firearm homicide victimisation disparities on four specifications of independent variables: (1) absolute measure only; (2) absolute measure and per cent Black; (3) absolute measure and Black-white disparity ratio and (4) absolute measure, per cent Black and disparity ratio.

Results For all six measures of structural racism the optimal specification included the absolute measure and Black-white disparity ratio and did not include per cent Black. Coefficients for the Black-white disparity were statistically significant, while per cent Black was not.

Conclusions In the presence of structural racism measures, the inclusion of per cent Black did not contribute to the explanation of firearm homicide disparities in this study. Findings provide empiric evidence for the preferred use of structural racism measures instead of race.

  • firearm
  • socioeconomic status
  • income inequaility
  • health disparities

Data availability statement

Data are available in a public, open access repository. All data relevant to the study are included in the article. All data used for this study are available in a public, open access repository. The compiled data used for this study are available in online supplemental files 3 and 4; the corresponding data dictionary is available in online supplemental file 5.

https://doi.org/10.1136/ip-2022-044788

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    WHAT IS ALREADY KNOWN ON THIS TOPIC

    • Research has established a relationship between structural racism and disparities in firearm homicide victimisation. However, these studies often adjust for the per cent of the population who are Black, which in some cases may not be an accurate proxy for structural racism.

    WHAT THIS STUDY ADDS

    • This study contributes empirical evidence for carefully considering whether measures of race, racial composition (eg, per cent Black) or racism are most appropriate in specific contexts.

    HOW THIS STUDY MIGHT AFFECT RESEARCH

    • Our study suggests that the inclusion of per cent Black does not always contribute further to the explanation of firearm homicide disparities, and researchers should consider using measures of structural racism when available.

    Introduction

    The USA experiences 19 000 firearm homicides annually.1 Black individuals are 8.1 times more likely to die from firearm homicides than white individuals.1 Despite this pervasive disparity, research has not always centred and named racism as a causal mechanism for racial inequities,2–4 due in part to biases in the academic publication process that undervalue the decades of work on racism completed by scholars of colour.2 As such, recent research has contributed to existing scholarship on racism and disparities in health outcomes3 by empirically establishing racism, particularly structural racism, as a determinant of both firearm homicide5 6 and disparities in firearm homicide.7–11

    Structural racism refers to ‘the historically contingent and persistent ways in which social systems and institutions generate and reinforce inequities in access to power, privilege and other resources among racial and ethnic groups deemed to be superior and those viewed as inferior.’12 Structural racism may influence disparities in firearm homicide through several pathways. Racist historical policies and practices, such as slavery, settler colonialism, Jim Crow laws, immigrant exclusion and genocide continue to be embedded in contemporary policies and practices governing social systems and institutions, including welfare, education, voting, criminal, healthcare and immigration.13–17 The integration of structural racism in these systems, which are responsible for allocating political, social and economic resources, perpetuates inequities.13–16 For example, racist housing policies such as redlining have created residential segregation, where people of colour are concentrated into under-resourced communities.18 Evidence suggests that residential segregation leads to crime19 by inhibiting employment,13 reducing school resources,13 diminishing public investments18 and perpetuating labelling of neighbourhoods as ‘hazardous and deleterious.’20 Similarly, research has established a link between levels of residential segregation and disparities in firearm homicide at the state and city level.6 9–11 21–23

    Studies measuring the effect of an exposure (eg, specific state policy) on disparities in state-level firearm homicide often adjust for measures of structural disadvantage, such as poverty, education, unemployment, labour force participation and index crime arrest rates.7–9 Similarly, the state’s per cent of the population who are Black (‘per cent Black’) is often adjusted for in this research.5 24 However, the per cent Black may not be the optimal measure, depending on the causal pathway of interest, as measurements of race may not be an accurate proxy for racism.2 25 26

    We sought to identify an improved approach to examining the association between measures of structural racism and racial disparities in firearm homicide that focused on racism, rather than race, as a driving mechanism. Because structural racism is an ‘interactive, independent and compounding’18 process, we did not attempt to identify a single measure of structural racism, but rather a strategy for researchers to include measures of structural racism tailored to the unique context of their hypothesised causal pathway between their exposure of interest (eg, a specific policy) and disparities in firearm homicide. We conceptualised measures of structural racism as racial disparities in structural disadvantage. We hypothesised that adding a disparity ratio (eg, disparities in poverty) to model the association of a measure of structural disadvantage (eg, poverty) with racial disparities in firearm homicide would improve model fit compared with a model including only an absolute measure (eg, absolute poverty levels). Second, we hypothesised that further including per cent Black to model this association would not improve the model fit.

    Methods

    Design overview

    We examined the association of six measures for structural disadvantage and racism commonly used in literature on firearm homicide disparities. These measures included poverty, educational attainment, labour force participation, per cent of individuals living in rental housing, per cent of single-parent households and arrests for index crimes. We used data from 2010 to 2019 at the state level and included only disparities between Black and white individuals in this analysis due to relatively low population sizes of Native American, Asian and Native Hawaiian or Pacific Islander individuals. Additionally, because not all data sources distinguished Latin(x) individuals from other racial groups and because data were collected inconsistently for some variables, we did not analyse data separately for this population. We calculated state-level disparity ratios for firearm homicide and each measure of structural racism as the rate for Black individuals divided by the rate for white individuals. Black/white disparity ratios of firearm homicide have been used in other studies7 9 27; we extended this approach to measure structural racism as well. Details on data sources used are available in online supplemental file 1. Patients or public were not involved in the design, conduct, reporting,or dissemination plans of our research

    Supplemental material

    Data sources and measures

    Outcome

    We used the Centers for Disease Control and Prevention (CDC) Web-based Injury Statistics Query Reporting to obtain each state’s age-adjusted rates of firearm homicide for white and Black populations, both non-Hispanic, averaged across 2010–2019.1 The CDC does not report categories for which there are fewer than 10 deaths and recommends against using data with fewer than 20 deaths annually; therefore, the following 10 states were removed from analyses: Hawaii, Idaho, Maine, Montana, New Hampshire, North Dakota, South Dakota, Utah, Vermont and Wyoming. We calculated the disparity ratio for the outcome as the rate of age-adjusted firearm homicides for Black individuals divided by the rate for white individuals in each state. A ratio of greater than 1 indicates a state has a higher proportion of firearm homicide among Black individuals compared with white individuals. Because the Black-white disparity in firearm homicide was skewed, we used the natural log of the racial disparity in firearm homicide as the outcome.9

    Measures of structural disadvantage and structural racism

    The following six measures of structural disadvantage were selected by searching covariates used in the literature on disparities in firearm homicide, and we calculated measures of structural racism to be Black divided by white rates for each measure of structural disadvantage. Data for per cent of population below the federal poverty line (FPL), per cent of population aged 16 or older participating in the labour force, and per cent of individuals living in rental housing came from the American Community Survey (2010–2019).28 Educational attainment was defined as the 4-year adjusted-cohort graduation rate (ACGR) for Black and white students sourced from the National Center for Education Statistics .29 The ACGR is the per cent of public high school freshmen who graduate with a regular diploma within 4 years of beginning ninth grade. ACGR data by race were only available for the 2012–2013 to the 2018–2019 school year. The per cent of single-parent households was obtained from DiversityDataKids, which differentiates individuals who are non-Hispanic white.30

    Arrests for index crimes were conceptualised to be a measure of structural racism in the criminal legal system. These data were obtained using the ‘fbicrime’ R package,31 which uses the Application Programming Interface of the Federal Bureau of Investigation’s Uniform Crime Reporting (UCR) Programme.32 We chose to use data for arrests rather than for convictions because of the well-documented anti-Black racist practices in arrests of Black individuals compared with their proportion of crime offending.16 Reporting to the UCR programme is voluntary. Florida, Illinois and New York do not report data to the UCR programme and so were excluded from regressions using index crimes. Six additional states contained at least one state-year where data were not reported to the UCR programme; these state-years were also excluded from analysis.

    Per cent Black

    The per cent of the population who were Black was calculated using data from the American Community Survey.

    Analysis

    Data were analysed using the 2010–2019 averages. We chose to use this pooled approach rather than conducting a longitudinal analysis for two reasons. First, there was little variation in each measure across time (online supplemental file 1) except in states with small populations of individuals who were Black. Second, conducting a longitudinal analysis over only one decade would imply a relatively short lag for the relationship between each measure of structural racism and disparities in firearm homicide. However, contemporary structural racism arises from a centuries-long history of ongoing oppression and discrimination. Therefore, it is challenging to ascertain the lag period between an individual state’s structural racism metric (eg, disparities in poverty) and resulting racial disparities in firearm homicide.

    We used linear regressions with ordinary least squares (OLS) to examine the association between state-level measures of structural racism and Black-white racial disparities in log age-adjusted firearm homicide rates. To compare across specifications and identify the optimal specification, five measures of model fit were evaluated: log likelihood (with likelihood ratio test), Akaike information criterion (AIC), Bayesian information criterion (BIC), R2 and adjusted R2. We estimated four separate regressions for each measure of structural racism to compare model fit when adjusting for per cent Black compared with the Black-white disparity ratio of structural racism: (model 1) absolute measure only (eg, poverty); (model 2) absolute measure and per cent Black; (model 3) absolute measure and Black-white disparity ratio and (model 4) absolute measure, Black-white disparity ratio and per cent Black.

    Sensitivity analyses

    In addition to the four models presented in the main body of this paper, we also tested two other configurations: (1) models that included the disparity metric as a difference of Black-white rates of each independent variable and (2) models that included only the disparity ratio and per cent Black but not the absolute measure of each independent variable. The preferred model with each of these configurations was broadly the same as the results presented below. Results with point estimates and fit statistics are available in online supplemental file 2; data files are available in online supplemental file 3.

    Supplemental material

    Supplemental material

    Results

    During the period 2010–2019, the average age-adjusted firearm homicide rate was 4.35 per 100 000 person-years, ranging from 1.05 to 10.73 per 100 000 person-years (figure 1). The average Black-white firearm homicide disparity ratio was 8.35 (range 4.62–34.66), indicating on average that Black individuals were 8.35 times more likely to be a victim of firearm homicide than white individuals. The disparity ratio for poverty, education, rental housing and single parent households indicated greater socioeconomic disadvantage for the Black compared with the white population in all states.

    Figure 1
    Figure 1

    Maps of independent variables (measures of structural disadvantage = ‘absolute measure’ and measures of structural racism = ‘disparity ratio’) and dependent variable (disparities in firearm homicide victimisation). aFirearm homicide is the age-adjusted firearm homicide victimisation rate per 100 000 persons. bPoverty is defined as the per cent of individuals below the Federal Poverty Line. cEducation is defined as the 4-year adjusted-cohort graduation rate, which is the percentage of public high school freshmen who graduate with a regular diploma within 4 years of starting ninth grade. dLabour force participation is defined as the per cent of the population aged 16 and over participating in the labour force. eRental housing is defined as the percentage of occupied housing units that are rented rather than owned. fSingle-parent households is defined as the percentage of households with a single parent (male householder and no wife present or female householder and no husband present). gArrests for index crimes is the number of arrests per 10 000 persons.

    The regression coefficients from OLS estimating the association of the six independent variables (absolute measure and disparity ratio for each variable) with Black-white disparities in the log of age-adjusted firearm homicide rates are in table 1. Across the six independent variables, absolute measures of structural disadvantage were inconsistently associated with disparities in firearm homicide. Per cent Black (in models 2 and 4) was not statistically significantly associated with firearm homicide racial disparities, and the point estimates were near 0. In contrast, the disparity ratio of each measure of structural racism (in models 3 and 4) was always significantly associated with the outcome. For example, in model 4, every one-unit increase in the Black-white disparity in the per cent of the population living below FPL was significantly associated with a 60.00% (95% CI: 16.18% to 118.15%) increase in the Black-white disparity ratio in firearm homicide. However, in this model, the per cent Black was not significantly associated with Black/white disparity in firearm homicide (1.00%; 95% CI −1.00% to 2.02%).

    View this table:
    Table 1

    Coefficients from linear models testing predictive value of independent variables on Black-white disparities in the log of age-adjusted firearm homicide rates

    Goodness-of-fit statistics for each model suggested that the model including the absolute rate and disparity ratio for each measure of structural racism, but not per cent Black (model 3), was optimal (table 2). For all models, AIC and BIC were lowest for model 3. Additionally, for all models either the log likelihood was highest for model 3, or a likelihood ratio test indicated no statistically significant difference between models 3 and 4 (which included per cent Black in addition to the absolute measure and disparity ratio of each measure of structural racism). The adjusted R2 statistic also indicated that model 3 was preferred for all measures of structural racism except for education.

    View this table:
    Table 2

    Model fit statistics for models of Black-white disparities in log age-adjusted firearm homicide rate

    Discussion

    This study advances prior literature by contributing empirical evidence for the use of measures of structural racism, rather than race, in research on firearm homicide disparities. In this study, the addition of the per cent Black did not contribute further to the explanation of disparities in firearm homicide beyond measures of structural racism. Scholars have called for this shift in focus from using race as a risk factor for disparities in health to variables that capture the complexity of structural racism.2 4 17 26 33 Additionally, while empirical research connecting structural racism to firearm homicide and firearm homicide disparities is relatively nascent, researchers have investigated the influence of racism on racial health disparities for decades.34 This paper builds on these calls by specifically considering the example of disparities in firearm homicide.

    Research on firearm homicide has recently begun to emphasise the importance of considering racial disparities in firearm homicide as an outcome7 9 23 27 rather than only focusing on absolute rates. This shift in focus is critical due to the differential influence of policies and other exposures of interest on persons of colour. For example, a recent reanalysis of firearm restriction policies for respondents of domestic violence restraining orders found reductions in intimate partner homicide for white but not Black victims.35 Attention to only absolute rates of firearm homicide can obscure the true effect—or lack thereof—of policies and other exposures on persons of colour. In our study, Black individuals had a higher rate of firearm homicide than white individuals in every state; in the most disparate state, New Jersey, the rate for Black individuals was 35 times higher than for white individuals. As additional research is conducted to understand the aetiology of these disparities and investigate solutions, scholars should shift focus from the role of race, which is not intervenable, to the role of structural racism, which may be measured as racial disparities in structural disadvantage. Some studies may require the use of racial composition as a measure of racism. For example, in their assessment of the relationship between historical racist housing loan discrimination and contemporary violence, Jacoby et al used measures of racial composition at the census tract level as this metric contributed to the assignment of loans.6 Researchers should include the components of structural racism tailored to the unique context of their research study and should justify these choices when presenting their findings.

    Consistent with prior literature,7 9 27 while absolute measures of structural racism were only inconsistently associated with the outcome, the disparity ratios of these measures were consistently associated with disparities in firearm homicide. These findings are consistent with our conceptualisation of the pathway through which structural racism influences disparities in firearm homicide and suggest differences in state-level racial disparities in firearm homicide are due to state-level differences in structural racism. Specifically, states with the highest levels of structural racism (measured in this study by disparities in six measures of structural disadvantage) have the highest racial disparities in firearm homicide. Regression coefficients should be interpreted with caution, as we did not control for additional variables that may contribute to the association between measures of structural racism and firearm homicide disparities. Future research seeking to measure the effect of an exposure of interest (eg, a specific policy) on disparities in firearm homicide should use additional adjustment variables as appropriate.

    As the goal of this paper was to compare approaches to the inclusion of absolute measures and disparity ratios of structural racism, as well as per cent Black, we did not seek to identify a single index variable or latent construct that comprehensively measures structural racism or identifies the ‘strongest’ measure of structural racism. This approach has been used by other scholars; Unnever et al used principal factor analysis to create a single factor score for each state and found per cent Black was still statistically significantly associated with firearm homicide perpetration by Black and white offenders.8 These findings differ from our results This difference could be due to differing methodologies between studies or potential information loss that can occur when using composite measures. The use of composite measures additionally may create problems with collinearity and prohibit the identification of specific contributors to racial disparities, and therefore, reduce the ability to identify ameliorating interventions. For this study, we instead chose to develop a strategy that focused on the role of structural racism, rather than race, that still allows for flexibility in choosing structural racism metrics.

    This study has some limitations. We only calculated disparity ratios for the outcome and measures of structural racism for Black compared with white individuals, and not all data sets distinguished non-Latinx individuals from other racial groups. Additionally, the inconsistent reporting of ethnicity data from individual agencies reporting arrests for index crimes diminished the reliability of any analyses conducted from these data. Relatively low population sizes for Native American, Asian and Native Hawaiian or Pacific Islander individuals within most states rendered national analysis unfeasible. Understanding the burden of inequities these communities experience is also critical to developing solutions. Future studies should consider substate (eg, city level) analyses comparing geographical units with the highest level of racial subgroups of interest. This approach would also allow for replicative studies to assess whether our findings on the lack of contribution of per cent Black to our models holds at a smaller geographical unit. Finally, data on arrests for index crimes were unavailable for three states (Illinois, New York and Florida, USA), although these states’ data were available for the other five independent variables.

    This study has implications for both research and practice. The majority of research to date on racial disparities in firearm homicide has focused on the role of residential segregation and discriminatory housing policies.7 9 27 We found other measures of structural racism, including disparities in poverty, education, labour force participation and arrests for index crimes were also associated with disparities in firearm homicide. As researchers further characterise racist policies that create these inequities, our study contributes to the methodological approaches that can be used to measure their effect. Our findings suggest a close examination of whether the inclusion of per cent Black is appropriate in this research given the causal pathway of interest, as we found that measure to not contribute to the explanation of firearm homicide disparities beyond measures of structural racism (eg, Black-white disparity ratios of poverty, education) in this study. Once the specific policies that create these disparities are better characterised, researchers have a responsibility to participate in designing structural solutions to the long-standing consequences of structural racism, including disparities in firearm homicide.2 4 9 26

    Supplemental material

    Supplemental material

    Data availability statement

    Data are available in a public, open access repository. All data relevant to the study are included in the article. All data used for this study are available in a public, open access repository. The compiled data used for this study are available in online supplemental files 3 and 4; the corresponding data dictionary is available in online supplemental file 5.

    Ethics statements

    Patient consent for publication

    Ethics approval

    Institutional review board approval was not sought as all data used were deidentified and publicly available.

    References

    1. ↵
      WISQARS data visualization, 2020. Available: https://wisqars-viz.cdc.gov [Accessed 13 Sep 2021].
    2. ↵
      2. Boyd RW ,
      3. Lindo EG ,
      4. Weeks LD
      . On racism: a new standard for publishing on racial health inequities. Health Affairs Blog 2020 https://www.healthaffairs.org/do/10.1377/hblog20200630.939347/full/
    3. ↵
      2. Bailey ZD ,
      3. Krieger N ,
      4. Agénor M , et al
      . Structural racism and health inequities in the USA: evidence and interventions. Lancet 2017;389:1453–63.doi:10.1016/S0140-6736(17)30569-X pmid:http://www.ncbi.nlm.nih.gov/pubmed/28402827
      OpenUrlCrossRefPubMed
    4. ↵
      2. Hardeman RR ,
      3. Murphy KA ,
      4. Karbeah J'Mag , et al
      . Naming institutionalized racism in the public health literature: a systematic literature review. Public Health Rep 2018;133:240–9.doi:10.1177/0033354918760574 pmid:http://www.ncbi.nlm.nih.gov/pubmed/29614234
      OpenUrlPubMed
    5. ↵
      2. Houghton A ,
      3. Jackson-Weaver O ,
      4. Toraih E , et al
      . Firearm homicide mortality is influenced by structural racism in US metropolitan areas. J Trauma Acute Care Surg 2021;91:64–71.doi:10.1097/TA.0000000000003167 pmid:http://www.ncbi.nlm.nih.gov/pubmed/33797488
      OpenUrlPubMed
    6. ↵
      2. Jacoby SF ,
      3. Dong B ,
      4. Beard JH , et al
      . The enduring impact of historical and structural racism on urban violence in Philadelphia. Soc Sci Med 2018;199:87–95.doi:10.1016/j.socscimed.2017.05.038 pmid:http://www.ncbi.nlm.nih.gov/pubmed/28579093
      OpenUrlPubMed
    7. ↵
      2. Knopov A ,
      3. Rothman EF ,
      4. Cronin SW , et al
      . The role of racial residential segregation in Black-White disparities in firearm homicide at the state level in the United States, 1991-2015. J Natl Med Assoc 2019;111:62–75.doi:10.1016/j.jnma.2018.06.002 pmid:http://www.ncbi.nlm.nih.gov/pubmed/30129481
      OpenUrlPubMed
    8. ↵
      2. Unnever JD ,
      3. Stults BJ ,
      4. Messner SF
      . Structural racism and criminal violence: an analysis of State-Level variation in homicide. Race Justice 2021;140:215336872110152.doi:10.1177/21533687211015287
      OpenUrl
    9. ↵
      2. Wong B ,
      3. Bernstein S ,
      4. Jay J , et al
      . Differences in racial disparities in firearm homicide across cities: the role of racial residential segregation and gaps in structural disadvantage. J Natl Med Assoc 2020;112:518–30.doi:10.1016/j.jnma.2020.05.014 pmid:http://www.ncbi.nlm.nih.gov/pubmed/32641258
      OpenUrlPubMed
    10. ↵
      2. Benns M ,
      3. Ruther M ,
      4. Nash N , et al
      . The impact of historical racism on modern gun violence: Redlining in the city of Louisville, KY. Injury 2020;51:2192–8.doi:10.1016/j.injury.2020.06.042 pmid:http://www.ncbi.nlm.nih.gov/pubmed/32650980
      OpenUrlPubMed
    11. ↵
      2. Poulson M ,
      3. Neufeld MY ,
      4. Dechert T , et al
      . Historic redlining, structural racism, and firearm violence: a structural equation modeling approach. Lancet Reg Health Am 2021;3:100052.doi:10.1016/j.lana.2021.100052 pmid:http://www.ncbi.nlm.nih.gov/pubmed/34888540
      OpenUrlPubMed
    12. ↵
      2. Agénor M ,
      3. Perkins C ,
      4. Stamoulis C , et al
      . Developing a database of structural Racism-Related state laws for health equity research and practice in the United States. Public Health Rep 2021;136:428–40.doi:10.1177/0033354920984168 pmid:http://www.ncbi.nlm.nih.gov/pubmed/33617383
      OpenUrlCrossRefPubMed
    13. ↵
      2. Pager D ,
      3. Shepherd H
      . The sociology of discrimination: racial discrimination in employment, housing, credit, and consumer markets. Annu Rev Sociol 2008;34:181–209.doi:10.1146/annurev.soc.33.040406.131740 pmid:http://www.ncbi.nlm.nih.gov/pubmed/20689680
      OpenUrlCrossRefPubMedWeb of Science
    14. ↵
      2. Bonilla-Silva E
      . Rethinking racism: toward a structural interpretation. Am Sociol Rev 1997;62:465–80.doi:10.2307/2657316
      OpenUrlCrossRef
    15. ↵
      2. Bell DA
      . Race, Racism & American Law 6e. 6th edition. New York: Aspen Publishers, 2008.
    16. ↵
      2. Alexander M ,
      3. West C
      . The new Jim Crow: mass incarceration in the age of colorblindness. New York: The New Press, 2012.
    17. ↵
      2. Jones CP
      . Invited commentary: "race," racism, and the practice of epidemiology. Am J Epidemiol 2001;154:299–304.doi:10.1093/aje/154.4.299 pmid:http://www.ncbi.nlm.nih.gov/pubmed/11495851
      OpenUrlCrossRefPubMedWeb of Science
    18. ↵
      1. Brown University Center for the Study of Race and Ethnicity in America
      . How structural racism works project, 2020. Available: https://www.brown.edu/academics/race-ethnicity/programs-initiatives/signature-series/how-structural-racism-works-project [Accessed 27 Jul 2021].
    19. ↵
      2. Light MT ,
      3. Thomas JT
      . Segregation and violence reconsidered: do whites benefit from residential segregation? Am Sociol Rev 2019;84:690–725.doi:10.1177/0003122419858731
      OpenUrl
    20. ↵
      2. Woods LL ,
      3. Shaw-Ridley M ,
      4. Woods CA
      . Can health equity coexist with housing inequalities? A contemporary issue in historical context. Health Promot Pract 2014;15:476–82.doi:10.1177/1524839914533568 pmid:http://www.ncbi.nlm.nih.gov/pubmed/24812195
      OpenUrlCrossRefPubMed
    21. ↵
      2. Ulmer JT ,
      3. Harris CT ,
      4. Steffensmeier D
      . Racial and ethnic disparities in structural disadvantage and crime: white, black, and Hispanic comparisons. Soc Sci Q 2012;93:799–819.doi:10.1111/j.1540-6237.2012.00868.x pmid:http://www.ncbi.nlm.nih.gov/pubmed/25035523
      OpenUrlCrossRefPubMed
    22. ↵
      2. Light MT ,
      3. Ulmer JT
      . Explaining the gaps in white, black, and Hispanic violence since 1990: accounting for immigration, incarceration, and inequality. Am Sociol Rev 2016;81:290–315.doi:10.1177/0003122416635667
      OpenUrlCrossRef
    23. ↵
      2. Beard JH ,
      3. Morrison CN ,
      4. Jacoby SF , et al
      . Quantifying disparities in urban firearm violence by race and place in Philadelphia, Pennsylvania: a cartographic study. Am J Public Health 2017;107:371–3.doi:10.2105/AJPH.2016.303620 pmid:http://www.ncbi.nlm.nih.gov/pubmed/28103077
      OpenUrlPubMed
    24. ↵
      2. Kim D
      . Social determinants of health in relation to firearm-related homicides in the United States: a nationwide multilevel cross-sectional study. PLoS Med 2019;16:e1002978.doi:10.1371/journal.pmed.1002978 pmid:http://www.ncbi.nlm.nih.gov/pubmed/31846474
      OpenUrlCrossRefPubMed
    25. ↵
      2. Greenberg M ,
      3. Schneider D
      . Violence in American cities: young black males is the answer, but what was the question? Soc Sci Med 1994;39:179–87.doi:10.1016/0277-9536(94)90326-3 pmid:http://www.ncbi.nlm.nih.gov/pubmed/8066496
      OpenUrlCrossRefPubMed
    26. ↵
      2. Adkins-Jackson PB ,
      3. Chantarat T ,
      4. Bailey ZD
      . Measuring structural racism: a guide for epidemiologists and other health researchers. Am J Epidemiol 2021:kwab239.
    27. ↵
      2. Siegel M ,
      3. Sherman R ,
      4. Li C , et al
      . The relationship between racial residential segregation and Black-White disparities in fatal police shootings at the City level, 2013-2017. J Natl Med Assoc 2019;111:580–7.doi:10.1016/j.jnma.2019.06.003 pmid:http://www.ncbi.nlm.nih.gov/pubmed/31256868
      OpenUrlPubMed
    28. ↵
      Explore census data. Available: https://data.census.gov/cedsci/ [Accessed 14 Sep 2021].
    29. ↵
      Common core of data (CCD). Available: https://nces.ed.gov/ccd/tables/ACGR_RE_and_characteristics_2012-13.asp [Accessed 14 Sep 2021].
    30. ↵
      1. CKAN
      . Households that are single parent (percent) by race/ethnicity. Available: https://data.diversitydatakids.org/dataset/11001_2_p-households-that-are-single-parent-percent-by-race-ethnicity [Accessed 12 Jan 2022].
    31. ↵
      fbicrime: an R warpper for FBI crime API. Available: https://htmlpreview.github.io/?https://github.com/SUN-Wenjun/fbicrime/blob/master/doc/fbicrime.html [Accessed 14 Sep 2021].
    32. ↵
      1. CDE
      . Crime data API. Available: https://crime-data-explorer.fr.cloud.gov/pages/docApi [Accessed 14 Sep 2021].
    33. ↵
      2. Chae DH ,
      3. Nuru-Jeter AM ,
      4. Lincoln KD
      . Conceptualizing racial disparities in health: advancement of a Socio-Psychobiological approach. Du Bois Rev 2011;8:63–77.
      OpenUrl
    34. ↵
      2. Paradies Y ,
      3. Ben J ,
      4. Denson N , et al
      . Racism as a determinant of health: a systematic review and meta-analysis. PLoS One 2015;10:e0138511.doi:10.1371/journal.pone.0138511 pmid:http://www.ncbi.nlm.nih.gov/pubmed/26398658
      OpenUrlCrossRefPubMed
    35. ↵
      2. Wallin MA ,
      3. Holliday CN ,
      4. Zeoli AM
      . The association of federal and State-level firearm restriction policies with intimate partner homicide: a Re-analysis by race of the victim. J Interpers Violence 2022;37:NP16509–33.doi:10.1177/08862605211021988 pmid:http://www.ncbi.nlm.nih.gov/pubmed/34144667
      OpenUrlPubMed

    Supplementary materials

    Footnotes

    • Twitter @VivianHLyons

    • Contributors KMC and AR-R conceptualised and designed the study. KMC collected and analysed all data. All authors contributed to interpretation of findings. KMC led drafting of the manuscrpt, and all authors contributed to critical revisions. All authors approved the final version of the manuscript for publication and agree to be accountable for all aspects of the work. KMC and AR-R are the guarantors of the work, accept full responsibility for the work, had access to the data, and controlled the decision to publish.

    • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

    • Competing interests None declared.

    • Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

    • Provenance and peer review Not commissioned; externally peer reviewed.

    • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.