Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2024 Nov 1.
Published in final edited form as: Health Place. 2023 Sep 27;84:103114. doi: 10.1016/j.healthplace.2023.103114

Association between racial residential segregation and walkability in 745 U.S. cities

Ben R Spoer 1, Sarah E Conderino 1, Taylor M Lampe 1, Rebecca H Ofrane 1, Elaine De Leone 1, Lorna E Thorpe 1, Virginia W Chang 2, Brian D Elbel 1
PMCID: PMC10842472  NIHMSID: NIHMS1935367  PMID: 37774640

Abstract

Despite higher chronic disease prevalence, minoritized populations live in highly walkable neighborhoods in US cities more frequently than non-minoritized populations. We investigated whether city-level racial residential segregation (RRS) was associated with city-level walkability, stratified by population density, possibly explaining this counterintuitive association. RRS for Black-White and Latino-White segregation in large US cities was calculated using the Index of Dissimilarity (ID), and walkability was measured using WalkScore. Median walkability increased across increasing quartiles of population density, as expected. Higher ID was associated with higher walkability; associations varied in strength across strata of population density. RRS undergirds the observed association between walkability and minoritized populations, especially in higher population density cities.

Keywords: Walkability, Built Environment, Segregation, Cities, Population Density, Health Disparities

Introduction

A growing body of evidence has established that the built environment is an important determinant of health1,2. Studies have identified associations between higher neighborhood walkability (measuring features of the built environment that facilitate walking3,4), and lower prevalence of chronic disease5–8. Two recent studies reported a higher density of racial/ethnic minoritized individuals in more walkable areas of cities4,9, despite typically higher prevalence of chronic diseases among these same minoritized populations that can in part be prevented by physical activity, including walking10–12. Important questions arise from these findings, including whether this observed association is evident in racially segregated cities only, in which racial/ethnic minoritized populations are differentially distributed13. We theorize that in segregated cities, high minoritized population neighborhoods have higher population density and higher density of amenities, increasing walkability in those neighborhoods and potentially in the city overall. Exploring this hypothesis will shed light on whether racial residential segregation (RRS) – an important contributor to both health and built environment disparities13,14– is related to this counterintuitive association between higher walkability and higher percent minoritized residents.

The present analysis examines whether city-level (RRS) is associated with city-level walkability. It also examines this potential association across strata of city-level population density. If racial residential segregation were associated with walkability only in high population density cities, it could be the case that population density- which is a core component of most walkability metrics and tends to be high in minority neighborhoods in segregated cities 15,16 – actually drives the hypothesized association.

We examine RRS among two US minoritized groups: non-Hispanic Blacks (Black) and Latinos. We examine these groups because they comprise a substantial minority of the total US population and experience RRS17–19. We hypothesize that city-level walkability will increase as city-level RRS increases, driven by highly walkable, high density minority neighborhoods within cities.

Methods

The sample of cities was selected in 2017 for the creation of the City Health Dashboard (‘the Dashboard’, www.cityhealthdashboard.com)20. The sample consisted of US Census incorporated places with population >50,000 in the 2017 American Community Survey (ACS), plus the largest city in Vermont (Burlington, 2017 population 42,445) and West Virginia (Charleston, 2017 population 48,017), to ensure all states were represented (n=756 cities). We removed cities with fewer than 10 census tracts to eliminate unstable estimates. Cities missing analytic variables were also excluded (final n=745).

City-level walkability was measured using 2019 WalkScore data. WalkScore is a validated walkability metric calculated based on density of amenities nearby a location of interest3,21, on a scale of 0 to 100, 100 being most walkable. City-level WalkScore was measured as the population-density weighted average of WalkScore values for 500 square-foot city blocks in a given city22. WalkScore values were purchased from Redfin, the company that calculates WalkScore, and were not altered during analysis. WalkScore is well fit to test our hypothesis because it factors into its calculation the two primary drivers through which we theorize RRS and walkability are linked- population density and walkable neighborhood amenities. Other walkability measures typical include population density, but not distance to amenities23, in their calculations.

City-level RRS was measured using the Index of Dissimilarity (ID)24. ID measures the percentage of a group’s population that would have to relocate to a different neighborhood within a city in order to achieve an even distribution of that group’s members across the city’s neighborhoods24. A higher percentage indicates greater segregation. ID calculation methods are described elsewhere24. There are numerous potential ways to measure RRS, including measures of evenness, exposure, concentration, centralization, and clustering25. Evenness measures, like ID, produce high values in cities where there is substantially uneven distribution of two populations. Our hypothesis can be interpreted as hypothesizing that cities with highly uneven racial population distributions will have higher walkability. Given this, ID is the RRS measure best fit to test our hypothesis.

Conversely, measures of exposure focus on potential for interactions between members of different groups; interpersonal interactions are most relevant for health drivers related to social interactions, like finding job and educational opportunities, social cohesion, and others. They are not as well suited for the study of environmental drivers like walkability. Measures of centralization capture how close minority populations live to a hypothetical central business district, which may not exist in many cities. Measures of clustering and concentration factor in the geographic size of smaller areal units, in this case census tracts, and therefore are not comparable across the study sample; there is substantial variation in population density across the study sample, and census tract size is roughly inversely proportionate to population density26.

We calculated city-level ID for Black-White segregation and Latino-White segregation using census tracts to proxy neighborhoods (data source: 2019 ACS population estimates). Black is defined as ‘Black or African American alone, Not Hispanic or Latino’, White is defined as ‘White alone, Not Hispanic or Latino’, and Latino is defined as ‘Hispanic or Latino’.

City-level linear regression models assessed associations between ID and walkability. Models controlled for percentage of adults with at least a high school diploma, percentage of families living below the poverty level, and US Census region (South, Northeast, Midwest, and West), defined using 2019 ACS data. We controlled for these variables because each has been related to both RRS and walkability in previous literature13,27. Analyses were stratified by quartile of city population density and by quartile of city population percentage Black or Latino to examine whether observed associations could be explained by either population density (calculated as mean population count per square mile excluding water area, using 2019 ACS data) or percentage minoritized population. We also examined tract-level correlations between percentage Black/Latino and walkability to explore whether potential associations between RRS and walkability were driven by higher walkability in high minoritized population neighborhoods. Model fit was assessed by examining plots of residuals. All analyses were conducted using SAS version 9.4.

Results

Median walkability was 38.1 (IQR: 31.1–49.4), median Black-White ID was 39.5 (IQR: 32.6–47.4), and median Latino-White ID was 32.0 (IQR: 25.2–39.4). As expected based on the measure, median walkability increased with increasing city population density, from 29.3 (IQR: 23.3–34.7) for low density cities to 62.6 (IQR: 53.1–70.7) for high density cities. Census tract-level percent Black and Latino were significantly, albeit weakly, correlated with walkability (r=0.10 and 0.18, respectively; p<.001 for both).

When controlling for region and socioeconomic factors, a one-unit increase in Black-White ID was associated with a 0.35-unit increase in walkability, and a one-unit increase in Latino-White ID was associated with a 0.22-unit increase in walkability (Table 1). The strength of association between ID and walkability varied when stratified by population density. For Black-White ID, the strongest association was observed in high density cities (β=0.34, p<0.001). The association between Latino-White ID was strongest in mid density cities (β=0.32, p<0.001). Broadly, the association between RRS and walkability increased with population density.

Table 1.

City-level associations between segregation and walkability stratified by population density and percent Black or Latino

Model* N B† SE† p-value
Black-White ID †
Full Sample 745 0.35 0.04 <0.001
Population Density Quartile
 Low Density (<2027) 186 0.29 0.05 <0.001
 Low-Mid Density (2027-<3048) 186 0.14 0.05 <0.001
 Mid Density (3048-<4758) 186 0.26 0.06 <0.001
 High Density (≥4758) 187 0.34 0.07 <0.001
Percentage Black Quartile
 Low (<2.7%) 186 0.02 0.10 0.83
 Low-Mid (2.7-<6.8%) 186 0.30 0.11 0.008
 Mid (6.8-<17.2%) 187 0.50 0.09 <0.001
 High (≥17.2%) 186 0.49 0.06 <0.001
White-Latino ID
Full Sample 745 0.22 0.05 <.001
Population Density Quartile
 Low Density (<2027) 186 0.24 0.06 <.001
 Low-Mid Density (2027-<3048) 186 0.16 0.06 0.008
 Mid Density (3048-<4758) 186 0.32 0.06 <.001
 High Density (≥4758) 187 0.16 0.08 0.044
Percentage Latino Quartile
 Low (<7.7%) 186 0.13 0.09 0.180
 Low-Mid (7.7-<16.4%) 186 0.19 0.11 0.092
 Mid (16.4-<32.6%) 187 0.37 0.09 <0.001
 High (≥32.6%) 186 0.21 0.09 0.026
Open in a new tab
*

All linear regression models control for region, poverty, and high school graduation.

†

ID = Index of Dissimilarity, B = Beta coefficient, SE = Standard Error

ID and walkability also varied when stratified by city percentage minoritized population, as expected. There was no association between Black-White ID and walkability in cities with low percentage (<2.7%) Black residents (β=0.02, p=0.83). Similarly, there was no association between Latino-White ID and walkability among cities with low (<7.7%) or low-mid percentage (1.7-<16.4%) Latino residents. In other strata, for both groups, coefficients were between 0.21 – 0.50 and statistically significant.

Discussion

The present research explored possible associations between city-level RRS and walkability. Though this is not the first time investigators have explored an association between race and walkability, RRS is rarely examined, and this potential association has not been examined among a sample of cities as large as that examined here.

We observed a positive association between RRS and walkability. Median Black-White and Latino-White ID scores were similar, as were full sample regression coefficients. Observed associations persisted across population density strata. Our findings suggest ID is associated with walkability across strata of population density for both minoritized groups examined. This indicates the association between ID and walkability cannot be fully explained by population density. Our primary hypothesis – that RRS would be associated with walkability- appears to be accurate.

Given population density is a part of the WalkScore calculation, we sought to explore how the observed association changed across population density strata. The association persisted across most strata of population density, but the estimate was attenuated in most strata compared to non-stratified effect size estimates. This suggests that population density plays an important role in the association between RRS and walkability, but does not fully explain the association.

For both Black-White and Latino-White ID, associations were weakest or non-significant in cities with low and low-mid percentage minoritized populations. However, significant associations were observed in mid (Black and Latino) and high (Black only) percentage minoritized residents cities, suggesting city-level percentage minoritized partially, but does not fully, explain the association between ID and walkability.

There was not a strong tract-level correlation between percentage Black or Latino residents and walkability. Though this simple correlation cannot conclusively disprove our secondary hypothesis – that city-level associations between RRS and walkability are driven by dense, highly walkable, high minoritized population neighborhoods- the correlation results do not necessarily support our secondary hypothesis either.

Though the secondary hypothesis concerns tract-level drivers of city-level associations, and other investigators have used tract-level RRS measures28, tract-level RRS would be inappropriate to test the present hypothesis because census tract boundaries are likely too small to capture RRS at the scale at which we are interested26,29. Tract boundaries are drawn to encompass approximately 4,000 individuals30, which may be only a few city blocks in densely populated cities. Also, ID is calculated based on the demographic composition of smaller geographic units nested within a larger unit of interest, e.g. census tracts nested within cities. Calculating tract-level ID would require the use of census blocks, which may be only one city block in size in dense areas, or census block groups; there can be a maximum of nine block groups in a tract. The small size of these geographic units, and the small number of these units located in each tract, may result in insufficient heterogeneity within each tract to produce meaningful ID values. We believe measuring the association between walkability and RRS is best at the city-level, an approach also employed by other investigators31.

These results do not provide a straightforward explanation for the observed association between RRS and walkability. A better understanding of what factors drive higher city- and neighborhood-level walkability, including what drives the observed association between RRS and walkability- can guide planning of walkable urban environments, which could in turn encourage more physical activity. Additional research should more thoroughly examine the role population density plays in this association, as well as other potential drivers, like higher total retail demand in high minoritized population neighborhoods, and other built environment factors like block length and intersection density23.

Our study has limitations. Although higher WalkScore is generally associated with more walking, it does not include subjective measures of the walking environment that may affect walking behavior, like residents’ opinions on neighborhood safety or quality of sidewalks, which may be lower in impoverished or high minoritized population neighborhoods32. Furthermore, other environmental health hazards may be driving higher disease prevalence in high minoritized neighborhoods despite high walkability, including for example poor air quality33, low healthcare access34,35, social stressors36, and others.

Conclusions

This research surfaces important new questions about the relationship between RRS, walkability and physical activity. Studies directly observing walking behaviors in diverse urban settings may provide further insight into whether and how neighborhood walkability attributes induce walking behavior in more and less racially segregated cities. This knowledge could then be leveraged to further inform urban design interventions to increase physical activity, particularly in segregated cities.37

Minoritized populations live in highly walkable US cities and neighborhoods

Despite this, minoritized populations also suffer high prevalence of chronic disease

Racial residential segregation and population density are associated with walkability

However, these factors could not fully explain this counterintuitive association

Racial residential segregation may proxy walkability-related city characteristics

Funding:

This work was supported by the Robert Wood Johnson Foundation grant number #78440 and the National Institutes of Health grant number R01DK108682

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  • 1.Northridge ME, Sclar ED, Biswas P. Sorting out the connections between the built environment and health: a conceptual framework for navigating pathways and planning healthy cities. Journal of urban health. 2003;80(4):556–568. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Kim B, Spoer BR, Titus AR, et al. Life Expectancy and Built Environments in the US: A Multilevel Analysis. American Journal of Preventive Medicine. 2023;64(4):468–476. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Duncan DT, Aldstadt J, Whalen J, Melly SJ, Gortmaker SL. Validation of walk score for estimating neighborhood walkability: an analysis of four US metropolitan areas. Int J Environ Res Public Health. Nov 2011;8(11):4160–79. doi: 10.3390/ijerph8114160 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.King KE, Clarke PJ. A disadvantaged advantage in walkability: findings from socioeconomic and geographical analysis of national built environment data in the United States. Am J Epidemiol. Jan 1 2015;181(1):17–25. doi: 10.1093/aje/kwu310 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Howell NA, Tu JV, Moineddin R, Chu A, Booth GL. Association between neighborhood walkability and predicted 10-year cardiovascular disease risk: The CANHEART (cardiovascular health in ambulatory care research team) cohort. Journal of the American Heart Association. 2019;8(21):e013146. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Creatore MI, Glazier RH, Moineddin R, et al. Association of neighborhood walkability with change in overweight, obesity, and diabetes. Jama. 2016;315(20):2211–2220. [DOI] [PubMed] [Google Scholar]
  • 7.Jones AC, Chaudhary NS, Patki A, et al. Neighborhood walkability as a predictor of incident hypertension in a national cohort study. Frontiers in public health. 2021;9:2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Yang S, Chen X, Wang L, et al. Walkability indices and childhood obesity: A review of epidemiologic evidence. Obesity reviews. 2021;22:e13096. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Conderino SE, Feldman JM, Spoer B, Gourevitch MN, Thorpe LE. Social and Economic Differences in Neighborhood Walkability Across 500 U.S. Cities. Am J Prev Med. Sep 2021;61(3):394–401. doi: 10.1016/j.amepre.2021.03.014 [DOI] [PubMed] [Google Scholar]
  • 10.Hales CM, Fryar CD, Carroll MD, Freedman DS, Aoki Y, Ogden CL. Differences in Obesity Prevalence by Demographic Characteristics and Urbanization Level Among Adults in the United States, 2013–2016. JAMA. 2018;319(23):2419–2429. doi: 10.1001/jama.2018.7270 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Beckles GL, Chou C-F. Disparities in the prevalence of diagnosed diabetes—United States, 1999–2002 and 2011–2014. Morbidity and Mortality Weekly Report. 2016;65(45):1265–1269. [DOI] [PubMed] [Google Scholar]
  • 12.Fang J, Gillespie C, Ayala C, Loustalot F. Prevalence of self-reported hypertension and antihypertensive medication use among adults aged≥ 18 years—United States, 2011–2015. Morbidity and Mortality Weekly Report. 2018;67(7):219. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Williams DR, Collins C. Racial residential segregation: a fundamental cause of racial disparities in health. Public Health Rep. Sep-Oct 2001;116(5):404–416. doi: 10.1093/phr/116.5.404 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Kwate NOA, Loh JM, White K, Saldana N. Retail redlining in New York City: racialized access to day-to-day retail resources. Journal of Urban Health. 2013;90(4):632–652. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Rugh JS, Massey DS. SEGREGATION IN POST-CIVIL RIGHTS AMERICA: Stalled Integration or End of the Segregated Century? Du Bois Review: Social Science Research on Race. 2014;11(2):205–232. doi: 10.1017/S1742058X13000180 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Bond Huie SA. The components of density and the dimensions of residential segregation. Population Research and Policy Review. 2000;19(6):505–524. [Google Scholar]
  • 17.Aaronson D, Hartley D, Mazumder B. The Effects of the 1930s HOLC “Redlining” Maps. American Economic Journal: Economic Policy. 2021;13(4):355–92. doi: 10.1257/pol.20190414 [DOI] [Google Scholar]
  • 18.Anderson KF, Lopez A, Simburger D. Racial/Ethnic Residential Segregation and the First Wave of SARS-CoV-2 Infection Rates: A Spatial Analysis of Four U.S. Cities. Sociological Perspectives. 2021;64(5):804–830. doi: 10.1177/07311214211041967 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Lynch EE, Malcoe LH, Laurent SE, Richardson J, Mitchell BC, Meier HCS. The legacy of structural racism: Associations between historic redlining, current mortgage lending, and health. SSM - Population Health. 2021/06/01/ 2021;14:100793. doi: 10.1016/j.ssmph.2021.100793 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Gourevitch MN, Athens JK, Levine SE, Kleiman N, Thorpe LE. City-level measures of health, health determinants, and equity to foster population health improvement: the city health dashboard. American journal of public health. 2019;109(4):585–592. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Duncan DT, Aldstadt J, Whalen J, Melly SJ. Validation of Walk Scores and Transit Scores for estimating neighborhood walkability and transit availability: a small-area analysis. GeoJournal. 2013/04/01 2013;78(2):407–416. doi: 10.1007/s10708-011-9444-4 [DOI] [Google Scholar]
  • 22.Walkscore. Walk Score Methodology. Accessed October 17th, 2022. https://www.walkscore.com/methodology.shtml
  • 23.The development of a walkability index: application to the Neighborhood Quality of Life Study. British journal of sports medicine. 2010;44(13):924–933. [DOI] [PubMed] [Google Scholar]
  • 24.Massey DS, Denton NA. The dimensions of residential segregation. Social forces. 1988;67(2):281–315. [Google Scholar]
  • 25.Bureau USC. Housing Patterns: Appendix B: Measures of Residential Segregation. [Google Scholar]
  • 26.Lee BA, Reardon SF, Firebaugh G, Farrell CR, Matthews SA, O’Sullivan D. Beyond the census tract: Patterns and determinants of racial segregation at multiple geographic scales. American sociological review. 2008;73(5):766–791. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Chiu M, Shah BR, Maclagan LC, Rezai M-R, Austin PC, Tu JV. Walk Score® and the prevalence of utilitarian walking and obesity among Ontario adults: a cross-sectional study. Health Reports. 2015;26(7):3. [PubMed] [Google Scholar]
  • 28.Salow AD, Pool LR, Grobman WA, Kershaw KN. Associations of neighborhood-level racial residential segregation with adverse pregnancy outcomes. American journal of obstetrics and gynecology. 2018;218(3):351. e1–351. e7. [DOI] [PubMed] [Google Scholar]
  • 29.Kramer MR, Cooper HL, Drews-Botsch CD, Waller LA, Hogue CR. Do measures matter? Comparing surface-density-derived and census-tract-derived measures of racial residential segregation. International journal of health geographics. 2010;9:1–15. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.U.S. Census Bureau. United States Census 2000; Participant Statistical Areas Program Guidelines. 1997. [Google Scholar]
  • 31.Russell EF, Kramer MR, Cooper HL, Gabram-Mendola S, Senior-Crosby D, Jacob Arriola KR. Metropolitan area racial residential segregation, neighborhood racial composition, and breast cancer mortality. Cancer Causes & Control. 2012;23:1519–1527. [DOI] [PubMed] [Google Scholar]
  • 32.Rajaee M, Echeverri B, Zuchowicz Z, Wiltfang K, Lucarelli JF. Socioeconomic and racial disparities of sidewalk quality in a traditional rust belt city. SSM - Population Health. 2021/12/01/ 2021;16:100975. doi: 10.1016/j.ssmph.2021.100975 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Pratt GC, Vadali ML, Kvale DL, Ellickson KM. Traffic, Air Pollution, Minority and Socio-Economic Status: Addressing Inequities in Exposure and Risk. International Journal of Environmental Research and Public Health. 2015;12(5):5355–5372. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Qato DM, Daviglus ML, Wilder J, Lee T, Qato D, Lambert B. ‘Pharmacy deserts’ are prevalent in Chicago’s predominantly minority communities, raising medication access concerns. Health Affairs. 2014;33(11):1958–1965. [DOI] [PubMed] [Google Scholar]
  • 35.Liu M, Kao D, Gu X, Holland W, Cherry-Peppers G. Oral Health Service Access in Racial/Ethnic Minority Neighborhoods: A Geospatial Analysis in Washington, DC, USA. International Journal of Environmental Research and Public Health. 2022;19(9):4988. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Cutts BB, Darby KJ, Boone CG, Brewis A. City structure, obesity, and environmental justice: An integrated analysis of physical and social barriers to walkable streets and park access. Social science & medicine. 2009;69(9):1314–1322. [DOI] [PubMed] [Google Scholar]
  • 37.Loh THL CB. The economic power of walkability in metro areas. The Avenue The Brookings Institute; 2019. July 12, 2019. Accessed October 17th, 2022. https://www.brookings.edu/blog/the-avenue/2019/07/12/the-economic-power-of-walkability-in-metro-areas/

RESOURCES