Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 Jul 1.
Published in final edited form as: Health Place. 2020 Jun 15;64:102361. doi: 10.1016/j.healthplace.2020.102361

Do Investments in Low-Income Neighborhoods Produce Objective Change in Health-Related Neighborhood Conditions?

Stephanie Brooks Holliday 1, Wendy Troxel 1, Ann Haas 2, Madhumita (Bonnie) Ghosh-Dastidar 2, Tiffany L Gary-Webb 3, Rebecca Collins 1, Robin Beckman 2, Matthew Baird 2, Tamara Dubowitz 1
PMCID: PMC8055100  NIHMSID: NIHMS1604851  PMID: 32838886

Abstract

This study examined the effect of neighborhood investments on neighborhood walkability, presence of incivilities, and crime in two low-income, primarily African American neighborhoods in Pittsburgh, USA. During the study period, one of the neighborhoods (the intervention neighborhood) received substantially more publicly-funded investments than a demographically matched comparison neighborhood. Comparisons between the neighborhoods showed a significant difference-in-difference for all three outcomes. The intervention neighborhood experienced significantly more change related to improved walkability and decreased incivilities. However, the control neighborhood experienced better crime-related outcomes. Analyses that focused on resident proximity to investments found similar results. This highlights the nuances of neighborhood investment, which is important to consider when thinking about public policy.

Keywords: neighborhood, walkability, built environment, crime

There has been growing interest in revitalizing urban neighborhoods that have been subject to historical disinvestment with the goal of improving neighborhood conditions, access to services, and ultimately the wellbeing of residents (Tatian et al., 2012). It is often assumed that community investments result in improvements to the neighborhood. Although investments can take many forms, including federal grant programs, local public investments, and private initiatives often pooled with public dollars, researchers have highlighted a need to examine whether strategic allocation or processes related to development implementation can “trigger the revitalization of distressed, low-income urban neighborhoods” (Galster et al., 2006). In particular, it is unclear whether investments actually impact those features which may directly impact residents’ livability, such as objectively-assessed walkability and incivilities, and incident reports of crime.

Walkability has been shown to be associated with increased physical activity, lower obesity, and higher self-rated health (Saelens et al., 2003, Coombes et al., 2010). Factors that contribute to walkability of a neighborhood may include aesthetic conditions (e.g., trees along the streets, glass-fronted shopfronts), features of the infrastructure (e.g., wide sidewalks, parking lots); as well as street connectivity and low traffic volume (Negron-Poblete et al., 2016, Talen and Koschinsky, 2013). Such factors are often assessed by self-reported resident perceptions and sometimes by audits of streets (Lee and Talen, 2014). However, there is a dearth of research examining change in objectively-measured walkability following neighborhood investments of this nature.

Neighborhood incivilities, or indicators of physical and social disorder (Rossen et al., 2011), can include litter, graffiti, and poorly maintained structures (Brown et al., 2004b). The presence of incivilities has been associated with higher levels of crime (Brown et al., 2004b) and lower perceived safety among neighborhood residents (Austin et al., 2002) and, in turn, associated with worse physical and mental health (Ellaway et al., 2009, Gapen et al., 2011, Hill et al., 2005). There is some evidence that investing in neighborhoods can decrease the presence of incivilities. For example, a Birmingham, AL study of HOPE VI, a large effort by the U.S. Housing and Urban Development to revitalize public housing, found that at the neighborhood level, there was no significant decrease in perceived disorder among residents, but those who lived closest to the HOPE VI target area reported less disorder (Dulin-Keita et al., 2015). Another study examined the effect of a new subdivision in a distressed neighborhood (Brown et al., 2004a). Before the subdivision was built, the presence of objectively measured incivilities was highest in areas close to the redevelopment site. After it was built, incivilities were reduced in the zones closest to the redevelopment, though not eliminated.

Neighborhood crime can have pervasive impacts on health and quality of life of residents, including risk for anxiety and depressive disorders (Stockdale et al., 2007); physical health concerns (e.g., reduced physical activity (Janke et al., 2016), coronary heart disease (Browning et al., 2012)); and worse self-rated health (Agyemang et al., 2007). Features of the built environment associated with more crime include the presence of mixed (vs. residential) land use, street density, bus stop density, and the average height of buildings (Sohn, 2016, Twinam, 2017). Moreover, the Crime Prevention through Environmental Design (CPTED) approach suggests that design principles in the built environment can be used to deter crime, such as pedestrian walkways, lighting outside buildings, and landscaping that allows for visibility (Carter et al., 2003, Thani et al., 2016). Studies of multicomponent CPTED interventions have found decreases in calls for police service, burglary, and violent crimes, though the effect on drug and violent crimes is less-well established (Carter et al., 2003, Cozens et al., 2005, Welsh and Farrington, 2008).

Present Study

Although research has highlighted the effects that investments in neighborhoods can have on neighborhood conditions, there has been limited longitudinal assessment of changes in the neighborhood environment before and after such investments. Most work has examined the impact on resident perceptions rather than objective neighborhood features, which may in turn influence residents’ satisfaction or perceptions of livability of the community. The present study contributes to this literature by exploring the effect of neighborhood investments at least partially funded with public dollars (i.e., at least $1 of public funding) on objectively-rated neighborhood conditions in two neighborhoods – an intervention and control neighborhood – within a natural experiment.

We focused on two low-income, primarily African American neighborhoods in Pittsburgh, Pennsylvania: the Hill District and Homewood. The neighborhoods were matched at baseline with a very similar sociodemographic profile based on census data (United States Census Bureau, American Community Survey data 2005–2009). During the study period of 2013 to 2016, investments that were at least partially publicly funded valued $193,628,994 in Hill District. Such initiatives included the building of a full service supermarket, a new energy innovation center (dedicated to workforce development and business incubation), two new housing developments totaling 194 units, and a community center (Dubowitz et al., 2019). There were also efforts to improve the streetscape (e.g., improved sidewalks, plantings, and street crossings) and park and greenspace renovations. Publicly-funded investments were also made in Homewood (the control neighborhood) during this same time, including senior housing with 41 new units, a public housing development with 40 new units, and a new business hub hosting local small businesses and non-profits, though the total investment was much lower ($47,516,268) (Dubowitz et al., 2019).

A previous study by our research team comparing the two neighborhoods before and after investments found that residents of both neighborhoods experienced improved perceptions of aesthetics, safety, and having many places within easy walking distance of home; however, the magnitude of changes across the neighborhoods were equivalent, and there were no significant differences in resident health outcomes (Dubowitz et al., 2019). Analyses examining the effect of resident proximity to investment across neighborhoods yielded similar results.

Many of the changes to the neighborhoods were theorized to influence walkability, incivilities and crime, three outcomes that our team hypothesized to be important to numerous health outcomes. Yet, documentation and assessment of the impact of investments on these outcomes was lacking. The current study builds on this gap by examining the effect of neighborhood investment on three objective neighborhood conditions: walkability and incivilities, measured by street audits, and crime, measured by City of Pittsburgh Police documentation of all reported incidents of crime. We examined these outcomes as a function of participants’ neighborhood (i.e., were greater improvements observed in the intervention neighborhood?) and as a function of participants’ proximity to investment (i.e., were there greater improvements in residential addresses closest to the investments regardless of neighborhood?). We hypothesized that conditions in Hill District would improve significantly more than conditions in Homewood, and that areas closest to the investments would experience significantly greater improvements.

Methods

This study is part of the Pittsburgh Research on Neighborhood Change and Health (PHRESH), a series of studies that have leveraged a natural experiment opportunity designed to examine the impact of neighborhood-level investments in the Hill District and a demographically similar neighborhood, Homewood. A full description of the study design and the geographic areas considered to be neighborhoods is described elsewhere (Dubowitz et al., 2015a, Dubowitz et al., 2015b). The study included longitudinal data from person-level surveys, audits of street segments within both neighborhoods, and crime data.

To enroll households, in 2011 a stratified random sample of residential addresses was drawn from the two neighborhoods, with oversampling of households in the intervention neighborhood (Hill District). The primary food shopper of each household was recruited into the study. In 2013, 1003 households still lived within the study neighborhoods and participated in data collection. Of the 2013 completes, 45 (4.5%) of participants had died, 56 (5.6%) had moved out of the study area, and 17 (1.7%) were physically or mentally unable to complete the survey in 2016. Of the households eligible to complete a 2016 survey, we were unable to contact 121 and another 88 refused to participate for an analytic sample size of 676 households. Although there were differences between participants attritors and non-attritors with respect to certain sociodemographic characteristics, attrition weights were used to account for these differences (described in more detail below). Additionally, due to item missingness in the outcome variables at either time point or in the proximity to investment indicator, analytic sample sizes may vary as described below.

We examined neighborhood conditions based on participant location. More specifically, we examined changes in the neighborhood that took place within, on average, a ¼-mile radius of each enrolled household. This enabled us to use person-level variables to control for differences between the two neighborhoods at baseline and as a result of attrition. Data came from three sources: a street segment audit; crime data from the City of Pittsburgh police department; and a survey administered to each participant. All procedures were approved by the local Institutional Review Board.

Measures

Street Segment Audit Data

At each data collection period, street segment audits were conducted by four trained data collectors using a tool adapted from the Bridging the Gap Street Segment Tool (Kelly et al., 2007, Slater et al., 2013). Using geographic shapefiles provided by ESRI and the City of Pittsburgh, we located all streets at baseline located within the study neighborhoods as well within ¼ mile of each neighborhood boundary. Twenty-five percent of the street segments (i.e., roads that extended between one intersection and another intersection or dead end) were randomly sampled and included as part of our street segment data.

Each household’s segment characteristics were averaged within a ¼ mile buffer, a radius based on being able to collect a minimum number of segments for each household enrolled in the study. The following objective neighborhood outcomes were measured in both 2013 and 2016.

Walkability.

Assessed via neighborhood audits, as described above, the 15-item walkability index assessed traffic signs at intersections, pedestrian crossings, sidewalks, lighting, transit, and mixed use. Scores range from 0 to 22, and higher scores indicate greater walkability.

Incivilities.

Incivilities were also assessed via audit. We assessed aesthetics, litter, vacant housing, bars on windows, broken windows, and the data collector reporting that they did not feel safe on a given segment. Scores range from 0 to 8 and higher scores indicate greater presence of incivilities.

Crime

We obtained incident-level crime data from the City of Pittsburgh police department. All crime incidents reported by the City of Pittsburgh were geocoded and street network distances from each household to each approximate crime location were calculated using ArcGIS 10.2 software. For each resident, we summed the total number of property and violent crimes that occurred in the year of survey administration (for both 2013 and 2016) within 1/10th of a mile of a residence.

Survey Data

Participants completed a 60-minute in-person survey at both wave, using interviewer-administered computer assisted personal interviewing. Participants received an incentive of $25 for completing a survey.

Although participant neighborhoods were demographically similar, individual-level characteristics may be associated with proximity to investment. Thus, covariates collected via survey were included to enhance comparability. Covariates had a low rate of missingness and were imputed where missing to retain all observations. Participants reported on: age, gender (male or female), household income, marital status (married or unmarried), level of education (categorized as less than high school education, high school education, some college, or college degree or higher), any children in the household (yes or no), and years in neighborhood.

Data Analysis

Participants in the study were characterized in two ways: first, according to their neighborhood (Hill District or Homewood), and second, based on participant distance from investments (based on baseline address). Building on our prior work (Dubowitz et al., 2019), we created a binary variable to reflect whether a household was within one-tenth of a mile of a neighborhood investment effort during the study period or farther than one-tenth of a mile from any investments. We chose a proximal distance because both neighborhoods are relatively small geographic areas and prior unpublished analyses showed that proximal neighborhood environments were associated with relevant health outcomes (i.e., sleep). We were unable to ascertain proximity to investment for three participants, and these cases were not included in analyses of change by proximity to investment. Our analyses were conducted as intent-to-treat, such that participant neighborhood was assigned based on location at baseline.

We estimated difference-in-differences by each neighborhood using linear regression of each outcome on 1) study year, 2) neighborhood, and 3) an interaction between study year and neighborhood, with this third term estimating the difference-in-difference. Baseline covariates were included. Models were weighted to account for attrition between baseline and follow-up to ensure that results generalize to the baseline sample. Attrition weights were derived as the inverse probability of response at follow-up, estimated using a logistic regression model with socio-demographics and additional baseline characteristics as predictors. Models accounted for correlations between repeated measurements of each participant using person-level random effects. After selecting for complete data, the analytic sample for 577 individuals for walkability and incivilities and 637 for crime.

We ran a similar set of models to compare difference-in-differences by proximity to investment. We employed a doubly-robust approach for these models, adjusting for covariates by including them as predictors in the models and by applying inverse probability of treatment weighting (IPTW) using propensity scores. Propensity scores were estimated using generalized boosted models to make each group [near investment, farther from investment] comparable to the combined set of respondents. The R package TWANG (Toolkit for Weighting and Analysis of Nonequivalent Groups) was used to create these IPTWs (Ridgeway et al., 2017, Austin, 2011). After applying these, adequate balance (absolute standardized mean difference <0.25) was achieved for all covariates. The estimation of IPTWs and assessment of sample balance analysis included attrition weights; product weights (IPTW x attrition weights) were used in modeling.

Results

Demographic characteristics appear in Table 1. Participants’ mean age was 55 years (SD = 15.2). Though the mean age was similar in both Homewood and Hill District, participants who lived closer to investments tended to be older than those who lived farther from investments (61.5 vs. 53.7 years old, respectively). Participants were largely female (77.4%) and had an annual household income of $21,400. There were no neighborhood differences in education; however, those who lived closer to investments were more likely to have less than a high school education. On average, residents of the Hill District had lived in their neighborhood longer than residents of Homewood (35.3 vs. 23.5 years, respectively).

Table 1.

Baseline characteristics of study participants.

Overall % or M (SD) By neighborhood
 By proximity to investment
Hill District % or M (SD) Homewood % or M (SD) p-valuea Near < 0.1 mile % or M (SD) Near < 0.1 mile % or M (SD) p-valuea
N 676 481 195 119 554
Age (years) 55.0 (15.2) 55.5 (15.4) 53.9 (14.6) 0.23 61.5 (14.7) 53.7 (15.0) <.01
Male 22.6% 20.7% 26.6% 0.13 20.7% 23.1% 0.59
African American/Black 95.7% 95.7% 95.8% 0.93 94.2% 96.2% 0.34
Annual Household Income ($) 21,400 (19,500) 21,300 (19,200) 21,700 (20,300) 0.77 20,600 (19,800) 21,600 (19,500) 0.64
Married/living with partner 21.2% 19.0% 25.7% 0.11 9.4% 23.3% <.01
Education 0.18 <.01
 < High school 12.0% 12.9% 10.2% 23.1% 9.8%
 High school 39.8% 41.7% 35.7% 37.8% 40.2%
 Some college 33.4% 32.5% 35.4% 28.7% 34.5%
 College 14.8% 12.9% 18.8% 10.4% 15.4%
Any children in household 27.7% 25.7% 32.0% 0.15 18.6% 29.5% 0.02
Years in neighborhood 31.5 (22.6) 35.3 (23.0) 23.5 (19.2) <.01 34.2 (24.1) 31.0 (22.3) 0.16
Proximity to investment < 0.1
mile 16.6% 21.5% 6.6% <.01 n/a n/a n/a
Hill District 67.6% n/a n/a n/a 87.2% 63.7% <.01
Open in a new tab
a

based on results of t-test or chi square

Table 2 presents the results of the difference-in-difference analyses by neighborhood. There were significant difference-in-differences for all of our outcomes of interest. For walkability, respondents in the Hill District experienced a significant increase in walkability (mean (M) = 1.55; p<0.001), whereas respondents in Homewood did not experience a significant change (mean net difference = 1.29; p < 0.001). Although incivilities remained stable in the Hill District, in Homewood there was a significant increase of 0.35 incivilities observed on the audited street segments (p < .001), for a net difference of 0.40 (p < .001) fewer incivilities observed in Hill District. By contrast, for all categories of crime (total crime, property crime, and violent crime), Hill District fared worse than Homewood. Comparing 2013 to 2016, the number of annual property crimes within 0.1 mile of a Hill District respondent’s home increased by 1.59 crimes, on average, relative to a Homewood respondent’s home (p < .05). For violent crime, Hill District respondents experienced a significant increase of 1.25 crimes (p < .001), whereas Homewood respondents experienced a significant decrease of −1.44 (p < .001), for a net difference of 2.69 crimes (p < .001).

Table 2.

Difference-in-difference models by neighborhood

Hill District (intervention neighborhood) Homewood (comparison neighborhood) Difference-in-Difference
Baseline mean (SE) Change 2013–2016 (SE) Baseline mean (SE) Change 2013–2016 (SE) HD Change - HW Change
Walk Score 7.78 (0.08) 1.55 (0.10) *** 7.11 (0.12) 0.26 (0.14) 1.29 (0.17) ***
Incivility 4.56 (0.04) −0.05 (0.05) 5.41 (0.06) 0.35 (0.07) *** −0.40 (0.08) ***
Total crime 22.16 (0.81) 2.23 (0.74) ** 25.22 (1.28) −1.61 (1.08) 3.84 (1.30) **
Property crime 7.46 (0.32) 2.34 (0.38) *** 9.28 (0.49) 0.74 (0.56) 1.59 (0.68) *
Violent crime 6.74 (0.35) 1.25 (0.34) *** 9.48 (0.54) −1.44 (0.50) ** 2.69 (0.61) ***
Open in a new tab
*

p<.05,

**

p<.01,

***

p<.001;

Note: all models adjust for covariates and include weighting to account for attrition between baseline and follow-up. Covariates included age, gender, household income, marital status, level of education, any children in household, and years in neighborhood.

Table 3 presents the results of the difference-in-difference analysis by proximity to investment. Walkability increased significantly both near investment (by 1.67) and farther from investment (by 1.00), but the magnitude of the increase was larger for those living within one-tenth of a mile of investments versus those living farther (difference-in-difference = 0.67, p < .001). The number of observed incivilities decreased near investments (M = −0.28) but increased farther from investments (M = 0.15); the net difference was −0.43 (p < 0.001). The number of property crimes increased significantly for those living near investments, but there was no significant change for those living farther from investments (mean net difference = 6.37, p < .01). There was no significant change in violent crime based on investment proximity, and the difference-in-difference was not significant.

Table 3.

Difference-in-difference models by proximity to investment

Intervention (Proximity to investment < 0.1 mile) Comparison (Proximity to investment 0.1 mile or more) Difference-in-Difference
Baseline mean (SE) Change 2013–2016 (SE) Baseline mean (SE) Change 2013–2016 (SE) Intervention change – Comparison change
Walk Score 7.74 (0.12) 1.67 (0.11) *** 7.56 (0.09) 1.00 (0.10) *** 0.67 (0.15) ***
Incivility 4.28 (0.06) −0.28 (0.05) *** 4.95 (0.04) 0.15 (0.05) ** −0.43 (0.07) ***
Total crime 30.09 (1.31) 4.34 (0.97) *** 21.38 (0.81) 0.16 (0.88) 4.18 (1.30) **
Property crime 9.65 (0.45) 6.93 (0.46) *** 7.62 (0.33) 0.56 (0.42) 6.37 (0.62) ***
Violent crime 9.53 (0.55) −0.14 (0.47) 7.11 (0.36) 0.43 (0.42) −0.58 (0.63)
Open in a new tab
*

p<.05

**

p<.01

***

p<.001

Note: all models adjust for covariates and include attrition, as well as propensity score weighting, to account for sample attrition between baseline and follow-up and for imbalance in baseline covariates between the two groups (lives < .1 mile of investment, lives >= .1 mile of investment), respectively. Covariates included age, gender, household income, marital status, level of education, any children in household, and years in neighborhood.

Discussion

This study examined the impact of neighborhood-level developments on objectively-measured physical and social neighborhood conditions in two disinvested neighborhoods in Pittsburgh. Households in the Hill District (the intervention neighborhood) experienced significant improvements in walkability, with no significant change for households in Homewood. Analyses that focused on proximity to investment demonstrated that residents who lived in close proximity to investments also experienced significantly greater increases in walkability than residents who lived farther from investment, regardless of neighborhood. Neighborhood revitalization efforts often intend to improve characteristics such as walkability, but documentation of whether such improvements lead to such changes is lacking and inconsistent. For example, prior research found that the introduction of mixed-income housing and creation of green spaces did not have a significant effect on resident perceptions of walkability.(Dulin-Keita et al., 2015) This may reflect a difference between resident perceptions and objectively measured walkability, but it may also be that improvements in walkability were driven by other changes in the built environment (e.g., improvements in sidewalks, introduction of the community center, new retail outlets).

Hill District fared better than Homewood with respect to incivilities; however, this was driven by an increase in incivilities in Homewood. Consistent with previous research (Dulin-Keita et al., 2015, Brown et al., 2004a), though, areas closest to the investments experienced significant decreases in incivilities compared to areas farther from investments. These findings suggest that promixity to investment may be a more important factor in reducing incivilities. They also suggest that a range of neighborhood investments can reduce neighborhood disorder, including those measured in our study and previous work (e.g., introduction of new housing and businesses, creating green spaces, improving the streetscape).

Contrary to expectations, we found that the Hill District had worse crime-related outcomes than Homewood, as did areas closest to the neighborhood investments. There has been some literature demonstrating that gentrification efforts can result in increased rates of crime. Researchers hypothesize that gentrification introduces a destabilizing process that results in initial increases in crime, though longer-term decreases are often observed (Kirk and Laub, 2010). It is possible that the neighborhood investments had some destabilizing effects. Alternatively, new amenities in these neighborhoods, such as retail stores, have increased the presence of residents and visitors, which may result in higher levels of property-related crimes; similarly, certain aspects of walkability (e.g., increased street connectivity) may also increase crime (Foster et al., 2010). It would be valuable to continue to study these neighborhood over times to determine the longer-term impact on crime.

This study has certain limitations. First, a fair number of participants were lost to follow-up in 2016, though attrition weights were used to account for differential attrition. Second, because 2013 was the first year that street segment audits were conducted, we were unable to explore whether there were any trends toward improvement in neighborhood conditions in these neighborhoods prior to that time, and if the time trends in years prior to 2013 were parallel in the two neighborhoods. During the course of this longitudinal study, we have downloaded extensive Census SF1 data (1990, 2000, 2010) and ACS 5-year census tract-level estimates on variables related to demographic, socioeconomic, and household composition characteristics. Using data from ACS 2005–2009 and ACS 2010–2014, we have been able to test for parallel trends in the two neighborhoods on these sociodemographic characteristics, which are related to neighborhood outcomes. We have observed a few differences between the two neighborhood, though these differences shrank over time. Ideally, we would have had Census data at the time points preceding the “pre” assessment for test of parallel trends, but had to use ACS 5-year point estimates, which have high sampling error. Finally, though we opted to focus on one-tenth of a mile for proximity-focused analyses, looking at additional distances (e.g., living within half a mile or one mile from investments) could reveal additional trends related to the influence of proximity to investments.

Despite these limitations, understanding the effect of place-based investments on walkability, incivilities, and crime has potential implications for the health and policies that impact neighborhood residents. Neighborhood revitalization efforts such as those observed in the present study generally target low-income areas that have been disinvested. Individuals from disadvantaged neighborhoods are at increased risk for outcomes such as cardiovascular disease morbidity and mortality (Bosma et al., 2001, Diez Roux et al., 2001). In addition, in part due to patterns of residential segregation, African Americans are disproportionately exposed to neighborhood disadvantage and co-existing neighborhood stressors (Williams and Collins, 2001). By addressing upstream socio-environmental factors that contribute to health, neighborhood-level investments have the potential to effect change on a large scale, as they have the reach to impact an entire community. Neighborhood-level interventions also address the broader contextual influences that impact population health and contribute to the persistence of racial/ethnic health disparities (Browning et al., 2012).

Our findings, based on longitudinal and objective data collection, demonstrated that neighborhood-level investments have the potential to improve walkability and incivilities, both of which have been associated with downstream health outcomes in neighborhood residents, including health behaviors (e.g., physical activity) (Janke et al., 2016, Saelens et al., 2003) and physical and mental health (Ellaway et al., 2009, Gapen et al., 2011, Stockdale et al., 2007). As this study continues, an important next step that we will be taking will be to examine whether these changes in neighborhood conditions have the expected effects on health. Another future direction could be to explore how cost savings associated with improved health offset the costs of investments. Ultimately, this work contributes to understanding how changes to the built and social environment can impact the livability of neighborhoods, in turn informing decisions about how and where to invest in neighborhood revitalization.

Highlights.

  • We examined the effect of neighborhood investments on neighborhood conditions.

  • This study built upon a natural experiment of two low-income neighborhoods.

  • Neighborhood investments were associated with improved walkability and incivilities.

  • However, investments were associated with increased crime

Acknowledgments

This study was funded by the National Heart, Lung, and Blood Institute (R01HL122460 and R01HL131531) and the National Cancer Institute (R01CA164137).

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. AGYEMANG C, VAN HOOIJDONK C, WENDEL-VOS W, LINDEMAN E, STRONKS K & DROOMERS M 2007. The association of neighbourhood psychosocial stressors and self-rated health in Amsterdam, the Netherlands. Journal of Epidemiology & Community Health, 61, 1042–1049. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. AUSTIN DM, FURR LA & SPINE M 2002. The effects of neighborhood conditions on perceptions of safety. Journal of criminal justice, 30, 417–427. [Google Scholar]
  3. AUSTIN PC 2011. An introduction to propensity score methods for reducing the effects of confounding in observational studies. Multivariate Behavioral Research, 46, 399–424. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. BOSMA H, DIKE VAN DE MHEEN H, BORSBOOM GJJM & MACKENBACH JP 2001. Neighborhood Socioeconomic Status and All-Cause Mortality. American Journal of Epidemiology, 153, 363–371. [DOI] [PubMed] [Google Scholar]
  5. BROWN BB, PERKINS DD & BROWN G 2004a. Crime, New housing, and housing incivilities in a first ring suburb: Multilevel relationships across time. Housing Policy Debate, 15, 301–345. [Google Scholar]
  6. BROWN BB, PERKINS DD & BROWN G 2004b. Incivilities, place attachment and crime: Block and individual effects. Journal of environmental psychology, 24, 359–371. [Google Scholar]
  7. BROWNING CR, CAGNEY KA & IVENIUK J 2012. Neighborhood stressors and cardiovascular health: crime and C-reactive protein in Dallas, USA. Soc Sci Med, 75, 1271–9. [DOI] [PubMed] [Google Scholar]
  8. CARTER SP, CARTER SL & DANNENBERG AL 2003. Zoning out crime and improving community health in Sarasota, Florida:“crime prevention through environmental design”. American Journal of Public Health, 93, 1442–1445. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. COOMBES E, JONES AP & HILLSDON M 2010. The relationship of physical activity and overweight to objectively measured green space accessibility and use. Social science & medicine, 70, 816–822. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. COZENS PM, SAVILLE G & HILLIER D 2005. Crime prevention through environmental design (CPTED): a review and modern bibliography. Property management, 23, 328–356. [Google Scholar]
  11. DIEZ ROUX AV, MERKIN SS, ARNETT D, CHAMBLESS L, MASSING M, NIETO FJ, SORLIE P, SZKLO M, TYROLER HA & WATSON RL 2001. Neighborhood of Residence and Incidence of Coronary Heart Disease. New England Journal of Medicine, 345, 99–106. [DOI] [PubMed] [Google Scholar]
  12. DUBOWITZ T, DASTIDAR MG, RICHARDSON AS, COLABIANCHI N, BECKMAN R, HUNTER GP, SLOAN JC, NUGROHO AK & COLLINS RL 2019. Results from a natural experiment: initial neighbourhood investments do not change objectively-assessed physical activity, psychological distress or perceptions of the neighbourhood. International Journal of Behavioral Nutrition and Physical Activity, 16, 29. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. DUBOWITZ T, GHOSH-DASTIDAR M, COHEN DA, BECKMAN R, STEINER ED, HUNTER GP, FLÓREZ KR, HUANG C, VAUGHAN CA & SLOAN JC 2015a. Diet and perceptions change with supermarket introduction in a food desert, but not because of supermarket use. Health Affairs, 34, 1858–1868. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. DUBOWITZ T, NCUBE C, LEUSCHNER K & THARP-GILLIAM S 2015b. A natural experiment opportunity in two low-income urban food desert communities: research design, community engagement methods, and baseline results. Health Education & Behavior, 42, 87S–96S. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. DULIN-KEITA A, CLAY O, WHITTAKER S, HANNON L, ADAMS IK, ROGERS M & GANS K 2015. The influence of HOPE VI neighborhood revitalization on neighborhood-based physical activity: A mixed-methods approach. Social Science & Medicine, 139, 90–99. [DOI] [PubMed] [Google Scholar]
  16. ELLAWAY A, MORRIS G, CURTICE J, ROBERTSON C, ALLARDICE G & ROBERTSON R 2009. Associations between health and different types of environmental incivility: a Scotland-wide study. Public health, 123, 708–713. [DOI] [PubMed] [Google Scholar]
  17. FOSTER S, GILES-CORTI B & KNUIMAN M 2010. Neighbourhood design and fear of crime: A social-ecological examination of the correlates of residents’ fear in new suburban housing developments. Health & place, 16, 1156–1165. [DOI] [PubMed] [Google Scholar]
  18. GALSTER G, TATIAN P & ACCORDINO J 2006. Targeting investments for neighborhood revitalization. Journal of the American Planning Association, 72, 457–474. [Google Scholar]
  19. GAPEN M, CROSS D, ORTIGO K, GRAHAM A, JOHNSON E, EVCES M, RESSLER KJ & BRADLEY B 2011. Perceived Neighborhood Disorder, Community Cohesion, and PTSD Symptoms among low Income African Americans in an Urban Health Detting. American Journal of Orthopsychiatry, 81, 31–37. [DOI] [PubMed] [Google Scholar]
  20. HILL TD, ROSS CE & ANGEL RJ 2005. Neighborhood Disorder, Psychophysiological Distress, and Health. Journal of Health and Social Behavior, 46, 170–186. [DOI] [PubMed] [Google Scholar]
  21. JANKE K, PROPPER C & SHIELDS MA 2016. Assaults, murders and walkers: The impact of violent crime on physical activity. Journal of health economics, 47, 34–49. [DOI] [PubMed] [Google Scholar]
  22. KELLY CM, SCHOOTMAN M, BAKER EA, BARNIDGE EK & LEMES A 2007. The association of sidewalk walkability and physical disorder with area-level race and poverty. Journal of epidemiology and community health, 61, 978–983. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. KIRK DS & LAUB JH 2010. Neighborhood change and crime in the modern metropolis. Crime and justice, 39, 441–502. [Google Scholar]
  24. LEE S & TALEN E 2014. Measuring walkability: A note on auditing methods. Journal of Urban Design, 19, 368–388. [Google Scholar]
  25. NEGRON-POBLETE P, SÉGUIN A-M & APPARICIO P 2016. Improving walkability for seniors through accessibility to food stores: a study of three areas of Greater Montreal. Journal of urbanism: International research on placemaking and urban sustainability, 9, 51–72. [Google Scholar]
  26. RIDGEWAY G, MCCAFFREY D, MORRAL AR, BURGETTE L & GRIFFIN BA 2017. Toolkit for weighting and analysis of nonequivalent groups: A tutorial for the twang package [Online]. The Comprehensive R Archive Network. Available: https://cran.r-project.org/web/packages/twang/vignettes/twang.pdf [Accessed October 17, 2019]. [Google Scholar]
  27. ROSSEN LM, POLLACK KM, CURRIERO FC, SHIELDS TM, SMART MJ, FURR-HOLDEN CDM & COOLEY-STRICKLAND M 2011. Neighborhood incivilities, perceived neighborhood safety, and walking to school among urban-dwelling children. Journal of physical activity and health, 8, 262–271. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. SAELENS BE, SALLIS JF, BLACK JB & CHEN D 2003. Neighborhood-based differences in physical activity: an environment scale evaluation. American journal of public health, 93, 1552–1558. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. SLATER SJ, NICHOLSON L, CHRIQUI J, BARKER DC, CHALOUPKA FJ & JOHNSTON LD 2013. Walkable communities and adolescent weight. American journal of preventive medicine, 44, 164–168. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. SOHN D-W 2016. Residential crimes and neighbourhood built environment: Assessing the effectiveness of crime prevention through environmental design (CPTED). Cities, 52, 86–93. [Google Scholar]
  31. STOCKDALE SE, WELLS KB, TANG L, BELIN TR, ZHANG L & SHERBOURNE CD 2007. The Importance of Social Context: Neighborhood Stressors, Stress-Buffering Mechanisms, and Alcohol, Drug, and Mental Health Disorders. Social science & medicine (1982), 65, 1867–1881. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. TALEN E & KOSCHINSKY J 2013. The walkable neighborhood: A literature review. International Journal of Sustainable Land Use and Urban Planning, 1. [Google Scholar]
  33. TATIAN PA, KINGSLEY GT, PARILLA J & PENDALL R 2012. Building successful neighborhoods. Washington, DC: The Urban Institute. [Google Scholar]
  34. THANI SKSO, HASHIM NHM & ISMAIL WHW 2016. Surveillance by design: assessment using principles of crime prevention through environmental design (CPTED) in urban parks. Procedia-social and behavioral sciences, 234, 506–514. [Google Scholar]
  35. TWINAM T 2017. Danger zone: Land use and the geography of neighborhood crime. Journal of Urban Economics, 100, 104–119. [Google Scholar]
  36. WELSH BC & FARRINGTON DP 2008. Effects of improved street lighting on crime. Campbell systematic reviews, 13, 1–51. [Google Scholar]
  37. WILLIAMS DR & COLLINS C 2001. Racial residential segregation: a fundamental cause of racial disparities in health. Public Health Reports, 116, 404–416. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES