Work and Home Neighborhood Design and Physical Activity

Abstract

Purpose:

To investigate relations of perceived worksite neighborhood environments to total physical activity and active transportation, over and above home neighborhood built environments.

Design:

Observational epidemiologic study.

Setting:

Baltimore, Maryland-Washington, DC, and Seattle-King County, Washington metropolitan areas.

Participants:

One thousand eighty-five adults (mean age = 45.0 [10.2]; 46% women) recruited from 32 neighborhoods stratified by high/low neighborhood income and walkability.

Measures:

The Neighborhood Environment Walkability Survey assessed perceptions of worksite and home neighborhood environments. Accelerometers assessed total moderate-to-vigorous physical activity (MVPA). The International Physical Activity Questionnaire assessed total active transportation and active transportation to and around work.

Analysis:

Mixed-effects regression tested relations of home and worksite neighborhood environments to each physical activity outcome, adjusted for demographics.

Results:

Home and worksite mixed land use and street connectivity had the most consistent positive associations with physical activity outcomes. Worksite traffic and pedestrian safety were also associated with multiple physical activity outcomes. The worksite neighborhood explained additional variance in physical activity outcomes than explained by the home neighborhood. Worksite and home neighborhood environments interacted in explaining active transportation to work, with the greatest impacts occurring when both neighborhoods were activity supportive.

Conclusion:

Both worksite and home neighborhood environments were independently related to total MVPA and active transportation. Community design policies should target improving the physical activity supportiveness of worksite neighborhood environments and integrating commercial and residential development.

Keywords

built environment exercise transport perceived environment neighborhood safety walkability walking

Purpose

The built environment has been associated with physical activity, active transportation, and health markers in numerous studies.^1
-3 Such research has focused primarily on built environment attributes of the home neighborhood. However, many people encounter and are likely influenced by environment attributes outside of their home neighborhoods.^4,5 Neighborhood environments around the worksite, in particular, may be important determinants of physical activity behaviors and health because the average employed American spends 7.6 h/d on most days of the week at work.⁶

A primary way that worksite neighborhood environments are likely to influence health is by supporting or inhibiting engagement in active transportation to and around work. Few studies have investigated associations between worksite neighborhood environments and physical activity,^7

-10 but similar features would be expected to be related to physical activity, as found in studies of home neighborhood environments. Features of worksite neighborhood environments that have been associated positively with physical activity included residential density,¹⁰ mixed land use,^7,8 street connectivity,^7
-9 and pedestrian and traffic safety.⁹

Limitations of existing worksite neighborhood studies were that they involved small sample sizes or samples with little variability in environments. Furthermore, little is known about how worksite and home neighborhood environments work together in explaining physical activity, or whether worksite neighborhood environments contribute a meaningful amount of additional explained variance in physical activity than home neighborhood environments alone. This information would show whether the impact of the built environment on health is underestimated by only focusing on home neighborhood environments and inform the extent to which community design policies and practices should consider the 2 locations together.

The aim of the present study was to investigate whether perceived neighborhood built environment attributes around the worksite were associated with total physical activity and active transportation, and whether they explained additional variance in these behaviors over and above perceived built environment attributes around the home.

Methods

Design

The Neighborhood Quality of Life Study (NQLS) was an observational epidemiologic study designed to examine relationships between built environments and physical activity in adults. Neighborhoods were defined as clusters of contiguous census block groups. Thirty-two neighborhoods were included from the Baltimore, Maryland-Washington, DC, and Seattle-King County, Washington metropolitan areas, evenly distributed by walkability (high/low) and income (high/low), where walkability was assessed using Geographic Information Systems (GIS)-based measures of residential density, street connectivity, retail floor area ratio, and land use mix. Details of neighborhood selection, walkability calculations, and results have been previously reported.^11,12

Sample

Participants were adults aged 20 to 65 years, recruited from households in the identified neighborhoods using marketing company lists, and contacted by phone and mail. Eligibility criteria included residing in a private residence (not a group facility), able to complete a survey in English, no major cognitive or developmental impairment, and ability to walk independently. Assessments were conducted from 2001 to 2005. The NQLS included 2199 participants, but only participants who reported working outside of the home were used in present analyses, resulting in a subsample of 1085 participants.

Measures

Demographic factors

Participant age, gender, ethnicity (non-Hispanic: white vs other), education (college degree: yes vs no), number of motor vehicles per adult in household, number of people in household, years at current address, marital status (married/cohabitating vs other), and region of residence (Seattle/King County or Baltimore/Maryland) were collected by survey.

Perceived neighborhood environment around home

The Neighborhood Environment Walkability Survey (NEWS) assessed perceptions of residential density (6-item scale), land use mix access (7-item scale), street connectivity (5-item scale), walking/cycling facilities (7-item scale), traffic safety (5-item scale), pedestrian safety (7-item scale), and crime safety (4-item scale) within a 10- to 15-minute walk from the participant’s home. With the exception of residential density (which is a weighted-sum algorithm), item responses ranged from 1 to 4 where 1 indicated strongly disagree and 4 indicated strongly agree. Scale scores consisted of a mean of the items, with some items reverse coded so that higher scale scores on all items reflected a more activity-supportive environment. The NEWS has been shown to have good reliability and validity in multiple studies.^13
-15

Perceived neighborhood environment around work

An adapted and shortened NEWS was developed for this study to assess the perceived environment around participants’ worksites. Questions were taken from the original NEWS but changed to reflect the neighborhood within a 10- to 15-minute walk from participants’ worksites rather than homes. The residential density scale was omitted. The worksite version assessed land use mix access (2-item scale), street connectivity (2-item scale), walking/cycling facilities (2-item scale), aesthetics (3-item scale), traffic safety (1 item), pedestrian safety (1 item), and crime safety (1 item).

Total moderate to vigorous physical activity

ActiGraph accelerometers (Manufacturing Technology Incorporated, models 7164 and 71256; Pensacola, Florida) were used to objectively measure participants’ total moderate to vigorous physical activity (MVPA). Accelerometers have been shown to have validity for estimating physical activity in adults.¹⁶ The epoch was set at 60 seconds, and participants were required to wear the accelerometer for at least 8 hours per day and 5 total days. A valid hour of wear time contained no more than 30 consecutive minutes of zero counts. Data were cleaned and scored using MeterPlus software version 4.0 (http://www.meterplussoftware.com). Minutes per week of MVPA were calculated using previously established cut points for adults (≥1952 counts/minute).¹⁷

Total active transportation

Self-reported walking and bicycling for transport was assessed using the 4-item scale from the International Physical Activity Questionnaire survey long version which has established reliability and validity.¹⁸ Items assessed frequency and duration of walking and bicycling for transportation within the past week, from which minutes/week of total active transportation were calculated.

Active transportation around work

Participants were asked how many days in the past month they walked from work to the following locations: (1) food store, (2) retail store, (3) school/day care center, (4) bank/credit union, (5) post office, (6) restaurant/café, (7) gym/health club/recreation facility, (8) park, and (9) public transportation/park and ride. These items were summed to represent the total number of trips per month the participant traveled from work to the aforementioned destinations. This measure was developed for the present study.

Active transportation to/from work

Participants were asked how many days in the past month they traveled to work by (1) walking or (2) biking. Only 15.1% of participants reported any walking or biking to/from work, so a dichotomous variable was created to represent any versus no active transportation to work.

Analysis

Mixed effects regression models were used to investigate the relation of home and work perceived neighborhood environment factors to physical activity, with home neighborhood entered as a random effect cluster variable (because recruitment of participants was neighborhood based). Linear models were estimated for total MVPA, total active transportation, and active transportation around work, while a logistic model was estimated for active transportation to/from work (4 models total, with all home and work environment variables entered in each model). All models adjusted for the 9 demographic factors listed in the measures section, with continuous covariates mean-centered and dichotomous covariates centered on zero (ie, 0.5 vs −0.5). The total active transportation and active transportation around work outcomes were natural log transformed to better approximate normal distributions, and regression coefficients from the respective models were transformed using the following formula so they could be interpreted in their original units: (Exp(coefficient)−1) * intercept. The perceived neighborhood environment scales were normalized to have a mean of zero and a standard deviation of 1 (ie, z-score) before being entered into the models so magnitude of associations could be compared across the scales. Total percentage of variance explained by the (a) home environment variables (not adjusted for the work variables) and (b) the home and work neighborhood environment variables was estimated as (1−(intercept + residual from full model)/(intercept + residual from base model)) * 100 from the covariance estimates, where the base model included only the aforementioned demographic covariates and the neighborhood clustering effect.^19,20 In accordance with ecological models,²⁰ interactions between the home and work perceived neighborhood environment in relation to participants’ physical activity were tested as (sum of scale z-scores for home) * (sum of scale z-scores for work). Interactions with a P value <.150 were plotted to minimize type II error. SPSS version 20 was used for the analyses.

Results

Participant characteristics and descriptive statistics for all study variables are presented in Table 1, and the relation of home and work perceived neighborhood environment characteristics to participants’ physical activity is presented in Table 2. The proportion of variation in the physical activity outcomes attributable to within- versus between-neighborhood effects (ie, Intraclass Correlation Coefficient (ICC) from empty model * 100) was 9% for total MVPA, 11% for total active transportation, 9% for active transportation around work, and 17% for active transportation to/from work.

Table 1.

Sample Characteristics and Descriptive Statistics for Study Variables.

Study Sample (N)	1085
Sociodemographic covariates
Mean age (SD)	45.0 (10.2)
Percent women	46.0
Percent White non-Hispanic	76.0
Percent with college degree	68.9
Percent married or cohabitating	63.3
Percent from the Seattle area	57.7
Mean (SD) number of people in household	2.6 (1.4)
Mean (SD) number of vehicles/adult	1.0 (0.5)
Mean (SD) years living at current address	9.4 (8.6)
Perceived neighborhood environment independent variables^a
Mean (SD) residential density around home	224.8 (70.6)
Mean (SD) land use mix access around home	2.9 (0.6)
Mean (SD) street connectivity around home	2.8 (0.6)
Mean (SD) walking/cycling facilities around home	2.9 (0.7)
Mean (SD) aesthetics around home	3.1 (0.6)
Mean (SD) traffic safety around home	3.0 (0.7)
Mean (SD) pedestrian safety around home	2.9 (0.5)
Mean (SD) crime safety around home	3.4 (0.6)
Mean (SD) residential density around work	3.2 (0.9)
Mean (SD) street connectivity around work	3.3 (0.8)
Mean (SD) walking/cycling facilities around work	2.9 (0.8)
Mean (SD) aesthetics around work	3.0 (0.7)
Mean (SD) traffic safety around work	2.6 (0.9)
Mean (SD) pedestrian safety around work	3.2 (0.9)
Mean (SD) crime safety around work	3.1 (0.9)
Physical activity outcome variables
Mean (SD) total MVPA minutes/week	232.7 (163.6)
Mean (SD) total active transportation minutes/week^b	19.7 (10.7)
Mean (SD) active transportation around work trips/month^b	3.4 (4.2)
Percent with any active transportation to/from work	15.1

Abbreviations: MVPA, moderate-to-vigorous physical activity; SD, standard deviation.

^aRanged from 1 to 4, with the exception of residential density; higher numbers indicated greater walkablity or safety.

^bGeometric rather than arithmetic means and standard deviations are reported because the distributions were skewed.

Table 2.

Relation of Home and Work Perceived Neighborhood Environment Characteristics to Physical Activity in Working Adults (N = 1048-1071).

Characteristics	Total Accelerometer MVPA (minutes/week)			Total Active Transportation (minutes/week)^a			Active Transportation Around Work (trips/month)^a			Active Transportation To/From Work (any vs none)
Characteristics	B	95% CI	P	B	95% CI	P	B	95% CI	P	OR	95% CI	P
Intercept	220.7	207.0, 234.4	–	18.6	15.3, 22.6	–	3.6	3.1, 4.1	–	0.07	0.04, 0.13	–
Home neighborhood environment (z-scores)^b,c
Residential density	6.4	−4.8, 17.6	.265	5.6^d	1.7, 11.9	<.001	–	–	–	0.99	0.80, 1.23	.932
Land use mix access	17.5^e	5.7, 29.4	.004	7.7^d	3.0, 15.2	<.001	–	–	–	1.42^e	1.07, 1.88	.014
Street connectivity	10.3	−1.7, 22.3	.093	5.0^e	1.1, 11.5	.006	–	–	–	1.40^e	1.06, 1.84	.017
Walking/cycling facilities	5.0	−7.4, 17.5	.428	−1.0	−3.2, 3.0	.547	–	–	–	0.85	0.64, 1.12	.245
Aesthetics	3.7	−8.7, 16.0	.561	0	−2.4, 4.4	.981	–	–	–	1.12	0.85, 1.48	.413
Traffic safety	2.3	−9.3, 13.8	.697	0.8	−1.8, 5.2	.612	–	–	–	0.92	0.72, 1.19	.540
Pedestrian safety	−8.4	−20.5, 3.7	.175	1.6	−1.3, 6.6	.341	–	–	–	0.89	0.69, 1.15	.365
Crime safety	7.7	−4.4, 19.8	.210	−1.2	−3.2, 2.5	.453	–	–	–	0.99	0.76, 1.29	.940
Total variance explained by home neighborhood environment^f		2.88%			9.33%			–			9.29%
Work neighborhood environment (z-scores)^b,c
Land use mix access	−4.3	−15.1, 6.5	.432	5.8^d	1.9, 11.9	<.001	1.5^d	0.9, 2.9	<.001	1.58^d	1.20, 2.08	<.001
Street connectivity	3.3	−7.4, 13.9	.547	1.2	−1.3, 5.3	.433	0.8^d	0.3, 1.3	<.001	0.94	0.74, 1.19	.597
Walking/cycling facilities	1.0	−9.9, 11.9	.858	2.0	−0.8, 6.6	.212	0.2	−0.1, 0.7	.155	1.34^e	1.06, 1.71	.017
Aesthetics	−7.5	−19.1, 4.2	.208	1.9	−1.0, 6.6	.252	0.2	−0.1, 0.6	.350	1.08	0.84, 1.39	.537
Traffic safety	−4.0	−14.0, 6.0	.430	1.4	−1.0, 5.4	.296	0.3^e	0, 0.8	.038	1.26^e	1.02, 1.55	.032
Pedestrian safety	12.9^e	1.8, 24.0	.023	1.2	−1.4, 5.6	.431	0.5^e	0.1, 1.1	.003	0.93	0.73, 1.19	.566
Crime safety	10.9^e	0.2, 21.6	.045	−2.9^e	−4.2, −0.4	.028	−0.3^e	−0.5, 0	.038	0.81	0.65, 1.01	.058
Total variance explained by home and work neighborhood environment^f		7.11%			12.58%			16.88%			10.81%

Abbreviations: B, unstandardized coefficient; CI, confidence interval; MVPA, moderate-to-vigorous physical activity; OR, odds ratio.

^aOutcome variables were natural log transformed to better approximate normal distributions and B was calculated as (Exp(coefficient)-1)*intercept.

^bAnalyses adjusted for city, age, gender, education, ethnicity, vehicles per adult, marital status, people per household, time at address, and clustering of participants within block groups.

^cFor all perceived environment measures, higher numbers indicated greater walkability or safety.

^d P < .01.

^e P < .05.

^fOver and above the covariates included in the base model.

Residential density around home was associated positively with total active transportation (B = 5.6 min/wk). Land use mix access around home was associated positively with total MVPA (B = 17.5 min/wk), total active transportation (B = 7.7 min/wk), and active transportation to/from work (odds ratio [OR] = 1.42). Street connectivity around home was associated positively with total active transportation (B = 5.0 min/wk) and active transportation to/from work (OR = 1.40).

In the models investigating the neighborhood environment around work, 5 of the 7 environment characteristics were associated with active transportation around work, 3 of the 7 were associated with active transportation to/from work, and 2 of 7 were associated with total active transportation and accelerometer-measured total MVPA. Land use mix access around work was associated positively with 3 of the 4 physical activity outcomes: total active transportation (B = 5.8 min/wk), active transportation around work (B = 1.5 trips/month), and active transportation to/from work (OR = 1.58). Street connectivity around work was associated positively with active transportation around work (B = 0.8 trips/month). Walking/cycling facilities around work were associated positively with active transportation to/from work (OR = 1.34). Traffic safety around work was associated positively with active transportation around work (B = 0.3 trips/month) and active transportation to/from work (OR = 1.26). Pedestrian safety around work was associated positively with total MVPA (B = 12.9 min/wk) and active transportation around work (B = 0.5 trips/week). Crime safety around work was associated positively with total MVPA (B = 10.9 min/wk) and associated negatively with total active transportation (B = −2.9 min/wk) and active transportation to/from work (B = −0.3 trips/month).

Total variance explained by the home neighborhood environment attributes (over and above the 9 covariates) ranged from 2.88% to 9.29%, whereas after adding the work neighborhood environment attributes, total variance explained ranged from 7.11% to 16.88%. Thus, total variance explained was increased in each model when work neighborhood environment attributes were added on top of the home neighborhood environment attributes and covariates (increases ranged from 1.52% to 4.23%). Work neighborhood environment attributes explained 16.88% of the variance in active transportation around work.

One of the 3 interactions between the home and work neighborhood environment tested had a P value <.150. The relationship between the work neighborhood environment and active transportation to/from work was stronger when the home neighborhood environment was more supportive of physical activity, with the odds of active transportation to/from work being highest when both the home and work neighborhood environments were supportive (see Figure 1; interaction t = 1.57; P = .117).

Figure 1.

Interaction between the home and work perceived neighborhood environment explaining active transportation to/from work (P = .117).

Discussion

Findings from the present study indicate both work and home environments contribute to the explanation of multiple physical activity outcomes among adults who work outside the home. Neighborhood built environment attributes around the worksite explained participants’ overall MVPA and active transportation, particularly active transportation to/from and around work. The worksite neighborhood explained additional variance in these physical activity behaviors than explained by the home neighborhood alone, suggesting that research focusing on only the home neighborhood may underestimate the impacts of built environments on health outcomes. Community design and zoning policies affecting worksite neighborhoods should consider similar attributes as are now being more commonly incorporated when designing residential neighborhoods due to the accumulation of research^1
-3 and home buyer preferences: increased residential density, mixed land use, street connectivity, and pedestrian and cycling infrastructure.

Mixed land use was consistently associated with multiple physical activity behaviors. Home neighborhood mixed land use was associated positively with total MVPA, total active transportation, and active transportation to/from work. Worksite neighborhood mixed land use was associated with total active transportation, active transportation to/from work, and active transportation around work. These findings are in agreement with previous studies.^7,8 Mixed land use in the worksite neighborhood appears particularly important for active transportation around work, as indicated by the large association (a 1 standard deviation [SD] increase above the mean for mixed land use corresponded to a 42% increase in active trips around work). Having shops and services within close walking distance to work allows errands and other activities to be performed midday or before and after work and enables workers to reduce car use in meeting daily needs.²¹

Street connectivity around the home appeared important for total active transportation and active transportation to work, whereas street connectivity around the worksite was important for active transportation around work. In previous studies, worksite neighborhood street connectivity was associated with active transportation to work⁷ and active transportation around work.⁹ Street connectivity likely plays a particular role in active transportation when there is mixed land use in the worksite neighborhood or when the home neighborhood also has connected streets and is within walking or biking distance from the worksite. It is likely easier for workers to choose to walk, bike, or use transit when a worksite is located in close proximity and with good transportation connections to restaurants, shops, fitness facilities, or other daily needs, as opposed to being isolated and largely surrounded by parking and major roadways.

Although traffic and pedestrian safety in the home neighborhood were not associated with the physical activity behaviors assessed, perceived traffic and pedestrian safety in the worksite were associated positively with total MVPA (pedestrian safety), active transportation around work (traffic and pedestrian safety), and active transportation to work (traffic safety). Incorporating safe street crossings and traffic calming into worksite community design to support pedestrian safety and active travel could be accomplished by adopting and implementing complete streets policies.²²

Neighborhood aesthetics was not associated with any of the physical activity behaviors in the present study. In other studies, aesthetics has been more consistently associated with physical activity performed for leisure rather than transportation,³ likely because the latter is often performed out of necessity. Perceived safety from crime was unexpectedly negatively associated with total active transportation and active transportation around work but was positively associated with total MVPA. Findings regarding crime and perceived crime in the home neighborhood as a correlate of physical activity among adults have been inconsistent across multiple studies,²³ so further research and improved measurement are needed.

A major finding was that the worksite neighborhood environment explained additional variance in physical activity than explained by the home neighborhood environment alone. Some associations, such as for pedestrian and traffic safety, appeared to be specific to the worksite neighborhood. The interaction between the worksite and home neighborhood combined attributes suggests that active transportation to work requires both a supportive worksite and supportive home neighborhood environment. Specifically, the odds of active travel to work more than doubled when both neighborhoods were supportive (1 SD above mean) than when both were unsupportive (1 SD below the mean). The interaction results are an indicator that previous studies that only considered home environments underestimated the contribution of built environments to explaining physical activity. Perhaps including additional environmental settings, such as routes traveled or frequent recreation destinations could help clarify the contributions of built environments to physical activity.

Strengths and Limitations

The built environment measures relied on participants’ perceptions of their neighborhood and may not correspond to objective measures of the same environments. However, perceived neighborhood environment characteristics have been important determinants of physical activity, independent of objective characteristics, in previous studies.²⁴ The cross-sectional nature of this study precludes ability to determine causality between worksite neighborhood-built environments and physical activity. Strengths included assessment of objective physical activity and multiple measures of active transportation, use of a perceived worksite neighborhood environment tool that was adapted from the widely used NEWS,^13
-15 and a stratified sampling plan deployed in 2 contrasting east and west coast regions that maximized variation across income and home neighborhood environments.^11,12

Conclusions

Findings showed significant associations between several attributes of worksite neighborhood environments and physical activity when adjusting for home neighborhood environment, demographic, and socioeconomic factors. It is notable that activity-supportive worksite neighborhoods appeared to facilitate active transportation both to/from work and around the worksite neighborhood. The important role of worksite neighborhood environments documented in the present study demonstrates the benefits of incorporating worksites, and possibly additional environments, in studies examining the role of built and social environments in explaining physical activity and other health outcomes. Improving the supportiveness of multiple environments is likely to lead to greater increases in active travel and total physical activity than improving supportiveness in only one environment.

So What?

Changes to worksite neighborhoods to make them more walkable and accessible by modes other than private vehicles may provide a much needed opportunity to promote physical activity. Location of worksites within walking distance of shops and services, and further promoting active transport with connected streets, safe street crossings, and traffic calming could support increases in active travel and total physical activity. The introduction of residential development into commercial areas where it does not already exist has considerable promise as a way of promoting active living and reduced car dependence through short walkable commutes for those that work nearby.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Funding was provided by NIH grant R01HL067350.

References

Ding

Gebel

. Built environment, physical activity, and obesity: what have we learned from reviewing the literature? Health Place. 2012;18(1):100–105.

Bauman

Reis

Sallis

Wells

Loos

Martin

. Correlates of physical activity: why are some people physically active and others not? Lancet. 2012;380(9838):258–271.

Saelens

Handy

. Built environment correlates of walking: a review. Med Sci Sports Exerc. 2008;40(suppl 7):S550–S566.

Hurvitz

Moudon

. Home versus nonhome neighborhood: quantifying differences in exposure to the built environment. Am J Prev Med. 2012;42(4):411–417.

Burgoine

Monsivais

. Characterising food environment exposure at home, at work, and along commuting journeys using data on adults in the UK. Int J Behav Nutr Phys Act. 2013;10:85.

US Bureau of Labor Statistics. American Time Use Survey (2015) Washington, DC: US Department of Labor; [July 12, 2017]. https://www.bls.gov/tus/charts/work.htm. Accessed April 2, 2018.

Dalton

Jones

Panter

Ogilvie

. Neighbourhood, route and workplace-related environmental characteristics predict adults’ mode of travel to work. PLoS One. 2013;8(6):e67575.

Frank

Bradley

Kavage

Chapman

Lawton

. Urban form, travel time, and cost relationships with tour complexity and mode choice. Transportation. 2008;35(1):37–54.

Schwartz

Aytur

Evenson

Rodriguez

. Are perceptions about worksite neighborhoods and policies associated with walking? Am J Health Promot. 2009;24(2):146–51.

10.

Troped

Wilson

Matthews

Cromley

Melly

. The built environment and location-based physical activity. Am J Prev Med. 2010;38(4):429–438.

11.

Frank

Sallis

Saelens

. The development of a walkability index: application to the neighborhood quality of life study. Br J Sports Med. 2010;44(13):924–933.

12.

Sallis

Saelens

Frank

. Neighborhood built environment and income: examining multiple health outcomes. Soc Sci Med. 2009;68(7):1285–1293.

13.

Cerin

Leslie

Owen

Bauman

. An Australian version of the neighborhood environment walkability scale: validity evidence. Meas Phys Educ Exerc Sci. 2008;12(1):31–51.

14.

De Bourdeaudhuij

Sallis

Saelens

. Environmental correlates of physical activity in a sample of Belgian adults. Am J Health Promot. 2003;18(1):83–92.

15.

Saelens

Sallis

Black

Chen

. Neighborhood-based differences in physical activity: an environment scale evaluation. Am J Public Health. 2003;93(9):1552–1558.

16.

Welk

. Use of accelerometry-based activity monitors to assess physical activity. Phys Activity Assess Health Relat Res. (Ed. Welk

). Human Kinetics Publishers, Inc; 2002:125–141.

17.

Freedson

Melanson

Sirard

. Calibration of the computer science and applications, Inc. accelerometer. Med Sci Sports Exerc. 1998;30(5):777–781.

18.

Craig

Marshall

Sjöström

. International Physical Activity Questionnaire: 12-country reliability and validity. Med Sci Sports Exerc. 2003;35(8):1381–1395.

19.

Snijders

TAB

Bosker

. Multilevel Analysis: An Introduction to Basic and Advanced Multilevel Modeling. 2nd ed. Thousand Oaks, CA: Sage Publications, Inc; 2012.

20.

Sallis

Owen

Fisher

. Ecological models of health behavior. In: Glanz

Rimer

Viswanath

, eds. Health Behavior: Theory, Research, And Practice. 5th ed. San Francisco, CA: Jossey Bass; 2015:43–64.

21.

Cervero

. Built environments and mode choice: toward a normative framework. Transp Res D Trans Environ. 2002;7(4):265–284.

22.

Smart Growth America. National Complete Streets Coalition. Washington, DC: Smart Growth America; [July 12, 2017]. https://smartgrowthamerica.org/program/national-complete-streets-coalition/. Accessed April 2, 2018.

23.

Foster

Giles-Corti

. The built environment, neighborhood crime and constrained physical activity: an exploration of inconsistent findings. Prev Med. 2008;47(3):241–251.

24.

Gebel

Bauman

Sugiyama

Owen

. Mismatch between perceived and objectively assessed neighborhood walkability attributes: prospective relationships with walking and weight gain. Health Place. 2011;17(2):519–524.