Does the Built Environment Influence Physical Activity? Examining the Evidence -- Special Report 282 (2005) / Chapter Skim
Currently Skimming:
6 Current State of Knowledge
6 Current State of Knowledge
Pages 151-218
The Chapter Skim interface presents what we've algorithmically identified as the most significant single chunk of text within every page in the chapter.
Select key terms on the right to highlight them within pages of the chapter.
From page 151... ...
. In this chapter, the empirical evidence on the relationship be tween the built environment and physical activity is reviewed.
|
From page 152... ...
International studies were included, although the committee recognizes that the social and environ mental determinants of physically active behavior may not be fully comparable with nor the results transferable to the situation in the United States. The committee acknowledges the contribution of international scholars and the importance of international collab oration on research linking the built environment and physical activity.
|
From page 153... ...
. Measures of the built environment include population and employment density, 3 Part of the review that deals primarily with work site interventions (e.g., industrial plants, universi ties, and federal agencies)
|
From page 154... ...
, street pattern (e.g., grid, cul-de-sac) and connectiv ity, presence of bicycle paths, neighborhood type (e.g., traditional versus suburban planned unit development)
|
From page 155... ...
Most of the studies also control for more typical socioeconomic variables, such as age, gender, race, ethnicity, educational level, marital status, employ ment status, and income level. With regard to effects of the built environment, the study results reveal that a few measures are significantly correlated with physi cal activity (see Handy 2004, Table 3-7, and Table 6-2, pp.
|
From page 156... ...
was positively correlated with total physical activity levels. The characteristic of land use density is a good example of the complexities involved in linking the design of the built environment to travel behavior, such as walking in the neighborhood or walk ing to access transit.
|
From page 157... ...
In the physical activity literature, both perceived and objective measures of proximity and convenience of facilities, ranging from exercise equipment at home, to bicycle paths and trails, to parks, to local shopping and transit stops, are significantly and positively correlated with walking, other forms of exercise and recreation, and total physical activity. The importance of good access to and convenience of facilities and destinations in the decision to be physically active is certainly plausible from a theoretical perspective.
|
From page 158... ...
These limited findings about the importance of design could reflect either the small number of studies that examined these vari ables, particularly in the travel behavior literature, or poorly spec ified measures of design. They could also signal the lack of a significant relationship between design and physical activity, or a relationship that may depend on the particular type of physical ac tivity involved.
|
From page 159... ...
concludes, and the committee concurs, that a supportive built environment alone is not sufficient to influence physical ac tivity; nevertheless, it can play a facilitating role. ANALYSIS OF FINDINGS Effects of the Built Environment on Different Socioeconomic Groups Data on physical activity levels of the adult population from the large public health surveys discussed in Chapter 2 indicate that activity lev els decrease with age and are lower among women, ethnic and racial minorities, those with less education and low income levels, the dis abled, and those living in the southeastern region of the United States (CDC 2003)
|
From page 160... ...
Of the 28 studies in this literature reviewed by Handy (2004) , only nine describe the relationship between the built environment and physical activity separately for some demographic characteris tic.
|
From page 161... ...
Stair use provides a low-cost way to integrate physical activity into the daily routine, and there is some limited evidence that interventions to increase workplace stair use (e.g., motiva tional signs and music in the stairwell) can be effective, although the duration of the effect is unclear (Kerr et al.
|
From page 162... ...
. The attributes of the neighborhood built environment that may affect walking and other forms of physical activity have already been discussed.
|
From page 163... ...
, and unattended dogs. These distinctions are important because each source of danger is related to different safety concerns, which may in turn determine which characteristics of the built environment act to inhibit or encourage physically active behavior.
|
From page 164... ...
. Likewise, parental concerns about safety curtail children's activity levels, from use of public spaces such as parks and other play spaces (Valentine and McKendrick 1997; Sallis, McKenzie, et al.
|
From page 165... ...
, characteristics of the built environment and the transportation infrastructure are part of the story. Most pedestrianautomobile collisions involving children happen in re sidential areas near a child's home (Sharples et al.
|
From page 166... ...
Raising ex posure levels can increase the risk of injury from traffic unless mit igation measures are taken.7 Time The role of time in total physical activity levels is potentially im portant but poorly understood. Lack of time is often cited as a reason for not being more physically active, and modern life is time-pressured for many.
|
From page 167... ...
Evidence for Causality The correlation between certain characteristics of the built envi ronment and higher levels of physical activity does not prove that the built environment caused the physical activity. As noted in Chapter 5, for example, the issue of self-selection bias must be ad dressed: individual preferences for being physically active may de termine the decision to live in a walking- and cycling-friendly environment and account for some or all of the higher levels of physical activity often observed in these neighborhoods.
|
From page 168... ...
In both cases, the researchers found that when variables associated with residential location preferences were identified and examined separately from variables associated with the built environment, some, but not all, of the environmental variables ceased to be significantly correlated with nonmotorized travel. For example, Greenwald and Boarnet (2001)
|
From page 169... ...
Regression models were used to predict changes in travel behavior as a function of changes in neighborhood accessibility, while changes in life cycle and regional and workplace accessibility were held constant. Krizek found that residents who moved to 10The researchers estimated a nine-equation model for residential location (traditional or subur ban)
|
From page 170... ...
Another longitudinal study of the effect of changes in the built environment on nonmotorized travel -- a natural-experimental study of the impact of improvements in traffic safety on children walking and cycling to school -- was summarized in Chapter 4. The authors (Boarnet et al.
|
From page 171... ...
Together, the few studies reviewed here provide limited but provocative results concerning the complexity of causal connec tions between the built environment and physical activity levels. 13The researchers examined the total amount of physical activity, the location of that activity (i.e., at home, in the neighborhood, outside the neighborhood)
|
From page 172... ...
KNOWLEDGE GAPS In its assessment of the empirical evidence to date, the committee identified several areas in which it believes knowledge gaps cur rently exist. The following discussion of such gaps is not exhaus tive, but it covers many of the critical areas in which further research would help clarify the complex relationships between the built environment and physical activity.
|
From page 173... ...
Stratification Relationships among built environments, policy interventions, and physical activity outcomes likely vary by subpopulation, urban setting, climate, and other contextual factors. Future research needs to use study designs and populations suitable for examining dif ferences in various subgroupings.
|
From page 174... ...
2001 4,214 parents at 1996 recall Percent walking as 51 selected in- survey distrib- usual mode to fant schools in uted through school two regions in schools the United Kingdom Cervero 1996c 42,200 housing 1985 American Choice of walk or units in 11 met- Housing bike as principal ropolitan statisti- Survey, ques- commute mode cal areas; trips tionnaire on as unit of commuting, analysis cross sectional survey Cervero and 7,889 trips, trips as 2000 Bay Area Choice of walking Duncan 2003 unit of analysis Travel Survey, or biking (with 2-day activity variables for diary survey, weekend trip, cross- recreation/ sectional entertainment, eating/meal, social, shopping purposes)
|
From page 175... ...
Full-time homemaker Distance to school Percent walking: Only one car -distance (<0.5 mile -- 89.5% Southern county walk, 0.3% bike; 0.5 to 1 mile -- 66.4% walk, 1.2% bike; 1.1 to 2 miles -- 27.7% walk, 2.0% bike; >2 miles- 5.5% walk, 0.8% bike) (not statistically tested)
|
From page 176... ...
Summary of Existing Research- Travel Behavior Literature Active Travel Study Samplinga Survey Variable Cervero and 26 neighborhoods, 1990 U.S. Number of walk Gorham 1995c 14 pairs in San Census, trips to work Francisco Bay cross- Percent walk trips Area, 12 pairs in sectional to work Los Angeles region Cervero and 620 households for 1994 recall mail Choice of nonauto Radisch 1996c nonwork survey, surveys, one mode for non 840 households for work trips, work trips for work survey one for non- Choice of nonauto in six census work trips, mode for work tracts in two cross- trips neighborhoods sectional in East Bay in San Francisco Bay Area EPA 2003 709 trips to K-12 2001 Gainesville Choice of walking school, trips as Metropolitan to school unit of analysis Transportation Choice of biking to Planning school Organization Survey and 2000 Florida Department
|
From page 177... ...
Percent walk trips: +transit neighborhoods (1.2 to 13.4 percentage points more walk trips in San Francisco Bay Area, from 1.7 to 24.6 percentage points more walk trips in Los Angeles region) Household size Pedestrian versus Choice of nonauto mode for Vehicles per automobile neigh- nonwork trips: household borhood +traditional Annual salary of Choice of nonauto mode for respondent work trips: not significant (logit model)
|
From page 178... ...
1994 Palm Percent walk or from six com- Beach bike trips of all munities in Palm County, trips Beach County, Florida, Travel Florida Survey, 2-day travel diary survey, cross sectional Frank and Pivo 1,680 households, 1989 Puget Percent walk trips 1994c weighted to re- Sound Trans- for work trips for gional total in portation census tract Puget Sound, Panel survey, Percent walk trips Washington, re- 2-day travel for shopping gion; census diary survey, trips for census tract as unit of cross- tract analysis sectional Friedman et al. Selected zones 1980 Bay Area Average number 1994c from 550 zones Travel Survey, of walk trips per in San Francisco 1-day travel day per house Bay Area diary survey, hold cross- Average number sectional of bicycle trips
|
From page 179... ...
density at origin +land use mix at origin Driver's license (share Gross employment +land use mix at destination for tract) density at Percent walk trips for shopping: Trips made by destination +employment density at trip employed resident Land use mix at origin destination (share of trip ends (entropy measure)
|
From page 180... ...
Summary of Existing Research- Travel Behavior Literature Active Travel Study Samplinga Survey Variable per day per household Percent walk trips for zone Percent bike trips for zone (all by purpose) Greenwald and 1,091 residents 1994 Portland Number of walk Boarnet 2001 from Portland, Travel Survey, trips in 2 days Oregon, region 2-day travel diary survey, cross sectional Handy 1996ac 400 residents in 1992 recall Number of strolling four neighbor- phone survey, trips per month hoods in San cross- Percent of resi Francisco Bay sectional dents strolling at Area least once per month Number of walking trips to commer cial area per month Percent of resi dents walking to commercial area at least once per month Handy et al.
|
From page 181... ...
Age Population density in Number of walk trips: Gender block group +population density Race Population density in +retail density Income zip code +percent of network that is a Square of income Retail density in grid Number of children 1-mile grid cell +pedestrian environment factor under 16 Retail density in zip +median walk distance Cars per driver code +median walk speed (ordered Employees per Percent of network probit model) household that is a grid Workday Pedestrian environ ment factor (three point scale)
|
From page 182... ...
Summary of Existing Research- Travel Behavior Literature Active Travel Study Samplinga Survey Variable cial area per month Hanson and 278 households 1971 Uppsala Percent of all stops Schwab 1987 stratified by life 35-day travel by nonmotorized cycle stage in diary survey, modes Uppsala, cross- Percent of work Sweden sectional stops by non motorized modes Kitamura et al. 1,380 individuals in 1992 three-day Number of non 1997c five neighbor- travel diary motorized trips hoods in San survey Percent nonmotor Francisco Bay ized trips for all Area trips
|
From page 183... ...
weighted by distance, using Euclidean distance) Age Study area Number of nonmotorized trips: Gender Macro-scale descrip- +North San Francisco neighbor Education level tors (y/n)
|
From page 184... ...
Summary of Existing Research- Travel Behavior Literature Active Travel Study Samplinga Survey Variable Kockelman 1997c 9,000 households; 1990 Bay Area Choice of walk or trips as unit of Travel Survey, bike for all trips analysis 1-day travel by adults diary survey, cross sectional Krizek 2000 550 households 1989 and 1997 Percent of trips by that moved Puget Sound alternative mode between 1989 Transportation (transit, walk, and 1997 in Panel, 2-day bike) Puget Sound, travel diary Change in percent Washington, survey, longi- of trips by alter region tudinal native mode (transit, walk, bike)
|
From page 185... ...
Employment status 30 minutes by walk Professional job mode) in origin zone, destination zone Land use balance (en tropy index, six land use types)
|
From page 186... ...
Summary of Existing Research- Travel Behavior Literature Active Travel Study Samplinga Survey Variable Krizek 2003 550 households 1989 and 1997 Percent of trips by that moved be- Puget Sound walking tween 1989 and Transportation 1997 in Puget Panel, 2-day Sound, Wash- travel diary ington, region survey, longi tudinal McCormack et al. 663 households 1989 Puget Percent walk trips 2001 from throughout Sound Trans- for shopping region, split into portation trips for neigh three zones; 300 Panel, 2-day borhood households in travel diary Percent walk trips each of three survey, 1992 for all trips for mixed land use same survey neighborhood neighborhoods, implemented (only walk trips neighborhood as in three longer than unit of analysis selected 5 minutes)
|
From page 187... ...
, average block area; mea sured for 150-m grid cells, averaged over all grid cells within 0.4 km Regional accessibility (gravity measure) None Straight-line distance Percent walk trips for shopping to nearest commer- trips: cial street -distance to nearest commer Neighborhood type cial street Percent walk trips for all trips: +walkable neighborhood (17.7 to 18.1 versus 2.0 to 2.8)
|
From page 188... ...
bBuilt environment variables are objectively measured unless otherwise noted. cIncluded in Saelens, Sallis, and Frank 2003.
|
From page 189... ...
None Pedestrian environ- Percent walk or bike trips: ment factor (3-point +PEF (from 1.4% in low PEF to scale) : ease of 9.6% in high PEF to 18.6% in street crossing, central business district)
|
From page 190... ...
2001c 3,392 adults in 1996 Physical Walking versus Australia Activity Sur- not walking for vey for state exercise in last of New South 2 weeks Wales, cross sectional survey Berrigan and 14,827 adults NHANES III, Walk 1 or more Troiano 2002 20 years old or cross- miles 20 or older in United sectional more times per States survey month (y/n) Leisure-time physi cal activity other than walking 20 or more times per month (y/n)
|
From page 191... ...
activity limitation Leisure time physical activity: Region not significant (logistic regression) Age Exercise equipment at Active: Gender home (y/n)
|
From page 192... ...
Summary of Existing Research- Physical Activity Literature Physical Activity Study Sampling Survey Variable Brownson et al. 1,818 adults, 19992000 Meeting recom 2001 United States, cross- mendations for modified BRFSS sectional moderate or sampling plan, phone survey, vigorous activity oversampling of questions (y/n)
|
From page 193... ...
: no safe place, bad weather Gender Perception of safety from Active: Race/ethnicity crime in neighborhood +perceived safe from crime in Education level (4-point scale) neighborhood Income Income Observations of 18 char- Percent walking to work: University acteristics on 10-point +ecologic score (1-unit increase education scales; hierarchical lin- in score associated with Poverty ear modeling to create 25 percentage point increase Degree of urban- ecologic score for each in walking)
|
From page 194... ...
Summary of Existing Research- Physical Activity Literature Physical Activity Study Sampling Survey Variable Eyler et al. 2003 4,122 women 20012002 Meets recommen 20 to 50 years Women and dations for old from diverse Physical Activ- moderate or vig racial/ethnic ity Survey, orous activity groups (white, cross- versus does not African Ameri- sectional meet can, Latina, phone and Does any physical and Native interview activity versus American)
|
From page 195... ...
, conve- +access to local shopping nience of physical +emotional satisfaction with activity facilities (18) neighborhood Vigorous activity: +physical activity equipment in home +convenience of physical activity facilities +work site environment Racial/ethnic group Traffic (3-point scale)
|
From page 196... ...
Summary of Existing Research- Physical Activity Literature Physical Activity Study Sampling Survey Variable in suburb for Walking at recom less than 1 year, mended levels exercised as (y/n, based on recommended 12 or more ses at work, medical sions in 2 weeks condition likely totaling 360 min to affect physi- utes or more) cal activity, not proficient in English Giles-Corti and 1,803 adults 18 to 19951996 Exercising as rec Donovan 2002bc 59 years old in cross- ommended (y/n, Perth, Australia sectional in- based on walk person survey ing for recre ation and transportation, light-moderate physical activity, vigorous physi cal activity)
|
From page 197... ...
Household income extent of tree coverage +high physical environment Education level Access to built facilities score relative to low (43% (gravity measures by more likely) quartile, from GIS)
|
From page 198... ...
Summary of Existing Research- Physical Activity Literature Physical Activity Study Sampling Survey Variable Giles-Corti and 1,803 adults 18 to 19951996 Walking at recom Donovan 2003 59 years old in cross- mended levels Perth, Australia sectional in- (y/n, based on person survey 12 or more ses sions in 2 weeks totaling 360 min utes or more) Hovell et al.
|
From page 199... ...
, score (2.13 times as likely as Household income extent of tree coverage low score) Education level Access to built facilities +high access to attractive open (gravity measures by space (1.47 times as likely as quartile, from GIS)
|
From page 200... ...
Summary of Existing Research- Physical Activity Literature Physical Activity Study Sampling Survey Variable MacDougall et al. 1,765 adults in 1987 cross- Moderately active 1997c Adelaide, sectional mail versus inactive Australia survey by the (based on mod South erate activity, Australia vigorous sport, Community walking for Health Re- exercise)
|
From page 201... ...
: no Meets for suburban: safe place, bad +walking/jogging trails weather +indoor gym Meets for rural: +indoor gym +4 places to exercise None Some place to walk (y/n) : Meeting recommendations: Not home based: public +public park park, school track, gym +school track or fitness center, walk- +gym or fitness center ing or jogging trail, shop- +walking or jogging trail ping mall, other place +other place Home based: neighbor- +neighborhood streets or roads hood streets or roads, +neighborhood sidewalk neighborhood side walk, treadmill at home Age City of Chicago versus Walking: Gender suburb of Chicago ver- +Chicago versus small town or Race sus small city versus rural area Ethnicity small town or rural Marital status area Education Household income Below poverty line Neighborhood poverty, race, ethnicity, and education characteristics (continued )
|
From page 202... ...
Summary of Existing Research- Physical Activity Literature Physical Activity Study Sampling Survey Variable Rutten et al. 2001 3,343 adults, six MAREPS study, Level of vigorous European coun- 19971998 activity (seden tries (Belgium, cross- tary, not/some Finland, sectional what vigorous, Germany, phone survey vigorous, very Netherlands, vigorous)
|
From page 203... ...
Age High-walkability versus Moderate-intensity: Education level low-walkability +high-walkability (194.8 versus neighborhood 130.7 minutes) Total physical activity: +high-walkability (210.5 versus 139.9 minutes)
|
From page 204... ...
Summary of Existing Research- Physical Activity Literature Physical Activity Study Sampling Survey Variable Sallis et al. 1990c 2,053 adults with Sedentary versus mean of exerciser (based 48 years in on three or more San Diego exercise ses sions per week)
|
From page 205... ...
Race/ethnicity ties perceived as con Marital status venient (15 items, y/n) Smoking Neighborhood environ ment (scale)
|
From page 206... ...
ipate in more survey physical activity Stahl et al. 2001c 3,343 adults, MAREPS study, Active versus in 6 European 19971998 active (based on countries cross- participation in (Belgium, sectional any gymnastics, Finland, phone survey physical activity, Germany, or sports)
|
From page 207... ...
: area +local opportunities Education level offers many opportuni- (logistic regression) Country ties to be active, local clubs and other providers offer many opportunities, commu nity does not do enough for citizens and their physical activity Age Neighborhood features Bikeway use: Gender (y/n)
|
From page 208... ...
Summary of Existing Research- Physical Activity Literature Physical Activity Study Sampling Survey Variable BRFSS sampling sectional (based on mod plan survey erate activity and vigorous activity) NOTE: ANOVA = analysis of variance; BMI = body mass index; BRFSS = Behavioral Risk Factor Sur veillance System; MAREPS = Methodology for the Analysis of the Rationality and Effectiveness of Prevention and Health Promotion Strategies; NHANES = National Health and Nutrition Examination Survey; SES = socioeconomic status.
|
From page 209... ...
64436_TRB_169_238 4/25/05 10:20 AM Page 209 Current State of Knowledge 209 Controls/ Built Environment Confoundersa Variables Resultsb able scenery, fre quently see others exercising, high levels of crime Safe to walk or jog alone during the day (5-point scale) Barriers (5-point scale)
|
From page 210... ...
2004. The Built Environment and Physical Activity: Empirical Methods and Data Resources.
|
From page 211... ...
Transportation Research D, Vol.
|
From page 212... ...
2002a. Socioeconomic Status Differences in Rec reational Physical Activity Levels and Real and Perceived Access to a Supportive Phys ical Environment.
|
From page 213... ...
Cited in University of CaliforniaBerkeley Traffic Safety Center Issue of Safety, Physical Activity, and the Built Environment, Vol.
|
From page 214... ...
1997. Travel Behavior as Function of Accessibility, Land Use Mix ing, and Land Use Balance: Evidence from San Francisco Bay Area.
|
From page 215... ...
1997. Assessment of Influence of Land Use Transportation System on Travel Behavior.
|
From page 216... ...
1997. Assessing Per ceived Physical Environmental Variables That May Influence Physical Activity.
|
From page 217... ...
2002. Environmental, Policy and Cultural Factors Related to Physical Activity in Sedentary American Indian Women.
|
From page 218... ...
2002. Environmental, Policy, and Cultural Factors Related to Physical Activity in Well-Educated Urban African-American Women.
|
Key Terms
This material may be derived from roughly machine-read images, and so is provided only to facilitate research.
More
information on Chapter Skim is available.