TESTING ASSOCIATIONS BETWEEN PHYSICAL ACTIVITY AND THE BUILT ENVIRONMENT

1
TESTING ASSOCIATIONS BETWEEN PHYSICAL ACTIVITY
AND THE BUILT ENVIRONMENT
Department of City and Regional
Planning University of North Carolina, Chapel Hill
Daniel A. Rodríguez, Ph.D., Kelly Evenson, Ph.D.,
Kelly Clifton, Ph.D., Robert Schneider, AICP,
Asad Khattak, Ph.D., David Salvesen, Ph.D.,
Stanley Sedo, Ph.D.
Aims
Hypotheses
Study location (cont)
Measurement Methods (cont)

• Individual attitudes and the built environment
  contribute to explain MVPA behavior
• The higher the cost of being physically active
  at a given location, the lower the physical
  activity at that location, all else held equal
• Engaging in physical activity at one location is
  positively related to physical activity in other
  locations effects are complementary

• Test relationships between objectively measured
  MVPA and expanded range of objective measures of
  the built and natural environments
• Examine potential substitution effects among the
  locations where MVPA can take place
• Account for individual preferences and attitudes
  (e.g., about health, exercise, the urban and
  natural environments) to address issues of
  self-selectivity

• Data collection methods for primary data

Physical Activity
Environmental
Individual

• Accelerometer and diary given in person to
  participant mailed back

• Audit of TAZs where participants reside
  neighboring TAZs

• In-person questionnaire administered when
  protocol and accelerometer are given to
  participants

Source TransMilenio S.A.
Analysis
Research Design
Motivation

• Dependent variables
• Total weekday, weekend, and weekly MVPA time
• Amount of time devoted to MVPA at each location
• Total number of MVPA events lasting more than 10
  minutes
• Number of MVPA events and total time devoted to
  MVPA for utilitarian travel
• Number of MVPA events and total time devoted to
  MVPA for leisure and recreation
• Behavioral model of PA time allocation (see
  handout)
• Statistical methods
• Single-equation count models, interval regression
  models
• Single-equation multi-level models
• Multi-equation seemingly unrelated regression
  models of time allocations

BEI in study area, using principal components
analysis

• A quasi-experimental, cross-sectional research
  design in a single study area
• Two-stage cluster sampling
• Select two urban, two suburban, one exurban TAZ

• Existing ambiguities about relationship between
  micro-level environments and physical activity
• Paucity of research on interdependencies among
  different ways of integrating physical activity
  into ones daily life
• Disconnection among different disciplines in
  approaching relationships of interest

• Recruit participants from each TAZ
• Follow locations ethnic, sex, and income
  distribution of study area

Study location
Conceptual framework
Measurement Methods

• Montgomery Cty, Maryland (in the Washington DC
  Metropolitan Area)
• Excellent GIS and archival data
• Contains healthy mix of development types
  (various built environments)
• BEI index based on
• Household density
• Job density
• Density of sidewalks, walking paths, roads, and
  bus service
• of commuters taking transit, walking, bicycling
  to work
• Proximity to Washington Metro station

• If the environment is to influence MVPA it does
  so by affecting the cost of participating in
  such activities at different locations
• Examine allocation of time to various time uses
• Use socio-ecologic framework to identify relevant
  variables/relationships

• Three types of data utilized

Physical Activity
Environmental
Individual

• Accelerometer
• Modified activity diary (weeklong recall)

• Secondary data
• Vector and Raster GIS data
• Pictometry
• Cadastral data
• Zip USA
• Primary data
• Urban infrastructure and streescale
• Aesthetic and appeal
• Destination quality and mix

• Socio-demographics
• Attitudes
• Residential info.
• Health
• Employment
• Social-environmental factors

Progress to Date

• Stakeholder kickoff meeting on 12/12/03
• Research kickoff meeting and data demo on
  01/12/04
• Visit us at
• http//alr.unc.edu

This research project is supported by Active
Living Research, a program funded by the Robert
Wood Johnson Foundation