Background:
Active living research has focused on the relationship between the social and built environment, and walking as a physically active behaviour. Empirical evidence shows relationships between walking and characteristics of neighbourhood of residence such as higher residential densities and greater mixing of uses. Research frequently hints at the importance of public transit in creating healthy cities supportive of walking. However, little has been done to identify mechanisms, causal pathways and potential factors in understanding the individual role of transit use, transit infrastructures and policies that favour the use of public transit when controlling for other variables of interest. Research that have considered transit relied on the distance to the nearest transit stops or station, on types of transit service (rail or bus), and on dummy indicators of transit use.
Objectives:
We seek to assess the relationship between number of trips taken by various motorized modes and meeting recommended walking guidelines, when controlling for individual, household and neighbourhood built environment. We also assess the relationship between being a beneficiary of employer sponsored public transit program and meeting physical activity recommendations through walking.
Methods:
Using SMARTRAQ, a two day activity-based travel diary in Metro Atlanta (2001-2002), we compared employed individuals taking automobile trips as drivers, as passengers and those taking transit trips in their likelihood of meeting public health recommendation for walking, while controlling for demographics, neighbourhood environment, access to transit, and employer subsidized transit incentives. We use a randomly selected sub-sample of the survey population that was asked additional survey items on employer incentives (n=4136). Meeting public health recommendations through walking a mean 30 minutes or more a day was measured as walking 2.4 km or more in a day (at an approximate speed of 3mph or 5km/h). Walking distances on the street network from origins to destinations were derived by Geographical Information System (GIS), summed together and coded as three walking categories. Non-walkers, those that do not walk sufficiently, and those that walk sufficiently to meet recommendations were then introduced as outcome variable in a multinomial logistic regression. GIS derived built environment measures of density, land use mix, connectivity, length of sidewalks and distance to transit were also introduced in the regression.
Results:
When controlling for other covariates, each additional transit trip was found to have an OR of 3.87 (CI 95% 2.93-5.11) of meeting public health recommendation as opposed to not walking at all. Individuals using employer sponsored transit passes were also more likely to walk 2.4 km or more, OR 2.23 (CI 95% 1.27-3.90). Car availability had a strong negative relationship with meeting public health recommendations, OR 0.18 (CI 95% 0.07-0.42) and additional automobile trips as driver were negatively related to meeting recommendations, when compared to moderate walkers (OR=0.82, CI=0.72-0.93). Other individual and household variables had expected independent effects. Land use measures also followed expected directions, although a few proved non-significant. The first two categories of distance to the nearest transit stop (within 450 meters and between 450 and 1000 meters) only significantly differed from the reference category in predicting moderate walking, but not in meeting recommendations.
Conclusions:
A positive relationship was found between mean number of transit trips and meeting public health recommendation by walking. Using employer sponsored transit incentive was also positively related to meeting recommendations. Results provide interesting insights on how greater transit provision and policies and programs that favour transit use can be important considerations within the overall set of strategies to promote physical activity. Numerous measures of transit level and quality of service already exist and should be integrated to studies that attempt to predict walking behaviour. This is especially feasible for studies that use objective built environment measures derived from GIS systems. Public transit research is giving growing importance to the synergies between walking and using transit. In transportation terms, a trip’s overall quality is no better than its worst link. Improving walking environments is therefore of considerable importance in promoting transit use, since most trips start and end with walking trips. Failure to control for the presence, quality of transit and transit trips taken may potentially confounding the relationship between the built environment and walking because public transit is typically more present, and used, in the types of neighbourhoods where more walking is recorded. At a policy level, public transit is an important research avenue that requires more attention since only changing the built environment may not be enough to spur individuals to walk more.
Support:
The study was conducted while the second author was at Georgia Tech, Atlanta, GA. Data collection supported by the Georgia Department of Transportation (GDOT).
