Correlates of Bicycling for Recreation and Transportation: Ecological Approach

Anna K Porter, The University of Texas School of Public Health


Bicycling is known to be associated with significant health benefit. Few studies have examined potential correlates of domain specific bicycling, recreation and transportation, in the context of the United States. This dissertation aimed to assess the association between (1) intrapersonal and environmental variables and bicycling for recreation and transportation, (2) perceived social and built environment factors and bicycling for recreation and transportation, and (3) objectively measured built environment variables and bicycling for recreation and transportation. ^ The first journal article was a secondary analysis of survey data of adults aged 18 and above from the United States National Survey of Bicyclist and Pedestrian Attitudes and Behaviors. Using logistic regression methods, three outcomes were explored: bicycling (yes/no), bicycling for transportation or recreation versus no bicycling, and a subgroup of bicycling for transportation versus recreation. In the multivariable model, significant differences were seen when comparing bicyclists to non-bicyclists across the ecological levels, including differences by gender, age, race, education, income, employment status, hand having a bicycle lane and/or path within ¼ mile of the participants’ residence. Significant differences were also seen when examined by domain, and when transportation cyclists were compared to recreation cyclists. ^ The second and third journal articles were secondary analyses of data from a pilot study exploring associations between ecological factors and domain-specific bicycling among a convenience sample of adult residents (18–65 years) of Travis County, Texas and Jefferson County, Alabama who had ridden a bicycle at least once in the past year. In journal article two, Spearman correlation coefficients indicated that perceived social and built environment factors were correlated with recreational cycling frequency to a lesser degree than transportation cycling frequency, and that bicycle route choice was the strongest correlate of total bicycling frequency (rs=0.39) and recreation bicycling frequency(rs=0.28), particularly for females (rs=0.41 and rs=0.33). For transportation bicycling frequency, the strongest correlates included connectivity (rs=0.24), and safety from crime (rs=0.24). In journal article three, recreation bicycling frequency was not notably correlated with any objectively measured built environment variables. Using exploratory factor analysis, an index of transportation bikeability was developed, identified by bike lanes 3km, residential density 3km, population density 3km, ozone 1500m (inverse) distance to transit (inverse), parks 3km, and tree canopy coverage 1500m (inverse). This index was moderately correlated with transportation bicycling frequency (r s=0.45). ^ The results of these analyses suggest that multilevel ecological factors are associated with bicycling, and that these factors differ by bicycling domain. Future research should further explore what ecological factors influence bicycling behavior by domain in diverse, representative populations throughout the United States.^

Subject Area


Recommended Citation

Porter, Anna K, "Correlates of Bicycling for Recreation and Transportation: Ecological Approach" (2017). Texas Medical Center Dissertations (via ProQuest). AAI10274666.