How Representative of Bicycling Populations Are Smartphone Application Surveys of Travel Behavior?. Issue 1 (2016)

• Record Type: Journal Article
• Authors: Blanc, Bryan; Figliozzi, Miguel; Clifton, Kelly
• Is Part Of: Transportation research record. Volume 2587:Issue 1(2016)
• Page Start: 78
• Page End: 89
• Publication Date: 2016
• Subjects: Transportation -- Periodicals; Roads; Transport -- Périodiques; Routes -- Périodiques; Routes -- Conception et construction -- Périodiques; Roads; Transportation; 388.05
• DOI: 10.3141/2587-10
• Languages: English
• ISSNs: 0361-1981

Abstract

Bicycle transportation increasingly has become a central focus of urban regions invested in the improvement of livability, sustainability, and public health outcomes. Recently, transportation agencies across North America have deployed travel surveys with a smartphone application to gain a better understanding of bicycling travel behavior to forecast travel, to invest in infrastructure, and for a variety of other purposes. A potential limitation of data sets crowdsourced with smartphones is sampling bias (i.e., the demographic characteristics of the smartphone application users may not match the characteristics of the cycling population). Such a bias can be caused by the passive nature of sample recruitment, by differences in access to smartphone ownership or in familiarity with the technology, or both. This study examined the characteristics of several user samples from bicycle smartphone application deployments in North America. Differences between these samples were highlighted, and the smartphone samples were compared with cycling samples from travel survey data sets. Whenever possible, a statistical test was used to calculate the statistical significance of the differences between smartphone samples and traditional travel survey samples. Compared with travel surveys, smartphone applications tended to undersample females, older adults, and lower-income populations and to oversample some minority ethnicity populations. The analysis also revealed that, for cities in which travel survey sample sizes were small, smartphone applications could provide higher-resolution data and larger sample sizes of bicyclists. For transportation agencies, all of these findings are useful to plan future travel survey and sample recruitment efforts.