Environmental invitingness for transport-related cycling in middle-aged adults: a proof of concept study using photographs

Introduction: Current evidence on associations between modifiable environmental characteristics and transport-related cycling remains inconsistent. Most studies on these associations used questionnaires to determine environmental perceptions, but such tools may be subject to bias due to unreliable recall. Moreover, questionnaires only measure separate environmental characteristics, while real environments are a combination of different characteristics. To overcome these limitations, the present proof of concept study used panoramic photographs of cycling environments to capture direct responses to the physical environment. We examined which depicted environmental characteristics were associated to environments' invitingness for transportation cycling. Furthermore, interactions with gender and participants' cycling behavior were examined. Methods: Fifty-nine middle-aged adults were recruited through purposeful convenience sampling. During a home visit, participants took part in a structured interview assessing demographics and PA during the preceding seven days, followed by an intuitive choice task and a (cognitive) rating task, which both measured 40 photographed environments' invitingness to cycle along. Multi-level cross-classified analyses were conducted using MLwiN 2.26. Results: Both tasks' multivariate results showed that presence of vegetation was identified as the most important environmental characteristic to invite people for engaging in transportation cycling, even when the amount of vegetation was relatively small. In the bivariate analyzes, some differences between results of the cognitive rating task and the intuitive choice task were found, showing that invitingness measured by the rating task was associated with environmental maintenance and cycling infrastructure, whereas invitingness determined by the choice task was associated with more traffic-oriented characteristics. Moreover, only for the choice task's results, moderating effects of gender and participants' cycling behavior in the preceding week were observed. Conclusion: The present study provides proof of concept that capturing people's less cognitive, more intuitive responses to an environment's invitingness for transport-related cycling may be important for revealing environment-behavior associations. If replicated in future studies using larger samples, results of our innovative measurements with photographs, especially those on vegetation, can complete the existing knowledge on which environmental characteristics are important for transportation cycling in adults and could form a basis to inform health promoters and local policy makers. However, future studies replicating our study method in larger samples and other population subgroups are highly encouraged. Moreover, causal relationships should be explored.