Yes, size does matter (for cycling safety)! Comparing behavioral and safety outcomes in S, M, L, and XL cities from 18 countries

Although most actions aimed at promoting the use of active transport means have been conducted in ‘large’ cities, recent studies suggest that their cycling dynamics could hinder the efforts put into infrastructural, modal share, and cycling culture improvements. Aim: The present study aimed to assess the role of city sizes on riding behavioral and crash-related cycling outcomes in an extensive sample of urban bicycle users. Methods: For this purpose, a full sample of 5705 cyclists from >300 cities in 18 countries responded to the Cycling Behavior Questionnaire (CBQ), one of the most widely used behavioral questionnaires to assess risky and positive riding behaviors. Following objective criteria, data were grouped according to small cities (S; population of 50,000 or fewer), medium cities (M; population between 50,000 and 200,000), large cities (L; population between 200,000 and one million), and megacities (XL; population larger than one million). Results: Descriptive analyses endorsed the associations between city size, cycling behavioral patterns, and mid-term self-reported crash outcomes. Also, it was observed a significant effect of the city size on cyclists' traffic violations and errors (all p <.001). However, no significant effects of the city size on positive behaviors were found. Also, it stands out that cyclists from megacities self-reported significantly more violations and errors than any of the other groups. Further, the outcomes of this study suggest that city sizes account for cycling safety outcomes through statistical associations, differences, and confirmatory predictive relationships through the mediation of risky cycling behavioral patterns. Conclusion: The results of the present study highlight the need for authorities to promote road safety education and awareness plans aimed at cyclists in larger cities. Furthermore, path analysis suggests that “size does matter”, and it statistically accounts for cycling crashes, but only through the mediation of riders' risky behaviors.