About me
Examining bicycling activity after bicycle theft in the US
The risk of bicycle theft has been found to be a deterrent to bicycle ownership and use (Winters et al., 2015; Castillo-Manzano et al., 2015; Piatkowski & Marshall, 2015). But literature on quantitative estimates of bicycling activity of individuals after bicycle theft is rare. Limited evidence suggests that 45% bicyclists bicycled less often or stopped bicycling after theft (Cohen et al., 2024), 7%–31% of bicyclists did not replace their stolen bicycles (van Lierop et al., 2015; Cohen et al., 2024), and only 2.5%–15% of the stolen bicycles were recovered (van Lierop et al., 2015; Cohen et al., 2023). Only the study by Cohen et al. (2024) examined the relationships between bicycling activity after theft and factors such as socio-demographic characteristics and pre-theft bicycling activity. However, the generalizability of the findings of Cohen et al. (2024) is likely to be limited because the convenience sample that they used might not be representative of the general US population.
This study investigates bicycling activity after theft using weighted data that is representative of active bicyclists within the general US population. It examines the extent to which bicycling frequency after theft and replacement of stolen bicycle vary based on factors including socio-demographic characteristics, pre-theft bicycling frequency, recovery status of stolen bicycle, and bicycle insurance. We also examine the extent to which reporting bicycle theft and bicycle registration influence recovery of stolen bicycles.
We collected data from individuals in the US through a survey. We collected a representative sample (n = 430) in partnership with YouGov (an online polling company) in March 2024. We also collected a convenience sample (n = 3055) in partnership with Bike Index (a bicycle registration organization) between September 2023 and November 2024. We combined the representative sample and the convenience sample into a single dataset for analysis. The sample from YouGov is expected to be representative of active bicyclists within the general population in the US. We calculated weights for the combined dataset using raking based on the age, gender, race, household income, and education marginal distributions of the representative sample from YouGov. We used the weighted combined dataset for analysis to enhance the generalizability of our findings.
Preliminary results show that approximately 72% of the stolen bicycles were not recovered. Of the stolen bicycles there were not recovered, about 23% of bicycles were not replaced and about 20% of bicycles were replaced with cheaper or free bicycles. In approximately 13% of stolen bicycle cases, individuals stopped bicycling, and in about 30% of cases, individuals bicycled less often after theft. Using multilevel logistic regression modeling, this study will further examine the differences in post-theft bicycling activity based on factors such as socio-demographic characteristics, pre-theft bicycling frequency, stolen bicycle recovery, and bicycle insurance. Insights gained from this study can help transportation planners and policymakers design targeted interventions such as bicycle leasing with theft protection and provision of subsidized temporary rental bicycles that support bicycling activity of vulnerable groups of people who are at a greater risk of reducing bicycling after experiencing theft. The findings will also shed light on effective ways to recover stolen bicycles.
