ABSTRACT
The novelty of carsharing as an alternative to private car raises a number of operating logistics questions. This study seeks to determine factors affecting vehicle usage through multilevel regression analysis and vehicle availability through a logistic regression analysis using the Communauto carsharing network in Montreal, Quebec case study. The number of vehicles parked at a station has the most effect on availability, with a great variation during the seasons. Vehicle usage is affected by average vehicle age, and by member concentration in the vicinity of stations. This research can be beneficial for carsharing operators to build or expand their network.
ACKNOWLEDGMENTS
We would like to thank Communauto, and especially Michelle Delisle-Boutin and Marco Viviani, for their support through the duration of this study, and for providing the data used in this project. We would like to thank Nithya Vijayakumar, McGill University for revising the paper. Last but not least we would like to thank the three anonymous reviewers for their valuable feedback.
Notes
n = 8,673
n = 70,871
Note. Dependent variable: Monthly hours-reserved by vehicle n = 8,673.
***99% confidence level; **95% confidence level; *90% confidence level.
Note. Dependent variable: Outcome of the binary test.
n = 70,871; Log likelihood = −37,858.56; Model χ 2 = 22,484.96 McFadden pseudo-R 2 = 0.2290.
Minimum significance: ***98% confidence level.