The effect of Real-Time feedback and incentives on speeding behaviors using Telematics: A randomized controlled trial
Mohammad-Reza Malekpour,Seyyed-Hadi Ghamari,Erfan Ghasemi,Seyedamirhossein Hejaziyeganeh,Mohsen Abbasi-Kangevari,Kavi Bhalla,Nazila Rezaei,Saeid Shahraz,Arezou Dilmaghani-Marand,Seyed Taghi Heydari,Negar Rezaei,Kamran B Lankarani,Farshad Farzadfar
DOI: https://doi.org/10.1016/j.aap.2023.107216
Abstract:Speeding behaviour of drivers is highly correlated to their tangible consequence. Therefore, this study aimed to evaluate the effectiveness of telematics-based feedback and financial incentives in reducing speeding behaviors through a randomized controlled field trial. This randomized controlled trial included four groups of (1) control, (2) information-only, (3) gain-of-reward, and (4) loss-of-reward. While drivers of the control group were unobtrusively monitored using telematics devices, drivers of the information-only group received real-time, weekly, and monthly feedback via text message. In both groups with financial incentives, in addition to receiving feedback, drivers could receive payments at the end of each month based on observed speeding. The primary outcome was the distance traveled at speeds more than 10% (S10 + ) above the posted speed limit as a proportion of the total traveled distance. A total of 397 male taxi drivers consented to participate in the study. After the stratified randomization, the mean age of the participant was 46.0 (95% Confidence Interval: 43.8 to 48.2), 47.0 (44.7 to 49.3), 46.1 (43.7 to 48.5), and 48.8 (46.5 to 51.1) years for the control, the information-only, the gain-of-reward, and the loss-of-reward groups, respectively. The mean S10 + rate per 100 km was 0.9 (0.5 to 1.2) for the control, 0.8 (0.4 to 1.1) for the information-only, 0.7 (0.3 to 1.1) for the gain-of-reward, and 1.3 (0.4 to 2.2) for the loss-of-reward group at the start of intervention. During the intervention phase, the loss-of-reward group with 0.6 (0.5 to 0.7) had the lowest mean of S10 + rate, followed by the gain-of-reward group with 0.8 (0.7 to 0.8). The loss-of-reward and gain-of-reward groups were the most influenced groups by the intervention, with 38.0% (13.2 to 55.7; p-value < 0.01) and 29.4% (2.0 to 49.2; p-value = 0.04) less S10 + rate than the control group, respectively. It was figured out that providing information regarding drivers' behavior without appropriating motives or penalties would have a minuscule impact.In addition, the group of drivers who were treated with instant punishment was the most influenced group by the intervention.