Undercounting of cellphone use in police crash reports

Cox, Aimee E. / Teoh, Eric R. / MacInnis, Rebecca
Insurance Institute for Highway Safety
August 2026

Abstract
Introduction: There is widespread agreement that police-reported crash data undercount the number of crashes involving cellphones because cellphone use is difficult to identify as a crash contributor on-scene. This study investigated the extent of the underreporting by linking Cambridge Mobile Telematics crash data with data compiled from police crash reports to compare cellphone use as recorded by the responding officer with phone use as recorded by telematics.
Method: We linked nearly 17,000 crashes from police crash reports to crashes identified by telematics in four states between 2021 and 2024 by date, time, and location. Distraction as recorded by police was compared with phone use as recorded by telematics in the 30 seconds and 10 seconds prior to the moment that a crash was detected.
Results: Telematics data indicated that enrolled drivers had been using a cellphone within 10 seconds of 4% of single-vehicle crashes and 5% of two-vehicle crashes. Telematics data recorded phone use within 30 seconds of 7% of single-vehicle and 8% of two-vehicle crashes. Passive use of unlocked phones (such as navigation or watching a video) and manipulation (such as tapping or swiping) were the most detected, while handheld and hands-free calls were less common. Police crash reports reflected cellphone distraction in less than 1% of crashes. There were virtually no instances in which both datasets identified the same type of cellphone use as a crash factor.
Conclusion: This study suggests police crash reports may undercount cellphone use before a crash by a factor of at least 7. Practitioners should use caution when using crash data based on such reports to determine the extent that cellphone distractions contribute to motor vehicle crashes and injury.
Abstract Introduction: There is widespread agreement that police-reported crash data undercount the number of crashes involving cellphones because cellphone use is difficult to identify as a crash contributor on-scene. This study investigated the extent of the underreporting by linking Cambridge Mobile Telematics crash data with data compiled from police crash reports to compare cellphone use as recorded by the responding officer with phone use as recorded by telematics.
Method: We linked nearly 17,000 crashes from police crash reports to crashes identified by telematics in four states between 2021 and 2024 by date, time, and location. Distraction as recorded by police was compared with phone use as recorded by telematics in the 30 seconds and 10 seconds prior to the moment that a crash was detected.
Results: Telematics data indicated that enrolled drivers had been using a cellphone within 10 seconds of 4% of single-vehicle crashes and 5% of two-vehicle crashes. Telematics data recorded phone use within 30 seconds of 7% of single-vehicle and 8% of two-vehicle crashes. Passive use of unlocked phones (such as navigation or watching a video) and manipulation (such as tapping or swiping) were the most detected, while handheld and hands-free calls were less common. Police crash reports reflected cellphone distraction in less than 1% of crashes. There were virtually no instances in which both datasets identified the same type of cellphone use as a crash factor.
Conclusion: This study suggests police crash reports may undercount cellphone use before a crash by a factor of at least 7. Practitioners should use caution when using crash data based on such reports to determine the extent that cellphone distractions contribute to motor vehicle crashes and injury.