Show summary Hide summary
A fresh analysis from road-safety researchers finds mobile-phone monitoring apps detect far more pre-crash phone activity than police reports do. The gap raises new questions about distracted driving, insurer programs, and how much privacy drivers trade for discounts.
What the new IIHS analysis compared
Phone location data can incriminate you after a crash: police may not even notice
Insured driver ticketed after Flock camera falsely flagged no insurance
The Insurance Institute for Highway Safety matched nearly 17,000 crashes from 2021–2024 with data from safe-driving apps powered by Cambridge Mobile Telematics (CMT). Researchers looked at police crash reports and anonymized telematics records side by side.
- Data set covered crashes in four U.S. states.
- Telematics came from drivers who opted into programs.
- Researchers used aggregated, anonymized feeds for analysis.
How big is the difference?
The contrast was striking. Police reports noted phone use in under 1% of crashes. By contrast, telematics flagged phone activity in a larger share of incidents.
Telematics detected phone activity within 30 seconds of impact in about 7% of single-vehicle crashes, and about 8% of two-vehicle crashes. That equals at least seven times the rate captured in police paperwork.
Numbers by crash type and phone behavior
Two-vehicle crashes (15,301 incidents)
- Telematics (within 10 seconds): handheld calls 14, hands-free 53, manipulation 287, passive unlocked 337.
- Telematics (within 30 seconds): handheld calls 19, hands-free 87, manipulation 590, passive unlocked 534.
- Police reports: handheld calls 6, hands-free 1, manipulation 17, passive unlocked 15.
Single-vehicle crashes (1,543 incidents)
- Telematics (within 10 seconds): handheld calls 2, hands-free 0, manipulation 23, passive unlocked 40.
- Telematics (within 30 seconds): handheld calls 3, hands-free 3, manipulation 49, passive unlocked 52.
- Police reports: handheld calls 1, hands-free 0, manipulation 3, passive unlocked 5.
What the telematics really tells us
CMT’s algorithms can classify several behaviors. They separate handheld versus hands-free calls. They detect physical interaction like swipes or taps. They also note when the screen was unlocked.
- Screen unlocked could mean navigation, a music app, or video.
- “Manipulation” covers any tap, scroll, or swipe before impact.
- Time windows used in the study were 10 and 30 seconds before a crash.
Importantly, the presence of phone activity near a crash does not prove it caused the crash. The data shows correlation, not definitive causation.
Privacy, control, and who owns the data
Drivers voluntarily enroll in many insurer telematics programs. CMT says it does not sell raw driving data. Still, telematics systems collect detailed signals.
- Location and movement traces.
- Crash detection triggers.
- Phone interaction events and screen status.
Insurers that run these programs typically control the data for their offerings. CMT’s privacy policy also allows disclosure when required by law or legal process.
Potential implications for drivers and policy
Telematics can offer researchers a clearer picture of distracted driving than police reports do. At the same time, the same records create a granular timeline of what a driver did before a crash.
- Drivers may get discounts in exchange for sharing telematics.
- That sharing builds a searchable record of phone use and location.
- Law enforcement still needs legal authority to access detailed driving histories.
Before accepting a telematics program, drivers might weigh the value of an immediate discount against the long-term presence of detailed driving data held by an insurer.
Questions drivers should ask before opting in
- Who is the data controller for my driving records?
- How long will the insurer or vendor retain raw and aggregated data?
- Under what circumstances can the company share my information?
- Does enrollment affect how claims are handled or investigated?
- Can I view and delete the raw data linked to my account?
