Evidence quality 4.13/5
Eight-dimension review score against the quality rubric . Each dimension scored 1–5.
- D1 Source grounding
- 4/5
- D2 Source authority
- 5/5
- D3 Arithmetic
- 3/5
- D4 Uncertainty
- 4/5
- D5 Scope
- 4/5
- D6 Prose
- 4/5
- D7 Perception honesty
- 4/5
- D8 Caveat completeness
- 5/5
Pick challenger
The canonical number comes from a 2005 BMJ case-crossover study by McEvoy and colleagues: drivers making a hand-held phone call at the moment of a crash had 4.1 times the odds of being in that crash compared to a matched baseline window when they were not on the phone. Dingus et al. (PNAS, 2016) reached a similar conclusion using naturalistic driving data from the SHRP 2 passenger-car study, finding an odds ratio of 2.2 for talking on a handheld cell phone — lower than McEvoy, but still double baseline, and produced by a different method on a different dataset. Both are per-epoch figures: they describe risk during the specific seconds a call is active, not across an entire trip or a year of driving.
What these studies do not give is a lifetime number, and that gap is the reason this entry carries no reliable estimate rather than a headline probability. An odds ratio describes how much the crash odds rise during the seconds a call is active; turning that into a lifetime risk requires two further ingredients neither study provides — a baseline crash rate for the average driver and the fraction of total driving time spent on calls. A driver who takes one or two brief calls per trip is exposed to the elevated-risk window for only a small slice of total driving, so the per-epoch odds ratio overstates the lifetime effect by a large but unquantified amount. Without a sourced exposure-weighted multiplier, any single “1 in N lifetime” figure would be an invented number dressed up as a measurement, so none is published here.
What the McEvoy study did make uncomfortably clear is that the hands-free version of the same call carries OR 3.8, statistically indistinguishable from hand-held (OR 4.9) — its authors concluded that using a hands-free phone is “not any safer.” Most of the risk appears to come from the conversation itself, not from the physical act of holding a phone. The hands-free-versus-handheld distinction matters most for the manual element — lane-keeping degrades slightly when one hand is occupied — but the bulk of the distraction effect is cognitive, which is why legislation banning hand-held use alone tends not to produce the fatality reductions that broader distraction rules aim for. Federal crash statistics are also widely understood to undercount cell-phone involvement, because phone use is not reliably recoverable from crash investigations unless phone records are specifically subpoenaed, so the absolute counts in official data should be treated as floors, not full totals.
Related tidbits
A habitual hand-held caller faces roughly 1 in 140 lifetime odds of a crash tied to the behavior, about 4× the sober-driver rate. Many treat phone calls as far safer than texting and assume hands-free is fully safe.
Claim ledger
Every number below is what each source reported, with the verbatim quote we relied on and how we arrived at our figure. Click any link to verify directly.
-
[1] McEvoy SP, Stevenson MR, McCartt AT et al., BMJ — Role of mobile phones in motor vehicle crashes resulting in hospital attendance: a case-crossover study
Role of mobile phones in motor vehicle crashes resulting in hospital attendance: a case-crossover study- Statistic
Mobile phone use in the 10 minutes before a crash associated with a 4-fold increase in crash risk (overall OR 4.1, 95% CI 2.2–7.7); risk raised irrespective of device type (hands-free OR 3.8, 95% CI 1.8–8.0; hand-held OR 4.9, 95% CI 1.6–15.5) — hands-free not significantly different from hand-held- Excerpt
“"Driver's use of a mobile phone up to 10 minutes before a crash was associated with a fourfold increased likelihood of crashing (odds ratio 4.1, 95% confidence interval 2.2 to 7.7, P < 0.001)." ”
- Source data from
- 2005-08-20
- Accessed
- 2026-05-04 · archived copy
- Calculation
- McEvoy 2005 is the canonical case-crossover study: it compared drivers' own phone records in the 10-minute window before a crash versus a matched control window on the same day the prior week. The OR of 4.1 is the per-epoch relative risk while a call is in progress — not a per-trip, per-year, or lifetime figure, and the study reports no absolute crash rate or exposure-weighted multiplier. Critically, hands-free calls showed OR 3.8, statistically indistinguishable from hand-held (OR 4.9), supporting the view that manual distraction is not the dominant mechanism. Because this source supplies only a relative per-epoch odds ratio and no baseline crash frequency, it cannot on its own be converted into a lifetime probability — a key reason this entry carries no_reliable_estimate.
-
[2] Dingus et al., Proceedings of the National Academy of Sciences (PNAS) — Driver crash risk factors and prevalence evaluation using naturalistic driving data
Driver crash risk factors and prevalence evaluation using naturalistic driving dataSee all 4 Likelier entries citing this source →
- Statistic
Talking on a handheld cell phone: OR 2.2; texting: OR 6.1; dialing: OR 12.2; overall handheld cell phone interaction: OR 3.6 (all vs model driving in SHRP 2 passenger-car naturalistic sample)- Excerpt
“"The overall risk of interacting with a handheld cell phone is 3.6 times higher than model driving." ”
- Source data from
- 2016-03-08
- Accessed
- 2026-05-04 · archived copy
- Calculation
- Dingus 2016 provides the naturalistic-driving complement to McEvoy's case-crossover design. The 2.2x OR for talking on a handheld phone is lower than McEvoy's 4.1 partly because naturalistic studies capture near-crashes differently from hospital attendance databases. The two estimates bracket the plausible per-epoch range (roughly 2.2–4.1x). Like McEvoy, Dingus reports relative odds ratios against a model-driving baseline, not absolute crash rates per trip or per year, so it supports the direction and rough magnitude of the elevated risk but not a lifetime number. For context, texting's OR in the same dataset is 6.1 — nearly three times higher than talking on a handheld phone — which is the relevant comparison for drivers who think "just talking" is close to safe.