Why do people claim cellphone use increased the USA accident rate?

There are many post-crash reports of semi and auto drivers watching movies or p*rn or texting furiously as they massacre innocents.

I agree that if people merely used the audio (as police and truckers do with radios) there would not be a problem.

Hi badgolferman,

You are well aware that I'm trained as a scientist and as an engineer. The first thing we learn in science is to be driven empirically by data.

Unfortunately, most people are not data driven. They're driven by religion or emotion.

They make all their "scientific" decisions based on scary anecdotes. The news writes about the one sensationalist accident. Not the 100 boring accidents that occurred that very same day.

But most people believe only that sensationalistic accidents occur. Not boring accidents.

Yet, "The plural of anecdote is not data" is an expression in science.

You are aware that for the past 60 years of driving, when I pull up to the Costco pump, I ask the guy across from me what the difference is between the premium and the regular. I got the right answer twice.

If that's once a month, for 60 years, I have 718 wrong answers out of

Why? They're convinced their Honda Civic runs "better" and "with more power" on premium, or that using premium "once a month" cleans out the engine.

Why do they think something that absurd? Because it "sounds plausible" to them.

Same with car accidents, which are NOT cherry picked when we look at the statistics since the 1920s compiled by the US for every state.

I get it that everyone reads the news, where "dog bites car" gets more hits than "car hits dog", but anecdotal evidence is not science.

Anecdotal evidence is emotional. Just like "religious evidence" is emotional.

But it's not science.

The Catholic Church has more than 1,000 formally canonized saints, and the canonization process has officially recognized hundreds of alleged miracles associated with them. Hundreds of individual miracle claims do not become scientific evidence merely because there are hundreds of them.

I get it people think emotionally. Not scientifically.

To them, if the current propaganda is cell phones cause accidents, every accident they can find that has a cell phone is validation.

The same is true of thousands of people saying they saw a distracted driver with a cell phone in their hand (where, I guess the corollary, I suppose, is that these same people must never have seen a distracted driver before cell phones existed, i.e., for about a hundred years.

Weak minds will parrot propaganda every time. I don't even need to prove that, I hope.

But propaganda is not science.

This article discusses why people can give disproportionate weight to personal experiences and anecdotes even when larger, systematically collected datasets point in another direction.

"thinking anecdotally comes naturally, whereas thinking scientifically does not"

The question being asked here is about science. Facts. Good data. Real data. And, to my knowledge, only the USA has compiled that data.

Accident rate = number of motor-vehicle accidents ÷ vehicle miles driven Usually expressed as accidents per 1 million vehicle-miles traveled or equivalently per 100 million VMT (vehicle miles traveled).

Given that's the ONLY metric that matters in any first-order discussion, luckily the federal government explicitly recognizes number of accidents per 100 million VMT as a highway safety performance measure.

Unfortunately, our first data-collection issue is Census has historical motor-vehicle accident counts, while FHWA has the historical VMT series.

Since this is getting long, let's take that up separately, since we are not asking for anecdotal data, nor cherry-picked data, but for the accidents divided by miles driven, either of or both of which are accurate federal statistics which have been kept since the 1920s.

Oh god! Not this again...

Before anyone responds you need to address three key flaws in your OP.

1. You present no data nor evidence. As a "trained scientist" - lol! - you should do better than that. A vague link is not enough. Be explicit regarding what data there is, the source and what you think it supports and how.
2. Remember absence of proof is not proof of absence. A complex environment can mask one affect with another, competing, one.
3. An argument to say that we're always distracted and we shouldn't bother to target a solvable distraction isn't valid. All avoidable accidents and deaths on the road should be targeted and reduced.

If you do not address all three points. There's no reason to engage with this topic with you.

Hi Chris,

It's a perfectly valid question, especially since most people own purely religious belief systems which are not based on solid USA government facts.

Let's approach this problem, which, since most people only believe in the propaganda, clearly needs to be revisited, and with all the new valid data.

Let's look at the assessment by enumerating two sets of accurate facts. A. Census gives us the accident numerator since at least way back to 1937. B. FHWA has annual USA vehicle-miles-traveled all the way back to 1900. C. The accidents divided by the miles driven gives us the accident rate.

While the overall trends matter, for the first-order assessment, we only need to look at the statistics during the time period of the cell phone. 2. Pre-cellular baseline 1937-1982 No commercial US cellular service 2. Cellular introduced 1983 First commercial cellular service began 3. Early adoption 1984-1989 Cell phones existed but were still rare 4. Rapid adoption 1990-1999 Cellular subscriptions exploded 5. Mass adoption 2000-2004 >100 million U.S. wireless connections 6. Essentially ubiquitous 2005-2010 ~208M -> 296M wireless connections 7. Smartphone era 2011-present Smartphones become dominant

Hence, I posit it's reasonable for us to look the accident rate from January 1, 1983 (which we can reasonably define as the cellular era).

Here is more granular detail on the total USA wireless subscribers.

Roughly... 1990: 5.283 million wireless subscriber connections 1993: 16.009 million wireless subscriber connections 1994: 24.134 million wireless subscriber connections 1995: 33.786 million wireless subscriber connections 1996: 44.043 million wireless subscriber connections 1997: 55.312 million wireless subscriber connections 1998: 69.209 million wireless subscriber connections 1999: 86.047 million wireless subscriber connections 2000: ~109.5 million wireless subscriber connections 2001: ~128.4 million wireless subscriber connections 2002: ~140.8 million wireless subscriber connections 2003: ~158.7 million wireless subscriber connections 2004: ~182.1 million wireless subscriber connections 2005: ~207.9 million wireless subscriber connections By 2005, that's roughly 70% of the U.S. population in subscriber connections. By 2010 there were about 296 million wireless subscriber connections.

Questions we might ask ourselves BEFORE making any assessments, are Q1: What was happening to the accident rate before cellular phones existed? Q2: What happened to the accident rate during the 1983-2000 adoption curve? Q3: What happened to the accident rate after cellular phones became commonplace?

Since we're being scientific, what's our hypothesis if we BELIEVE that the cell phone caused accidents that wouldn't have happened prior to cell phones.

Is this the hypothesis we will strive to prove/disprove in our studies? The introduction and increasing prevalence of cell phones caused an increase in the rate of motor-vehicle accidents per vehicle-mile traveled, above the rate that would otherwise have occurred. That produces a very specific prediction, does it not: If cell phones cause accidents that would not otherwise have occurred, then, as cell-phone use spreads through the driving population, the accident rate should increase relative to the pre-cellular baseline.

If, what almost everyone believes is true, is born out by facts, then we should reasonably expect something like this in accident rate facts.

The stronger test is: Q: Did the accident rate after cellular adoption depart from the trajectory that was already occurring before cellular adoption?

What do folks think about that being the starting point hypothesis? If we proceed on that hypothesis, we should be able to assess the facts.

Familiar to what interface? Apple? Stock Android? Samsung? Nothing Phone?

Without regard to the larger questions you address, people driving while engaging in audio conversation, telephone or radio, seem not to be the actual problem.

Blasting that guy you have never met in a texting battle, playing video games, watching p*rn or movies and the like are of an entirely different risk level for distraction.

Legislatures, who regularly specify large hammers for disposing of an ant, conflate the two.

Hi Woozy Song,

Thanks for asking about that thread on beat sonic device screen mirroring.

While that thread talked about all the cellphone-to-dashboard automotive methods, mirroring of the exact phone on the dash should be familiar to the phone owner, since it's his/her phone that is being mirrored exactly.

Like this mirroring I do every day, all day, to operate my iOS and Android devices using the PC keyboard, mouse, monitor, speakers & clipboard.

It's familiar because it's literally the driver's exact phone on the dash.

Just as it's my exact phone about two feet tall on my monitor in this shot.

But this thread is designed to make people think about why there is no change in the accident rate in USA records which have been dutifully collected every year for every state for just a bit over a century.

For anyone to claim those well-respected records are "cherry picked", without providing any corollary records, is just a religious gripe.

What they're really saying is they hate the facts because those facts don't line up with their religious beliefs about cellphone use while driving.

The goal in this thread will be clear, where it is obvious I'm the only one who is actually providing facts (with cites, links, assessments & data).

The rest are clinging to their religious beliefs, which they feel, because they're religious, they can't be proven wrong when they claim all facts about cellphone use while driving must only come from biased propaganda.

I prefer the government data that has been collected for over a century. The number of accidents divided by the number of miles = accident rate

If cellphone use had *any* effect whatsoever, our hypothesis would be that there should be a spike or jump in the accident rate during those years. 1. Pre-cellular baseline 1937-1982 No commercial US cellular service 2. Cellular introduced 1983 First commercial cellular service began 3. Early adoption 1984-1989 Cell phones existed but were still rare 4. Rapid adoption 1990-1999 Cellular subscriptions exploded 5. Mass adoption 2000-2004 >100 million U.S. wireless connections 6. Essentially ubiquitous 2005-2010 ~208M -> 296M wireless connections 7. Smartphone era 2011-present Smartphones become dominant

The data is convincing. But religious zealots won't be swayed by facts. Propaganda is a well-known ferociously fierce competition to science.

Please explain the improvement in driver awareness, reaction time and overall safety on the public roads when changing from playing a video game on a telephone to playing a video game on the dash console.

If there's some advantage there, it escapes me.

Ditto for movies, p*rn and hot texting battles.

Hi Andrew,

I agree with you completely.

Remember back in 2008 when the Governor of NY, Andrew Cuomo, made a big propaganda show with the major cable companies and ISPs over kiddie p*rn, enabling AT&T and the others the excuse to summarily drop Usenet feeds?

Just as with the insurance companies in cahoots with the states, I don't doubt the cable companies approached Cuomo saying "hey, we want to drop Usenet, but let's do it together as a charade that we're 'doing our bit' to combat kiddie p*rn".

The fact is it was total theater and pure PR to claim Usenet was all p*rn.

They used a horrific emotionally-charged issue as cover to kill off an expensive, bandwidth-heavy legacy service that had nothing to do with the actual illegal content, looking like they were taking action while just ditching a service they simply didn't want to pay to maintain anymore .

I think you're arguing likewise, that lawmakers are unfairly grouping all forms of in-car distraction together under the same heavy laws, even though they represent vastly different levels of risk. a. Audio vs visual distractions b. Extreme distractions c. Overreaching laws

Standard audio conversations, whether on the phone, the radio, or talking to a passenger are not actually a major safety problem (unless it's that damned crying brat in the car seat that the driver just can's shut up.) :)

There's a difference between simple audio multitasking versus cognitive distractions like texting, playing video games, or watching movies/p*rn, stating these are on a completely different level of danger (although, I can't imagine anyone watching p*rn while driving, and, if they did, I'd argue that this is why I always got a 'good student' discount on my insurance and they don't because it's clear they make bad decisions).

The legislatures want to "appear" to be "doing something" about it, much like my Cuomo example which we all know in the USA because the cable companies used kiddie p*rn as the excuse for dropping our Usenet access.

The funny thing about "large hammers for disposing of an ant", is that we looked in the past at systematic reviews of studies which showed the effect of all US automotive safety laws that impacted the driver directly, and absolutely none of them (not even one!) showed any correlation whatsoever with accidents. Not even one. There was a second-order effect, but zero first order effects, even of laws such as seat belts, which makes sense, but there was zero correlation with accidents. None. Zip Nada.

That included cellphone laws as they went into effect in a staged manner per state, and there is no correlation whatsoever with reducing accidents.

Hence, I agree lawmakers pass broad, blunt laws that conflate low-risk audio habits with higher-risk cognitive habits without properly distinguishing between them.

Moving forward, our hypothesis should be that accident rates should have "changed" in some manner during the years that cell phones were introduced to driving, where we can delve into specific periods to look at accident rates.

1. Pre-cellular baseline 1937-1982 No commercial US cellular service 2. Cellular introduced 1983 First commercial cellular service began 3. Early adoption 1984-1989 Cell phones existed but were still rare 4. Rapid adoption 1990-1999 Cellular subscriptions exploded 5. Mass adoption 2000-2004 >100 million U.S. wireless connections 6. Essentially ubiquitous 2005-2010 ~208M -> 296M wireless connections 7. Smartphone era 2011-present Smartphones become dominant

If cellphones 'cause' accidents in a meaningful way, shouldn't we see "something" happening to the accident rate (which is good data kept for over a century) between the time before and the time during and after?

Yes? No?

People are experts at multitasking. They have no problem having one hand on the wheel, one hand holding their sandwich, and one hand on their beer. It's all perfectly safe. So how could anyone possibly be at risk just because they are paying attention to the person on Facebook whose political beliefs are clearly wrong instead of to the road? The road can take care of itself. And turn signals? Those other people have no business knowing where you're going. They should keep to themselves.

--scott

Hi Scott, I wonder if you even realize how absurd your argument is, to a scientist? You're attempting to ridicule your absurd distortion of the argument.

That's called beating a strawman. And claiming victory.

Stop it. Be scientific please.

a. You claim cellphones are a distraction (and they are) b. Out of, oh, there are about a hundred the government listed c. And then you claim, unilaterally, it's NUMBER ONE in that hundred???? d. When, clearly, it's not???? WTF? e. Worse, you seem to claim distractions are the ONLY cause of accidents! WTF? Not speed? Not drunken driving? WTF?

Who makes such absurd arguments by ridiculing the facts? No scientist would say what you just said, Scott. None.

Cellphones are not the major cause of accidents. They're not even the major distraction in accidents, in fact.

It's not even in the top ten for God's sake.

People have been handling distractions while driving since driving started. What's different now?

What is the claim that people are making here given nobody is claiming there aren't a zillion distractions while driving that everyone feels.

We covered, years ago, all the major distractions (there were about 100).

Cellphones barely made the top ten, and then it depended, as others have said, on what the driver was doing with the cell phone at any given time.

If cellphone use had *any* effect whatsoever, our hypothesis would be that there should be a spike or jump in the accident rate during those years. 1. Pre-cellular baseline 1937-1982 No commercial US cellular service 2. Cellular introduced 1983 First commercial cellular service began 3. Early adoption 1984-1989 Cell phones existed but were still rare 4. Rapid adoption 1990-1999 Cellular subscriptions exploded 5. Mass adoption 2000-2004 >100 million U.S. wireless connections 6. Essentially ubiquitous 2005-2010 ~208M -> 296M wireless connections 7. Smartphone era 2011-present Smartphones become dominant

Nobody ever said cell phones aren't one of the hundreds of distractions. What is your claim about the accident rate changes during that period?

The accident rate is the number of accidents divided by miles driven. a. FHWA gives annual USA vehicle-mile-traveled from 1900 through 2024. b. That's our denominator. c. NHTSA's General Estimates System (GES) began in 1988 and was specifically designed to estimate the number of USA police-reported motor-vehicle crashes that had any damage or injury. d. That's our numerator.

This is good data. It's not cherry picked.

What do YOU think it tells us about crashes during the cellphone period?

Well if it's like it is here, because it's intuitively obvious that it must increase the accident rate. We still regularly see people driving while holding and using their phones and obviously there would be less accidents if they did not do so.

In New Zealand it has been illegal to use your cellphone while driving (unless using a hands-free option) ... but the fine / penalty is so pathetic and chances of getting caught that many people still do it.

It isn't helped by almost all car manufacturers now having silly touchscreen systems in their cars, which means there's little difference between using that and a cellphone, other than the cellphone being a smaller screen size. :-(

Hi YourName,

Assuming no other factor was involved to cancel the arguments that people are claiming, if their hypothesis is cell phones cause accidents, then the accident rate should have skyrocketed at some point during the introduction of cellphones, right?

What's intuitively obvious is most accidents are NOT caused by cell phones. Speed and drunken driving are, almost certainly, vastly great represented.

Even so, given we know exactly when cellphones were introduced into the USA, those who argue cellphones meaningfully increased the USA accident rate have merely to show us valid USA statistics which back up their claim.

U.S. cell phones, crashes, miles, and crash rate - 1988-2024 Crash rate = NHTSA estimated police-reported crashes ÷ FHWA VMT x 100 million.

Year Wireless connections* Police-reported crashes VMT (billion) Crashes / 100M VMT

*Wireless figures are connections/subscriptions, not unique people. CTIA's historical series gives the exact 1988-2015 numbers/ For example, 1990 was 5,283,055 and 2015 was 377,921,241.

The crash figures through 2020 come from NHTSA's national GES/CRSS series. NHTSA's historical table explicitly starts at 1988 and gives 6,886,800 total crashes for that year.

The 2023 and 2024 numbers are from NHTSA's latest national estimates.

FHWA provides the national VMT series, including every year from 1988 through 2024.

Note sample designs changed between 2015 & 1016 for NASS GES & CRSS estimates. That doesn't invalidate the table. But it does mean that we should not interpret the jump from 202.4 in 2015 to 214.9 in 2016 as necessarily a real-world increase without looking closer at it.

Guess what? Yup. You guessed it.

Just like everyone intuitively believes "premium gas" gives their Honda Civic "more power", everyone believes the accident rate must have skyrocketed during the period that cell phones were introduced & after.

And yet... The numbers we can compare within the same GES system are valid:

Wireless connections:

Crash rate:

So while wireless connections increased roughly 56-fold from 1990 to 2010, the national police-reported-crash rate went down, not up.

Think about that fact...

Assuming no other factor was involved to cancel the arguments that people are making, if their hypothesis is "cell phones cause accidents," then the accident rate should have increased substantially at some point during the introduction and widespread adoption of cell phones, right?

What's intuitively obvious is that most accidents are NOT caused by cell phones. Speed, alcohol, weather, road conditions, vehicle conditions, driver experience, etc. are all potential causes.

Even so, given that we know approximately when cell phones were introduced into the USA, those who argue that cell phones meaningfully increased the USA accident rate have merely to show us valid USA statistics which support that claim.

Here are the national numbers.

U.S. CELL PHONES, CRASHES, MILES & CRASH RATE 1988-2024

Crash rate = NHTSA estimated police-reported crashes / FHWA VMT x 100 million.

Year Wireless Police-reported VMT Crashes/ connections* crashes billion 100M VMT

----------------------------------------------------------------

• Wireless figures are connections/subscriptions, NOT unique people. A person can have multiple connections, so these numbers should not be interpreted as "percentage of Americans who own a phone."

The historical CTIA series gives the exact annual wireless connection numbers for the earlier period. For example:

CTIA:

NHTSA's national crash series begins in 1988. It is based on the National Automotive Sampling System's General Estimates System (GES), which estimated the number of police-reported crashes in the USA.

NHTSA:

For the later years, NHTSA uses the Crash Report Sampling System (CRSS).

NHTSA:

FHWA provides the national vehicle-miles-traveled series.

FHWA:

There is a methodological break between GES and CRSS in 2016. NHTSA states that the estimates from the two systems are not directly comparable because their sampling designs differ.

Therefore, don't interpret the change from 2015 to 2016 as necessarily representing a real-world change in crashes.

The much cleaner comparison is entirely within the GES period:

WIRELESS CONNECTIONS

CRASHES / 100M VMT

So from 1990 to 2010, wireless connections increased approximately

During that same period, the national police-reported crash rate declined from approximately 302 crashes per 100 million vehicle-miles to approximately 182.

That's a decline of about 40%.

This does NOT prove that cell phones never cause individual accidents. Obviously, a driver can be distracted by a cell phone and crash.

It does, however, create a very simple empirical question for anyone claiming that cell phones caused a substantial increase in the overall U.S. accident rate:

Where is the corresponding increase in the NATIONAL ACCIDENT RATE?

The phones went from essentially nothing to hundreds of millions of connections.

The miles driven went up.

And the national crash rate, during the clean 1990-2010 GES period, went DOWN substantially.

Think about that fact.

Hi Nick,

If phones are causing a great number of accidents, we would expect it to show up in the USA statistics on the number of accidents per mile during the period just before, during & after the cellphone skyrocketing.

1. Pre-cellular baseline 1937-1982 No commercial US cellular service
2. Cellular introduced 1983 First commercial cellular service began
3. Early adoption 1984-1989 Cell phones existed but were still rare
4. Rapid adoption 1990-1999 Cellular subscriptions exploded
5. Mass adoption 2000-2004 >100 million U.S. wireless connections
6. Essentially ubiquitous 2005-2010 ~208M -> 296M wireless connections
7. Smartphone era 2011-present Smartphones become dominant

Yet, the accident rate went down, just as it had been trending for decades, and still trends downward, which needs to be explained if someone is claiming that skyrocketing cell phone ownership caused concomitant skyrocketing cellphone related accidents which wouldn't have been caused otherwise, right?

But we just showed the rates went down in the data for the United States which has been compiled very accurately over decades in use.

Bear in mind there is one important limitation in that the NHTSA's nationally estimated total-crash series begins in 1988.

For 1983-1987, we have federal national traffic data, but not the same GES total-crash series, so, for sure, we need to understand the data.

What we can put together cleanly is this:

Period Year Wireless Crashes VMT Rate connections billion /100M

---------------------------------------------------------------------- Cellular introduced 1983 ~0.2 M -- 1,653 -- Early adoption 1984 ~0.3 M -- 1,720 -- Early adoption 1985 0.34 M -- 1,775 -- Early adoption 1986 ~0.68 M -- 1,835 -- Early adoption 1987 ~1.2 M -- 1,921 -- Early adoption 1988 2.07 M 6,886,800 2,026 339.9 Early adoption 1989 3.51 M 6,652,815 2,096 317.4 Rapid adoption 1990 5.28 M 6,471,202 2,144 301.8

The VMT figures for 1983-2000 are available directly from the Federal Highway Administration's historical series. FHWA:

The national crash estimates beginning in 1988 come from NHTSA's GES series. The 1988 estimate is 6,886,800 crashes. The 2000 estimate is 6,393,624.

NHTSA:

The wireless numbers for the early-adoption period and 1990s come from the historical CTIA series.

Notice that this is the really important part when analyzing the dataset.

If we start the clean comparison at 1988, the progression is:

So during the period when wireless connections went from roughly

And that's not using fatalities as a proxy. These are total estimated police-reported crashes where you can't have a

From 1983-1987 is a bit harder because the dataset changed..

We have excellent national VMT:

But for those years, it's harder to get comparative crash reports.

Luckily, 1988 is still extremely early in cellular adoption where only about 2 million U.S. wireless connections existed, so we can then follow the same national crash series through the explosion to hundreds of millions of connections.

1. 1937-1982 -- Pre-cellular Historical baseline/context.
2. 1983 -- Cellular introduced First commercial U.S. cellular service.
3. 1984-1989 -- Early adoption Still extremely few subscribers.
4. 1990-1999 -- Rapid adoption ~5M --> ~86M connections.
5. 2000-2004 -- Mass adoption ~109M --> ~182M connections.
6. 2005-2010 -- Essentially ubiquitous ~208M --> ~296M connections.
7. 2011-present -- Smartphone era.

For the actual crash-rate calculation, however, we have to start around 1988, because that's where the nationally estimated NHTSA total-crash series begins.

That is the cleanest first-order comparison of cellular adoption versus total crash rate that the national data give us, where the trend is clearly going down, not up.

So the hypothesis that cellphone ownership rates skyrocketing caused accident rates to skyrocket, is busted, don't you think?

I don't know what the stats are but can state that some time ago I was rear-ended at a stop light by a young idiot who was texting. I have also seen young imbeciles run through red lights and stop signs while looking down at their phones. It's pretty common.

Given that most if not all states have seen fit to pass laws against hand-held phone use while driving it is likely that cell phone use has indeed increased the accident rate.

Why in the world would anyone want their dashboard to look like a smart phone? That's one of the worst automotive ideas I've heard in while. Also a smart phone interface is not something that is familiar to everyone. I wouldn't know what to do with an Android or iPhone if one were shoved in my face. (I know people who have no cell phone at all. I've been tempted to go that way myself.)

Last time I checked the average age of cars on the road in the U.S. was about 12 years. Some (like mine) are much older. Fewer people are buying new cars these days due to their cost and complexity. Making the dashboard look like a phone might well push even more people away.

My own rule is that my cell phone is OFF while driving. (And mine is truly dead to the world when it is turned off, though I could remove the battery to be really certain.) Excepting possibly a medical emergency there is nothing that can't wait.

But you can't just assume no other factor. Why would you think you could??

How can you possibly say that as mobile phones gradually became more popular other factors weren't, at the same time, reducing the accident rate?

You can't, well not unless you're stupid anyway.

I agree. Your God works in mysterious ways that even you, can't explain.

But you still believe very deeply in your staid religious views, don't you. But, where's your data?

Let's look at it. What are you claiming happens to all your missing accidents?

Maybe aliens took them?

HINT: The spike in accident rates you claim, simply never existed. You made it up.