Wired has an article about cell phones being just as dangerous as alcohol, when it comes to traffic accidents. That's pretty scary, when you consider the fact I see people talking on the phone on every trip I make. I only see drunk drivers one in every two trips.
Luckily, it's all bullshit.
In the US, total highway fatalities have remained essentially the same, maybe declined slightly since 1982, even while cell phone use has grown dramatically.
Meanwhile, alcohol is responsible for about half of those fatalities, about 20,000 per year.
If cell phones are just as deadly, why hasn't the number of highway fatalities increased by 50%? Why haven't highway fatalities increased AT ALL?
Maybe they meant some new type of dangerous, one where no one actually gets hurt.
Good points from Josh.
ReplyDeleteNon-fatal accidents have remained about the same for the last 14 years.
Other factors making driving safer might be masking the dangers of driving while talking over this period.
But I would expect those safety advancements to affect alcohol and chatting roughly the same.
I don't think Alcohol and Phones are going to effect driving in the same way. Anecdotally, I've used the phone while driving, as I suspect we all have, and while I'm distracted from dangerous situations, I'm usually able to drop the phone, focus my attention, and narrowly avoid an accident. A drunk person on the other hand can't shift attention and become less impaired - that's why driving drunk is illegal.
ReplyDeleteAlso, you said that the number of accident fatalities has stayed close to the same, but is that a per-capita/per-car measure? Or just the total number? In addition to car safety levels changing, the number of cars on the road has also increased significantly since 1982.
(a product of cars not only becoming safer, but also becoming cheaper and more reliable. Cars today sure are amazing, we should have a post on that.)
Here we go,
ReplyDeletein
1990 There were 193M vehicles, and 6.4M accidents on the highway, and 44k deaths due to driving, 51% of which were alcohol related.
in 2002 there were 234M vehicles, 6.3M accidents on the highway, 42k deaths due to driving, 41% alcohol related.
So, some numbers:
In 1990 there was 1 accident for every 30 cars. (right?) And in 2002, there was 1 accident for every 37 cars. (3.3% -> 2.6%) So Accidents per capita are down .7% (I think?)
All of this seems to support Tom's suppositions that cell-phones haven't increased the number of accident fatalities.
However, in the data that tom first linked to (U.S. Driving fatalities, total and alcohol related) we find some rather damning evidence:
While the number of highway fatalities has hovered around 41k, the percentage that alcohol was involved in has decreased significantly, about 20%. In 1982, 60% of driving fatalities were alcohol related, in 1992 it was 47%, and in 2004 it hit a low of 39%.
So what's causing the other 60% of accidents? Without more knowledge of statistics, and a more robust dataset, it's probably hard to say.
But it looks to me like that alcohol related fatalities per accident have decreased faster than non-alcohol related fatalities per accident, which means that non-alcohol related factors are becoming more salient as predictors of accident likelyhood. Correct?
(sorry for the long post.)
I should've emphasized the decline in alcohol related fatalities in the data.
ReplyDeleteBut (1) This decline is attributable to other factors, and (2) even while substantial, the decline is far too small to account for the number of fatalities we should expect, if talking is just as dangerous as drinking.
1. Decline Attributable to Other Factors
The biggest indicator for alcohol related fatalities is per capita alcohol consumption.
Per capita alcohol consumption has been sharply declining in the last couple decades.
Even if cell phones were responsible for that full 20% decline in alcohol related fatalities, that's still a pretty small impact compared to what we should expect.
2. The numbers we should expect
Let's say we talk while driving merely twice as often as we drive drunk. That being the case, if alcohol and phones are just as dangerous, we should expect 33,000 yearly fatalities to be caused by talking while driving.
We only had 43,000 highway fatalities last year, and 17,000 were already caused by alcohol. 33,000 would be 77% of highway fatalities.
And what if the ratio of talking to driving is more like I expect, 5 or 10 times or more?
Up to 170,000 highway fatalities (if not far more) should be caused each year by cell phone use. Highway fatalities should have skyrocketed.
Surely adding a source of driver distraction has some effect. But saying cell phones use is just like being drunk turns out to be a really extreme claim, and I find it highly unlikely.
I agree, talking on the phone is definitely not as dangerous as drinking and driving. The data seems quite clear. However, I do suggest that something is keeping highway fatality numbers higher than would be expected with the observed declines in alcohol consumption and alcohol related fatalities.
ReplyDeleteIn the study you referenced, per capita alcohol consumption changed very little between the years 1992 and 1996, and actually went up in 1996 while fatalities that same year reached an all time low. I suspect that if the researchers introduced an "air-bag saturation" variable they would find the measures of corelation between "per capita consumption" and "Drink Driving fatalities" significantly weakened.
Another significant thing to notice is the way that drinking remained steady in the mid 90s, but impairment violations went down. While this might be a result of police negligence, I suspect that it's actually harsher penalties, and changing social pressures reducing the number of drunk drivers.
So, the question remains, if cars are safer, and fewer people are driving drunk, why haven't traffic fatalities decreased more significantly? Cell phone distraction sounds like a good culprit to me, that or ennui.
The variety of options related to car cellphone usage also causes problems. Do the various hands-free solutions still cause the same problems? Clearly using the handset is more troublesome while driving a manual transmission. It's not as clear-cut of a condition as blood-alcohol level.
ReplyDeleteActually, it's basically the same problem as eating/drinking while driving. I wonder if that would be outlawed as well.
I think Lester's right. I'm not sure what variance is to be expected with the data, so it's hard to be sure, but it's plausible something else is propping up fatalities through the mid 1990s, and it's plausible that other factor is cell phones.
ReplyDeleteOn the other hand, this increase might be due to other factors. When was the national speed limit repealed? 1995?
Since PL points out that cell phones are clearly disadvantageous in some situations, my instinct is that we compensate when driving while talking by driving more slowly and attentively than we would otherwise. That conforms to my experience: when people have the phone to their ear, it seems more likely I'll see them going 30 in a 45 than 45 in a 30.
Good discussion.
http://www.infoplease.com/ipa/A0005118.html
ReplyDeleteThis graph might add to your discussion. Perhaps cell phone related accidents come under the Improper Driving heading.