6.14.2005

Statistics

This article from USA Today about traffic safety and speeding is yet another example of how people bend statistics to meet their needs. Let's break this down. First, the allegation. Or, really, the hypothesis.

"Motorists in most states get a free pass from police for driving 5 to 10 mph above posted speed limits, a policy that contributes to 'carnage' on the nation's roadways, an association that represents state highway safety agencies reported Monday."

Boy, they don't waste any time getting right to the meat of the matter do they? Let's see. What they're saying is, letting people drive 5 to 10 miles over the speed limit causes 'carnage'. You would assume that this would mean that they have some sort of causal evidence linking these two events. Read on.

"The number of speeding-related deaths is not declining despite major safety improvements in vehicles in recent years... Speeding is a major factor in about one-third of the 42,000 traffic deaths a year in the USA."

Since I don't know any better, let's assume all the numbers are as they say they are. It's possible they are fudging the numbers or some calculations, but that's okay. We'll give them that and take their stats at face value. So, okay the number hasn't decreased.

But how is this measured? Per capita, or percentage? Or straight numbers? Because if we are just saying that 100 people died 10 years ago and 100 people died last year, that's not apples to apples. In 10 years, the population has increased, and, you'd assume, the number of people driving has also increased. Therefore, we're actually having fewer accidents, per driver. That's a decrease, right?

Additionally, this doesn't take into account any population oddities, like, say, an increase of elderly drivers as the "baby boomer" generation plows headlong into old age and all the accompanying driving foilbles, like not being able to see or control your vehicle or remember who or where you are. That might have an affect. Just supposing.

Also, note the use of the phrase "major factor". Not "cause", but "factor". This means that they aren't even stating that speeding caused these deaths, only that it played a part. Of course, so did people, driving, asphalt, tires and cars. But seriously, if you consider that pretty much everyone who is driving is probably speeding, especially if you are going to consider even 5-10 mph over, I'm surprised the number isn't higher. I mean, if these were highway deaths, you'd figure the number would be like 90%.

So, what is a traffic death? I don't know. And neither do you. It's not spelled out. Does that include pedestrians? Parking lots or just streets? Clearly, it includes city and neighborhood driving, not just highways.

How many is 42000? Is that a lot of deaths considering how often we drive? Let's see. 42000 is 0.014% of the population. More people die of nephritis (I had to look it up; it's some sort of kidney imflammation) or influenza/pneumonia (65000+) than on the highways. That's funny, I never hear about the carnage of influenza. Or even Heart Disease (nearly 700,000).

Look, even one death is a tragedy. Probably. But the fact of the matter is, death is part of the game. If you are willing to get behind the wheel of a car, you are accepting the fact that there may be a little death waiting for you. It can happen and it's the risk we take for getting across town in an hour instead of a day.

In 2003, about 17000 accidents (40%) were alcohol-related (note, that doesn't even include drugs). (Did you know Rhode Island and Hawaii top the charts in DUI deaths, per accident? Crazy.) How many of these included speeding as well? It seems to me that someone who's blazingly drunk and not in control of his vehicle is probably also speeding.

Or, maybe not. But who can say? Without getting the full picture, we're already starting to draw conclusions that may not be warranted.

But enough about the generality of the numbers. Let's continue. Here's the part that kills me.

"The nation's traffic fatality rate last year was a record low of 1.46 deaths per 100 million vehicle miles traveled... But the number of people killed in accidents each year has remained fairly constant as the number of vehicles and miles driven increased."

Don't you see! The number of fatalities isn't decreasing! That constitutes a serious problem! (The spokesman even uses 'carnage' again.) Liar.

Look, blockhead, fatalities did go down. To say otherwise is pure, misleading bullshit. Your statement isn't untrue, but it's completely disingenuous. Considering the amount of time we spend on the road as a culture, we're at a record low for deaths. And this is after the nationwide 55mph speeding restriction was lifted. Figure that one out.

"The safety report found that police in 42 states routinely let drivers exceed speed limits."

Yes, in 42 states, police admitted to not wasting their time ticketing otherwise law-abiding citizens who are merely headed to work to pay their taxes, at 5 to 10 mph over the posted limits (except Atlanta, where, and I speak from experience, 85 is dandy, just don't hit nobody), and are instead concentrating on the morons who rocket their little crackerjack motorcycles between lanes at 100+mph or the cretins who are weaving down the road after spending all evening at Happy Jacks Pub and Grill.

Yes, if the cop is busy on the side of the road writing me a ticket for 68, he's not bagging the teens dragging down the express lane of 271. This is a side argument, so I'll stop there. But you understand what I mean, I'm sure.

So, now to the part where they link those 68'ers and 71'ers to the 'carnage'. Here's where they show that, of the 42000 deaths 38% were caused by folks going 5-10 mph over the speed limit.

""

No? Okay, then here's where they show that, in the states that don't let us scrape by, traffic accidents are fewer.

""

Nothing? Right, nothing. What we have here is a failure to prove our hypothesis. We have, approximately, the following.

42000 people died in 2003 in traffic accidents.

About 1/3rd of these were speeding-related.

Most states don't bother ticketing folks who are barely breaking the speed limit.

Let's see, did I leave anything out? I actually originally put in "Most people drive 5-10mph over the speed limit," but I realized, they didn't even show that. That's just an assumption I made by (a) reading this article and (b) driving.

Oh, yeah, also this one:

Traffic fatalities, on a per mile basis, are at a record low.

So, we loosened the speeding restrictions and the fatalities have gone down. That sorta puts a dent in that hypothesis. I suppose you could still argue that restricting speeding would further reduce fatalities, but I could also point to the era that they were reduced and fatalities were higher than they were right now. So, they'd need to sell that with a pretty good argument.

But they don't have one. We have four facts, none of which are interconnected.

Which is fine, except for this idiotic drive to force a connection. And to what end? I believe that they probably do want to save lives. There's the restrictionist person that believes that more laws are better because they help people control themselves. But we know that just isn't true. And, apparently in this case, I have the numbers to prove it.

But the cynic in me realizes that a major income source for state governments is traffic tickets. Yes, it's an actual income source on the budget, one the states are pretty reliant on. In fact, if the police were to stop giving tickets, they could do serious damage to the economy of that state.

And the cynic would also like to point out that this is not the National Safety Council or something like that putting out the report. It is, in fact, the Governors Highway Safety Association, which, I assume, is something the governors (of actual states) have sponsered. It is not false to say that increasing the number of traffic tickets by ticketing for what is normally considered the 'grace point' on the speedometer would also increase the income to increasingly cash-strapped state governments.

I mean, it just would.

So, where does that leave us? Once again, a specious argument, based on unsound wrangling of mathematics, ostensibly for 'our protection' (isn't it always?), but probably really just about the cash (isn't it always?).

I'd really like to see the report, just to see what it really says, versus what they are taking from it for their own ends.

posted by risser at 14:49

Nobody has anything to say.