Drinking and Driving for Public Safety

Last month, the Substance Abuse and Mental Health Services Administration released the results of a report that asked 124,000 adults if they had driven under the influence of alcohol in the last year.

The results were kinda' fun. Wisconsin finished with the worst results in the country, with one in four respondents having admitted to driving under the influence over the previous 12 months. The survey results led to articles like this one, proving the psychology and demographics of the state's residents to explain their risky behavior.

But as the National Motorists Association points out, Wisconsin's highway fatality rate is significantly lower than the national average. Riffing off how government typically manipulates data like this in the public health context, NMA satirically suggests a public health campaign encouraging a drink or two before getting behind the wheel.

What's not exactly clear is whether the government agency that conducted the survey defined "under the influence (and if they did, how they defined it), or if they left it up to the respondents to come up with their own definition. What does seem clear is that the state with the most drivers under the influence (or at least the state that's most honest about it) isn't exactly littering its highways with dead motorists.

In other DWI news, you might want to steer clear of San Antonio this weekend. Officials there announced this week that police will forcibly draw blood from any motorist who refuses a breath test.

Comments

[Legate Damar]

I've successfully steered clear of San Antonio for 31 years and counting. Not gonna stop now.

[Zeb]

OK, who, among those who like to drink, has NOT technically been guilty of DWI/DUI in the past year?

[Elemenope]

In other DWI news, you might want to steer clear of San Antonio this weekend. Officials there announced this week that police will forcibly draw blood from any motorist who refuses a breath test.

Honestly, are they *trying* to get a cop shot? Because attempting to hold down and draw blood from some dude's wife is liable to get his blood up for some regulatin'.

[ktc2]

They don't even think about that. They feel they can do whatever they damn well please to "little people" (ie non cops) without regard for morality, rights or law.

I'm loathe to encourage shooting cops but some of these bastards (anyone who would do the above described) deserve it.

[robc]

Zeb,

Me.

Next question?

I dont think I have ever been technically guilty of DWI/DUI in 21 years of driving.

[Elemenope]

I'm loathe to encourage shooting cops...

Me too, me too! My only point was that *they* didn't seem to mind encouraging it with such behavior.

[ed]

All sobriety and no DWI makes robc a dull boy.

[Legate Damar]

Zeb,

I haven't been technically guilty for at least 8 years. And I drink, well, not quite like a fish. But at least like some sort of baby amphibian.

[Michael Pack]

The fact they have to take blood to prove a crime has taken place tells you something.In other cases their matching fingerprint,DNA,ect to a crime scene.I guess in this case the crime scene is your body. robc,are you a no-drinker?Here in Ohio 1 beer can result in a DUI.

[SugarFree]

robc is probably correct. I've never seen anything in Louisville I would properly classify as driving.

BURN!

[Episiarch]

OK, who, among those who like to drink, has NOT technically been guilty of DWI/DUI in the past year?

Me. I absolutely will not give them the opportunity to fuck me royally. It sucks, but if you get snagged it can make your life hell. I just won't risk it. I guess their plan worked.

[B]

In the past decade, the number of impaired drivers involved in alcohol-related crashes has remained relatively stable - from 12,348 in 1996 to 12,491 in 2006. Those figures from the Department of Transportation cover drivers with a blood-alcohol level of 0.08 percent or higher."

OK...so nationally, about 15% of people surveyed own up to drinking and driving (which means more than that actually do it) but about 0.004% of the population are "impaired" by the govt's standards and involved in a crash? Even if you account for the non-drivers in the population, that's a pretty big disparity that suggests to me that driving at or just above 0.08 isn't very dangerous.

[Matt Moore]

Michael Pack - Doesn't Ohio have the same .08 limit as everywhere else? If so, you'd have to weigh less than 110 pounds and wait less than an hour to get a DUI off one drink.

[JW]

I live near DC. Until just a few months ago, *everyone* was guilty of DUI. One drink and you're sharing a room with Otis.

[Rhywun]

OK, who, among those who like to drink, has NOT technically been guilty of DWI/DUI in the past year?

Me. I don't have a car. And even if I did, I live in a city with 24 hour subways and lots of taxis.

[Matt Moore]

JW - Yeah, and that law (or that interpretation of the law) was ridiculous and tossed out.

[Kolohe]

OK, who, among those who like to drink, has NOT technically been guilty of DWI/DUI in the past year?

I dont think I have ever been technically guilty of DWI/DUI in 21 years of driving.
I haven't been technically guilty for at least 8 years.

In many (most?) states you can still be guilty of DUI even with a limit below 0.08.

This is a problem.

[Kolohe]

Blah, that should be:

"even if with a BAC below the limit of 0.08"

[kinnath]

Last weekend there was an article in the local newspaper. Seventy or so officers from multiple jurisdictions participated in a alcohol checkpoint where they stopped people on an Interstate in a rural location in Northeast Iowa -- got that -- freeway; rural Iowa.

They stopped 1000 cars on a Friday night and arrested one person (that's right one poor fucker) for DUI. They did, however, make a handful of narcotics arrests and did capture a couple of illegal aliens.

There is no safe place in America anymore.

[Naga Sadow]

Does driving a theme parks go cart under the influence count as DWI? If so, I'm boycotting Magic Mountain next time I'm in California.

[Michael Pack]

Matt,Their are 2 crimes when you arrested.One is DUI and the other is having a .08 bac or above.You can be charged even if you test .00.

[MP]

Seventy or so officers from multiple jurisdictions participated in a alcohol checkpoint where they stopped people on an Interstate in a rural location in Northeast Iowa -- got that -- freeway; rural Iowa.

I don't know if they still do this, but I remember a few years ago going through a customs roadblock on Interstate 91 in southern Vermont.

More here.

[Neu Mejican]

B,

An estimated 360 deaths are prevented each year in the United States as a result of the move from a 0.10 to 0.08 legal limit in recent years, and an additional 538 lives could be saved each year if the United States reduced the limit to 0.05, consistent with limits in most countries worldwide.

Journal of Safety Research
Volume 38, Issue 5, 2007, Pages 493-499

It all depends upon whether you view 360 deaths as an acceptable danger, of course.

During follow-up, 390 of 2,325 (16.8%) survivors were treated for injury after a new crash. The overall event rate was 34 per 1,000 subject-years. Four variables (age ?32 years, male sex, nighttime crash, and blood alcohol concentration >50 mg/dL) were identified as independent predictors of recurrent crash. After adjustment for sex, age, and nighttime, alcohol was the leading predictor (relative risk 3.73; 95% confidence interval 3.00 to 4.64). In the presence of the 4 variables, the recurrence rate was as high as 145 (117 to 175) events per 1,000 subject-years, and alcohol per se accounted for more than 75% of events. In the absence of the 4 variables, the rate was as low as 11 (7 to 17) events per 1,000 subject-years.

Annals of Emergency Medicine
Volume 46, Issue 2, August 2005, Pages 161-167

Underestimating risk can be problematic:
http://www.springerlink.com/content/91247nn1v73433h7/fulltext.html

[P Brooks]

It all depends upon whether you view 360 deaths as an acceptable danger, of course.

Or whether you believe better driver training might save more lives, with less restriction of freedom for everybody else.

[Elemenope]

Sorry, Neu Mejican, but since traffic stops are the single most common, and usually also the nastiest, interaction that most otherwise law-abiding people have with the Police, it is a context in which freedoms must be defended zealously.

"For the protection of Police" while executing traffic stops is one of many reasons why cops are now armed better than the fucking marines and have use of force doctrines which in reality boil down to "shoot first and ask questions later".

Everyone cannot be saved, but freedoms are difficult to recover once lost. I have had close losses due to drunk driving, so I'd be the first not to wish such a death on anyone; however, I still say that regulations must be balanced unfavorably against freedoms even if a tightening of those regulations might squeeze a little more safety out of them.

[herodotus]

It all depends upon whether you view 360 deaths as an acceptable danger, of course.

It all depends on which 360 people we're talking about.

[kinnath]

It all depends upon whether you view 360 deaths as an acceptable danger, of course.

From: http://www.soyouwanna.com/site/toptens/accidents/accidentsfull.html

There are so many interesting ways to die in America, that we felt it was just wrong to limit it to the most frequent causes, which are all boring diseases and infections and stuff. (Except in Alaska, where suicide generally makes it into the top ten.) You want to hear about the terrible calamities, the tragic consequences of an error in judgment or a general lack of coordination. Do we ever disappoint?

10. Machinery
Deaths per year: 350

We can thank the farmers of America for the inclusion of this particular misfortune as a cause of death. Between corn-huskers and wheat-threshers, is it a wonder? The reason it is last on the list is that there just aren't enough people in farming these days. Ironically, they have all been replaced by machines. Hmm accident, or deliberate act by wanton machinery? We may never know.

9. Medical & Surgical Complications and Misadventures
Deaths per year: 500

While we are incredibly insensitive people, we did not coin the term "medical misadventure"- the National Safety Council did. How is death by surgeon a "misadventure?" While we're not sure, we suspect that this number refers to elective surgeries that people undertake, such as liposuction. After all, the removal of a brain tumor is not usually considered to be an "adventure."

8. Poisoning by gases
Deaths per year: 700

There's nothing like the smell of napalm in the morning In this category, you mostly have deaths by carbon monoxide poisoning due to faulty operation of a heating or cooking appliance, or a standing automobile. We assume, however, that the noxious gasses emitted by Uncle Albert qualify too.

7. Firearms
Deaths per year: 1,500

We can thank our second amendment rights for all 1500 of these deaths; call it the "right to die" amendment. You probably don't want to know how many countries in the world do not even have "accidental death by firearms" on their top ten, or their top twenty. Suffice it to say that it's most of them. Of the 1500, you're looking at about 75% young males between the age of 14 and 25 (and getting younger every year), who unintentionally shoot themselves or someone else. For more information on the place of guns in society, click over to our pros and cons section.

6. Suffocation
Deaths per year: 3,300

Call this one the "Heimlich" section, as these deaths mostly resulted from blockages of the respiratory system by food or other objects.

5. Fires and burns
Deaths per year: 3,700

This would include deaths resulting from fires, such as smoke inhalation, falling beams, and sitting through Backdraft. Ironic that cancer is number two on the total deaths list, and a by-product of smoking is responsible for one of the top causes of accidental deaths. Are we getting the picture that this is a dangerous pastime? What kind of warnings do we have to put on these boxes, anyway?

4. Drowning
Deaths per year: 4,000

This includes all sorts of drownings in boat accidents and those resulting from swimming, playing in the water, falling in, or even having a bath. The human body is what, 70% water? And we begin our lives in a watery environment, there's lots of oxygen in water what's the deal? Something for the scientists to work on.

3. Poisoning by solids and liquids
Deaths per year: 8,600

These would be all your commonly recognized poisons, as well as such items as mushrooms, shellfish, drug overdoses, and problems with medicines-which is a wide category, and why it is so high on the list. What they leave out is things like food poisoning or salmonella, which they classify as "disease deaths" and place on another list.

2. Falls
Deaths per year: 14,900

Then we come to the America's Funniest Home Videos category of accidental death, including falls from ladders, down stairs, over curbs, off buses, into manholes, and through plate glass windows.

1. Motor vehicle crashes
Deaths per year: 43,200

The winner, by a ridiculously huge (and ever-increasing) margin is: death by car wreck.

= = = = =

So no, I don't see any justification to declaring nearly everyone that has one drink at dinner a ciminal if they get in a car and drive.

[MP]

It all depends upon whether you view 360 deaths as an acceptable danger, of course.

If it only saves one life.
It's for the children.
It could happen to you.

[kinnath]

I apologize for posting the whole thing, but I found the snark more amusing that the raw statistics.

[Mr. Average]

Officials there announced this week that police will forcibly draw blood from any motorist who refuses a breath test.

'Forcibly draw blood", eh? Cue images of flailing nightsticks smashing into drunken mouths and skulls. Who needs needles, anyway?

[kinnath]

Remember, when you stop stupid people from killing themselves, you are just polluting the gene pool.

[ed]

I drive drunk every day, and that's just on the way to work!
Relax. I'm an airline pilot!

[John]

"Even if you account for the non-drivers in the population, that's a pretty big disparity that suggests to me that driving at or just above 0.08 isn't very dangerous."

Bingo. You also have to figure that of the crashes that do occur with people with a BAC of .08 or above, how many of those involve people with really high BACs? A lot I would imagine. I would be interested to break the crash numbers down by BAC and see how many accidents there were involving BACs from .08 to .12 and those involving above .12. I would imagine that you would find that most of the wrecks involved .12 or higher.

If driving with a BAC of .08 were that dangerous people would stop doing it. Most people are not thrill seekers or suicidal. If the MADD scare stories were true, we wouldn't need laws to stop drinking and driving because only the stupid or the suicidal would do it.

[Neu Mejican]

John,

I would be interested to break the crash numbers down by BAC and see how many accidents there were involving BACs from .08 to .12 and those involving above .12. I would imagine that you would find that most of the wrecks involved .12 or higher.

The research supporting the 0.08 level was based on exactly that kind of breakdown. It was the 0.08 level that seemed the reasonable breaking point even though effects of alcohol can be measured down as low as 0.04.

So the findings were that most wrecks happen at above 0.08...you can make arguments for another cut-point, of course.

[Jim Bob]

I do not understand how forced blood test policies can survive if they are scrutinized next to a copy of the Constitution, with the 4th Amendment highlighted. If that isn't an unreasonable search I can't imagine what is unreasonable.

It seems the San Antonio police are saying, in effect, "Refuse this breath test and we're REALLY going to fuck with you. Come on, buddy- blow into the tube or come and get stuck with a needle. Your choice." I'm guessing most people will go with the tube.

[The Wine Commonsewer]

and the whole .08 thing that used to be .1. Now you're drunk but five years ago you weren't.

[The Wine Commonsewer]

I'm guessing most people will go with the tube.

Yeah, that scares me to death for a couple of reasons.

1. You can't get blood out of a turnip.

2. You can't find my veins. Not only would I make a bad junkie but only one in ten lab types can even get the needle in a vein. I cannot imagine how much pain I'd be in.

[Mr. Rodgers]

Can you say "Zero Tolerance". Sure you can.

[The Wine Commonsewer]

360 deaths

So lower the limit to .000000

Better yet, lower the speed limits to ten mph, that would save maybe 25,000 lives a year on it's own.

[Michael Pack]

Still,most accidents and deaths are caused by sober drivers and for a number of reasons.Most deaths due to DUI are at .15 and up.People driving at .08 are not any more likely to crash than a sober driver.High speed,sleepiness and non attentive driving are the biggest causes of accidents.

[B]

Neu Mexican,

Thank you for responding with science...you have no idea how much I appreciate that.

[kevrob]

Zeb | May 23, 2008, 1:20pm |

OK, who, among those who like to drink, has NOT technically been guilty of DWI/DUI in the past year?

I got let out of work early today (1:00ish), so I bicycled into town and stopped at a local tavern. 1 pint of Sam Adams Summer Ale, a plate of chicken wings and a basket of fries and an hour and a half later, I hied myself a few blocks to the public library. I rode half the distance, but walked the bike the other half. (There was some ugly traffic involving a gravel truck.) Would I have blown a 0.8?

Is there a Mothers Agains Drunk Bicyclists forming as we speak?

Kevin

[MP]

Better yet, lower the speed limits to ten mph, that would save maybe 25,000 lives a year on it's own.

The real answer is square tires.

[Michael Pack]

Kevin,a man in Columbus,Ohio was arrested for DUI on a bicycle..He was in his own yard.It was on Hit and Run a few weeks ago.

[B]

Naturally, the sensible way to decide whether saving ~1 life a day nationally is "worth it" is to weigh that against the cost of enforcement of the 0.08 limit. I don't have a calculator handy but I'm willing to go out on a limb and say that's a slam dunk.

But obviously most people get squeamish when you try to monetize the value of a human life. Which is one reason we end up with really stupid and paternalistic laws.

Just to be clear...I have no problem with throwing the book at someone who is clearly a danger behind the wheel due to alcohol (or anything else for that matter). I'm just not convinced that where the line is drawn now makes any real sense.

[bigbigslacker]

Neu Mejican, I bet more car wrecks occur at 0 than above .08

But that's not what you meant...

[Elemenope]

Better yet, lower the speed limits to ten mph, that would save maybe 25,000 lives a year on it's own.

And raise a bajillion dollars in ticket revenue.

[Neu Mejican]

B,

Naturally, the sensible way to decide whether saving ~1 life a day nationally is "worth it" is to weigh that against the cost of enforcement of the 0.08 limit.

I agree. But be sure you calculate that cost as the difference in cost of enforcing 0.08 vs. the cost of enforcing 0.10.

Cost has to include the false positive rate that everyone around here is concerned about.

FWIW, I think it is the risk of false-positives that have kept the US DUI criteria higher than that in other countries (0.05). Some may feel that 0.08 results in too many false positives, but I haven't seen good data on that issue since the change in the US. My sense is that if that cost were prohibitively high a public outcry would result.

[Neu Mejican]

bigbigslacker,

Neu Mejican, I bet more car wrecks occur at 0 than above .08

But that's not what you meant...

Of course...
But you knew what I meant.

[Neu Mejican]

Michael Pack,

People driving at .08 are not any more likely to crash than a sober driver.

That one screams for a citation.
Easy to assert as fact...but I don't buy it.

[Neu Mejican]

Hypothetical cost-benefit analysis:

enforcing 0.08 vs. 0.10 saves 360 lives, but results in 360 people receiving citations who would have harmed no one...

Is it worth it?
Does the benefit outweigh the cost?

Saves 360 lives, but results in 3600 such citations...

Saves 360 lives but results in 36 such citations...

Where do people think the balance point is?

[Neu Mejican]

Just to get this one off the table:

So lower the limit to .000000

From the research I have read (and we'll just use the numbers posted above)...this would save no more lives than the reduction to a BAC of 0.05 or so.

So if lowering to 0.05 saves something like 538 lives, then so would zero tolerance...unless zero tolerance led to a general ignoring of the law and increased drunk driving rates.

The science provides no justification for zero tolerance.

There are some philosophical reasons to reject any punishment for risky behavior unless and until that behavior results in an actual harm.

Between those two positions, it seems that an evidence-based cost-benefit analysis is the only way to design the policy.

For those arguing 0.10 or 0.15...show me the numbers that support your choice.

Michael Pack has asserted that they exist...surely they can be found.

[Michael Pack]

For the year 2002 in Ohio the number of deaths caused by bac.01-.10 were 0.In West Virgina,where I work,3.Almost all deaths were at .15 and above.Calif. had the highest at 47.

[Neu Mejican]

Some numbers to work with... a place to start:

"The effect size combined across all 19 locations showed statistically significant decline (p < .005) of 14.8% in the rate of drinking drivers in fatal crashes after the .08 laws were introduced. "

"It is possible to use the results of the meta-analysis to estimate the number of lives saved if all 50 states, rather than just 18 plus the District of Columbia, had a .08 law in the year 2000. To estimate the average number of fatalities associated with each drinking driver in a fatal crash, the number of fatalities within each crash (using the 2001 FARS data file) (NHTSA, 2002) were attributed proportionately to each driver involved in that crash; these proportional fatalities were then aggregated separately for nondrinking drivers (BAC = .00) and for drinking drivers. This procedure, when applied to the 32 states not included in this study, yielded 2795 fatalities that were attributed to the 3550 drivers with BAC levels from .01-.09 (a rate of 0.7875 fatalities per lower-BAC driver), and 9248 fatalities that were attributed to the 10,408 drivers with BAC levels of .10+ (a rate of 0.8885 fatalities per higher-BAC driver)."

A meta-analysis of .08 BAC laws in 19 jurisdictions in the United States
Accident Analysis & Prevention
Volume 37, Issue 1, January 2005, Pages 149-161

[Neu Mejican]

Michael Pack,

Anecdotes?
Or do those numbers have a source?

[Michael Pack]

According to MAAD there are 1,100,000,000 episodes of drinking and driving and 123,000,000 episodes of drunk driving.We should have bodies all over the roads.

[Michael Pack]

Sorry Neu it's RIDL.US.

[Neu Mejican]

Declaim.

Assert.

[yawn]

[MP]

"The effect size combined across all 19 locations showed statistically significant decline (p < .005) of 14.8% in the rate of drinking drivers in fatal crashes after the .08 laws were introduced. "

That doesn't tell you anything. You need to see a statistically significant change in the rate of fatalities across all drivers. Otherwise, all you know is that people are driving with a lower BAC, not that lives have been saved.

Of course, maybe the whole study tells you that. But you have access that apparently the rest of us don't. So throwing tidbits of large studies on the table to prove you point, without allowing for critical review, isn't particularly helpful.

Of course, neither is Michael Pack's unsubstantiated factual tidbits...

[Michael Pack]

MP ,check DUI BLOG and GETMADD for the numbers.

[Neu Mejican]

MP,

Yeah, it is a problem...but I certainly don't want to go posting the whole study in a comments section.

I was mainly posting that edit so that people can get a sense of the number of low BAC drivers involved in fatal accidents...

As for the 14.8% percent decline...

I am not sure I agree with you, but the problem of controlling for all the relevant factors is non-trivial in research like this.

[Neu Mejican]

MP,

More of the study that may get to your question...

"The number of drinking drivers in fatal crashes in those 32 states was then reduced by the effect size estimates from the composite analysis (14.8%), and then the attributable fatalities were re-calculated for these reduced numbers of drinking drivers. The "lives saved" estimate is the difference between the actual attributed fatalities and the estimated reduced attributable fatalities.

If all of the 32 states that did not have a .08 law as of 1 January 2000, had an "average" quality implementation of .08 (average being determined from the simple aggregate of the 19 jurisdictions analyzed in this study), it is estimated that 947 lives might have been saved in the year 2000. Two of the 32 states passed .08 laws in 2000. Thus, these estimates apply to only that fraction of the year 2000 during which the .08 law was not in place. Rhode Island enacted a .08 law in July, thus having the law for nearly 6 months in 2000, and Kentucky enacted a .08 law in October, thus having law for the final 3 months of 2000. This attribution method necessarily produces a smaller number than the official NHTSA number of "alcohol-involved fatalities" because, in our attribution method, a portion of the fatalities in alcohol-involved crashes are attributed to the nondrinking drivers. Of those crashes with ANY alcohol involvement, 32.5% of the fatalities were attributed to nondrinking drivers (BAC = .00). Therefore, our method produces a more conservative estimate of the lives saved than if we were to use the assumption that the fault for a crash lies entirely with the drinking driver (Voas et al., 2002 R.B. Voas, A.S. Tippetts and E.P. Taylor, The Illinois .08 law: an evaluation, J. Saf. Res. 33 (2002) (1), pp. 73-80. Article | PDF (85 K) | View Record in Scopus | Cited By in Scopus (5)Voas et al., 2002)."

[MP]

I am not sure I agree with you, but the problem of controlling for all the relevant factors is non-trivial in research like this.

I understand, and thus it makes the research non-definitive. As you've presented it, the 14.8% figure indicates that people change their DUI habits when penalties increase. That seems like an obvious consequence. It does not indicate that the change in DUI habits has thus saved lives. Only a drop in the aggregate (controlled for whatever variables you could account for) would indicate such.

[MP]

More of the study that may get to your question...

Nope...still doesn't show a reduction in fatalities. It just shows a reduction in fatalities where alcohol was a possible factor. Again, this is an obvious result from increasing penalties on DUI.

[MP]

The table on Page 21 from this (now dated) GAO review shows what I'm referring to. The table shows that while the number of alcohol related deaths declined 6.1%, the aggregate number of deaths increased 0.8%. Now, those figures aren't specific to an analysis of .08 laws, but it shows what I'm talking about. Any talk of "saving lives" needs to show actual lives saved, and not simply a reduction of driving deaths where alcohol was a factor.

There is no shortage of people who are unsafe drivers when sober. An unsafe driver when sober is also an unsafe driver when drinking. So a real study has to be able to show that drinking creates more unsafe drivers, or makes typically unsafe drivers even more unsafe. And to do that, the study needs to focus on aggregates, not simply number of people caught with booze in their system.

[Neu Mejican]

MP,

I am still not sure I agree.

Although I agree with this:

So a real study has to be able to show that drinking creates more unsafe drivers, or makes typically unsafe drivers even more unsafe.

Just to play around with this...

If there were 1000 deaths in year A and 2000 deaths in year B, but 500 of the deaths in year A involved 0.09 drivers, and only 100 of the deaths in year B involved 0.09 drivers, you have reduced the deaths involving 0.09 drivers.

So if 250 of the 500 deaths involving drivers are blamed on their drinking in year A and only 50 are blamed on drinking in year B there is a direct reduction in lives lost due to alcohol.

When the study I cite apportions blame for fatal crashes to both sober and non-sober drivers and finds that rate at which blame is assigned to non-sober drivers has decreased, they are making an argument that there is a reduction in deaths that can be assigned to non-sober drivers. Those deaths would be added to your aggregate death count in the absence of the effect, no matter whether it was overall higher or lower. It is not the aggregate rate that matters. In fact the aggregate rate needs to be accounted for very carefully.

Of course I might be missing something this late on a Friday...

[Art-P.O.G.]

2. Falls
Deaths per year: 14,900

Then we come to the America's Funniest Home Videos category of accidental death, including falls from ladders, down stairs, over curbs, off buses, into manholes, and through plate glass windows.

I laughed a little too hard at this.

[MP]

If there were 1000 deaths in year A and 2000 deaths in year B, but 500 of the deaths in year A involved 0.09 drivers, and only 100 of the deaths in year B involved 0.09 drivers, you have reduced the deaths involving 0.09 drivers.

That rate reduction may simply be explained by reducing the number of .09 drivers relative to the overall population. It does not mean that any lives were saved.

Here's my scenario. Let's say 50,000 people in a year die in auto accidents. And let's say that 5% of those are alcohol related. Now lets say that five years after the law reduces the BAC from .10 to .08, 50,000 die in an auto accident, but only 3% of those are alcohol related. Has the law been proven effective? Obviously, the change from 5% to 3% shows that less people are drinking while driving. But if 50,000 people are still dying, what have you accomplished?

Now your statement "blamed on drinking". What statistics are those? Remember, "alcohol related" statistics can't directly infer alcohol as a causitive factor. Yes, odds are that someone who blew a .2 can blame alcohol for their actions. But as the BAC goes down, it becomes increasingly difficult to infer that if the driver was stone cold sober, the accident would not have happened. That's why the aggregate numbers are so important. Only by looking at significant statistical changes in the total number of auto deaths can you see, holding other factors constant, if a BAC law has a real effect.

But the problem of course is that you can't effectively hold other factors constant. So you'd need to see a very significant change in aggregate deaths to make a convincing argument for the effectiveness of BAC laws.

[SM]

I gotta tell you guys... this thread makes me a little ashamed to be a libertarian.

I will care your personal freedom, and mine, until you are hurting someone else. There is no reason whatsoever to drive drunk. It is simply an unnecessary risk that benefits no one, including yourself. It's not your right and it's not safe.

[Anthony]

"Have you ever driven under the influence of alcohol?"

Wisconsin guy thinks to himself "I've had a beer or two at a ball game and driven home afterwards - I guess that means yes."

Alabama guy thinks to himself "I've never noticed that I drive funny after a beer or two - I guess that means no"

[Neu Mejican]

MP,

Thanks for the elaboration.
That confirms that I was understanding you correctly.

There is, of course, no path to causation in a study based on correlations (no matter the BAC). There are, however, good experimental studies showing significant impacts of alcohol on driving skill at levels well below 0.08. So the conservative blame apportionment of the study I cite is well motivated. (Converging sources of evidence exist to show an increased relative risk of accident when your blood alcohol is as low as -iirc- 0.04.).

Requiring that deaths from all causes are reduced significantly as a result of a change in BAC laws is not a valid test of the laws effectiveness, imho. You may not be able to hold all other factors constant, but you can account for the accidents that were clearly not alcohol related (sober drivers) using methods like the study I cited.

Your position is perfect as the enemy of the good, or something.

[Neu Mejican]

MP,

A bit of elaboration with a hypothetical.

Studies have demonstrated that for every sober driver on the road there is, let's say, 0.25 accidents.

For every BAC 0.05 driver there is 0.5 accidents.

For every BAC 0.08 driver there is 0.75 accidents.

For every 0.10 driver there is 0.9 accidents.

And so forth.

Your claim is that all the law does is reduce the number of drivers with BAC above 0.08...

If the accidents rates at various BAC are valid and well established, then reduction in proportion of drivers driving at a given BAC is a demonstration of the effectiveness, even if fluxuation in some other unknown cause is keeping the aggregate deaths up.

[MP]

Your claim is that all the law does is reduce the number of drivers with BAC above 0.08.

My claim is that a study that simply shows a reduction in accidents where alcohol was an indicator is not sufficient evidence to verify the effectiveness of the law. As you stated, that data requires extrapolations based on other evidence of driver effectiveness at different BAC rates.

I'll take this up again on another DUI thread. In the meantime, I'll do some more research and maybe read chapter 7 of this book.

[Neu Mejican]

MP,

Nice talking to ya.

Your comments forced me to think more carefully about the topic than I would have otherwise.

[wyzbok]

If Texas ever needed an enema, San Antonio is where they would put the hose.