Mother Jones - Comments for "Breath Tests Take a Hit"

breathalyzer

What part about 40 some thousand people a year dieing from DUI accidents a year would ever make you argue to take away a tool from the police, the only people trying to stop the carnage? Why do people continue to drink and drive, even if their BAC is technically .001 from .08? I think it it is telling about a society that someone would argue that they as an individual have more of a right to drive impaired, even if barely below the legal limit, than the rest of us have to drive on roads free from drunk and impaired drivers. If you want to argue technicalities, then let's set the limit at .000. Maybe that would clear it up for the "poor, innocent, mostly, but not completely" impaired drivers.

Kevin, your teetotallarity

Kevin, your teetotallarity is showing. And "anonymous physicist" explains why defense lawyers should never let a physicist/statistician/mathematician stay on the jury. Statistics are great, but they have nothing to do with the individual case. There is a one in a million chance the airplane you're in will go down, but try explaining that to passengers whose airplane is going down.

Interesting.

My fiancee recently researched an article on the subject of the inaccuracies inherent in breathalyzer tests. She told me that the test measures a specific compound that, yes, is found in alcohol, but it is a compound that is found not only in alcohol, but in other common substances as well. The reason her research stuck in my head is that one of those substances that will cause a false positive is found in asthma inhalers. I will never drive under the influence, but I keep an inhaler in my car, for obvious reasons.

Actually, using the Wikipedia breathalyzer entry as a starting point, _gasoline fumes_ can also cause a false positive. That seems highly problematic, given the circumstances under which these tests are administered.

JMA's first post got most of

JMA's first post got most of the statistics correct. All measurements have errors. There are a number of sources of errors in the breathalyzer, probably the largest one is the conversion from breath alcohol amount to BAC. Years ago, in CA, I was on a jury for a DUI case. The prosecutor put on an "expert" that explained how the breathalyzer worked. He stated that there was a 10% relative error. My guess is that the 10% was one standard deviation (SD). The expert didn't say, nor did the prosecutor or defense attorney ask.

The error is that large because of all the issues mentioned in earlier comments, gender differences, medical conditions, etc, as well as the physics of the situation. You have diffusion going on between the alveoli and the air space in the lungs. How fast the diffusion takes place is a function of physiology and the partial pressure differences between the alveoli and air space. Since air is being inhaled all the time, I doubt if you ever reach equilibrium.

Knowing the error in the measurement, it then becomes a straightforward statistics calculation to determine the probability that the true value was greater than or less than the cutoff limit. If you measured 0.08, then the probability that the true value was less is 0.5, i.e. half the time the person is not guilty. For 1 SD measurement (0.088), the probability of being not guilty is 0.158. For 2 SD (0.096), the probability is 0.022.

So what this boils down to is what is the probability level that you would call "beyond a reasonable doubt"? At 2 SD, there is 1 chance in 50 that the person was legally not drunk. Do you think this is an acceptable cutoff level?

For the case I was on, the defendant was 3 SD above the cutoff (which was 0.1 back then). This is a probability of 0.0013 that he was legally below the limit (1 chance in about 750). FYI, I voted guilty.

Note: The probabilities are the 1 sided prob values, since you don't care about the prob that the true value is greater than the measured value.

If the science is bogus, convict them anyway?

Bullshit.

There's an easy way around this. Hire a staff phlebotomist to take a damn blood sample. There's no confusion about that. Confusion arises out of an attempt to cheap out with a shortcut that doesn't really work.

Sorry, I'm not willing to punish innocent people in order to make it easy and cheap for the government. It is supposed to be hard. That's the system.

*****I say raise the limit

*****I say raise the limit so that all ambiguity is removed. Folks who are below the legal limit but still impaired, rare but very possible, could still be convicted based on behavior.*****

That's just nuts. THOUSANDS of people die on American roads each year because of alcohol impairment. Even in a large country like the US, that's a lot of unnecessary deaths. I think we should be moving to the opposite standard - ZERO tolerance toward mixing alcohol and driving. We play lip service to the notion that one shouldn't mix the two, but the fact our society tends to qualify this attitude with the modifier "drunk" -- as if people who really ARE drunk are capable of determining whether or not they're safe to drive. They're not.

We need to make it so that not only is "drunk" driving socially unacceptable, but drinking and driving period are unacceptable, socially. I've got absolutely nothing against alcohol. But if you're going to drink, take a freakin' cab. And if you are going to be driving, then have a coke.

I propose we move the BAC standard down to, say, .02%. Perhaps a fat fine and auto insurance points would be assessed from that level up to .08%, with harsher sanctions (loss of license, etc.) assessed for higher BAC levels. The simple fact is human beings who are consuming alcohol can't be trusted to determine whether or not they're safe to drive. Eventually none of this may be necessary, as driver safety sensors in cars don't allow the engine to be started by intoxicated drivers.

Rare?!?

Being below the legal limit but nevertheless impaired is "rare"? Really? It seems to me that one drink is slightly impairing, two drinks moreso, etc. I'm trying to figure out what you might mean: too impaired to drive safely, perhaps?

Won't jurors see it that way too?

If breathalyzers are "generally pretty accurate," and without having researched it in great detail I'd agree with that characterization, won't jurors be overwhelmingly likely to see it that way too? I'd rather trust a jury of the defendant's peers see through his lawyer's bullshit and convict than create an evidentiary rule that something "generally pretty accurate" is "accurate."

Read the opinion

If you read the actual opinion, http://www.courtinfo.ca.gov/opinions/documents/S157565.PDF, you find out two important bits of information:
1. This does not affect the admissibility of evidence under the 1994 decision. In a per se case, i.e., where they're charging you with having a BAC over .08%, you still cannot admit evidence concerning partition ratio variability, as it is still irrelevant because it's defined under the statute.
2. This case only concerns the generic DUI offense, that is driving "under the influence" of alcohol. Here, it may be relevant to deciding whether someone was generically driving under the influence of alcohol.

Basically, in the first case, you couldn't admit evidence concerning partition ratio variability, because it's part of the statute. In the second case, you can admit evidence, because it goes to the totality of the evidence concerning a particular individual. Overall, this is really important for cases on the edge - people who were at a .08% or so. This allows them to show that, hey, these tests which are set on pretty generic assumptions differ for that individual. The end result is that the breathalyzer says they have a higher BAC than they actually do.

This isn't a generic catchall for defense attorneys – it's a legitimate evidentiary issue that the California Supreme Court got right.

Innocent till proven guilty? Remember?

It's that Constitution thing again, getting in the way of putting bad folks in jail.

I've got an idea: just declare them "enemy combatants" and dispense with the need for all that inconvenient evidence all together.

But cops can always fall back to the old standby of DWBB (Driving While Black or Brown).

JMA is right

I think JMA has the strongest argument in this thread. It would be very simple to fix this by raising the "incontestable" breath test to two standard deviations above the legal limit.

Whether or not 0.8 is a reasonable limit is another question. I think not. It was 1.0 in my state (California) for a very long time.

I say raise the limit so that all ambiguity is removed. Folks who are below the legal limit but still impaired, rare but very possible, could still be convicted based on behavior. That would also address MaynardHandley's argument but allow the fence to be in a more reasonable place.

Who would you rather be on

Who would you rather be on the road with? A 16 yr old driver with BAC of 0.04 who is texting? Or a 40 year old guy at 0.08 with both hands on the wheel making sure he is aware of the situation. There are more important variables than BAC, and that is why random roadside stops/checkpoints are absurd and probably unconstitutional.

"Look at the difference in

"Look at the difference in anyone's life that the difference in blowing .079% vs .080% makes. "

Be very very careful with your argument here...
What exactly are you trying to argue --- legality or morality? Because most of us here are interested rather more in morality than in legality, and this is a pretty thin reed on which to hang a moral claim.

There is a Jewish concept known as "gezeirah", a "fence around the law". The idea is that, yes there are specific details in the law, but if your interest is in not breaking the law, then you don't spend your time walking 1 inch away from the law hoping that you just won't step over the line. Instead you create a fence, limiting yourself from doing things that may be legal, but which you still should not do so, so that you are assured you always stay on the good side of the law. The world (especially the US) could do with rather fewer amateur lawyers, and rather more living one's life according to gezeirah.

Or, to put it in simpler terms --- if you're trying to be a moral person (as opposed to some a--hole, a danger on the road and wasting your fellow citizens' time and tax dollars), then don't go out driving if your blood level is over say .05. Put your own fence down there, and spend rather less effort worrying about whether you are at .079 as opposed to to 0.81.

The basic logic I am seeing

The basic logic I am seeing here, by both Drum and commenters, is, quite frankly, retarded.

The existing system is ALREADY somewhat arbitrary. Presumably what one really wants in the law is something like "you can't drive if you have a certain level of impairment in skills relevant to driving". But that is an impractical standard, so we utilize a proxy, namely blood alcohol level. A defendant could argue that it is unfair that he is being prosecuted for particular blood alcohol level because he is a drunk, always has a non-zero blood alcohol level, and has learned to function just fine under those circumstances --- and he may well be correct, that he, at a level of .08, is a substantially safer driver than someone who drinks once every three years, and has a level of .07. This, like much about the law, is less than perfect, but we all seem willing to live with it.

At which point the question becomes "if we're willing to accept this level of arbitrariness, why not accept just a small bit more"? If the courts are going to be stupid about this (and make no mistake, this is an incredibly stupid and worthless decision), it's easiest enough for the legislature to fix by simply adding to the law. The arbitrary drunkenness proxy "blood alcohol level" is now augmented with a second proxy "breath alcohol level". And, sure, in various situations this is going to be unfair, blah blah. As I said, so freaking what? We already have an imperfect system, we're always going to have an imperfect system.

Or, to put it slightly differently:
If the ONLY ACCEPTABLE measurement of drunkenness is blood alcohol level, as measured by actual blood; then how do people propose that the police proceed? Surely, in that case, the ONLY POSSIBLE way they can enforce this law is to have them draw blood from every suspected drunk driver they encounter? And is that what we really want, even the various whiners up page complaining about how breathalyzers hate them?

As a practical matter this

As a practical matter this might be a pain in the ass, its true. The same could be said of the right to a trial in the first place, it is be much simpler if the police could just charge people and they were automatically found guilty, however that wouldn't be a particularly good system. Sometimes these machines are accurate sometimes they aren't and the defense should be able to argue that, since it is actually true.