Last Call for Drunk-Driving Posts

Time to wrap up this series on drunk driving, alcohol abuse, and measures we could take to better control them if we so desired. With help from authors Peter Anderson, Dan Chisholm, and Daniela C. Fuhr, I’ve written about measures like programs to persuade people to moderate their liquor consumption (not much evidence of effectiveness), to cut back on the times and places one can buy liquor (more promising), and to increase taxes on alcohol (would definitely work, but perhaps not as effectively as we would hope).

However, as I have noted, all of these measures would affect (perhaps “punish,” depending on your point of view) responsible drinkers, not just the problem cases likely to indulge in crime, domestic abuse or drunk driving.

Thus it may be more fair and perhaps more effective to target troublesome behaviors directly. What potential measures might work for DUI?

Since other forms of persuasion seem to have limited effectiveness, it should not come as a big surprise that designated driver programs have shown equivocal results (see this literature review from Susan M. Ditter and colleagues). Advertising, and programs conducted by drinking establishments (e.g. rewards like free soft drinks for designated drivers), have demonstrated a positive, but quite modest, effect. More study is needed before we can conclude that designated driver programs really work.

Education and persuasion have been shown to be more powerful when backed up by the muscle of the criminal justice system. Over 150 studies have looked at whether court-mandated treatment programs for convicted drunk drivers are effective (see this from E. Wells-Parker and colleagues). The evidence indicates they are. Those who complete such programs get into about 9 percent fewer alcohol-related crashes, and this figure probably understates the effect. Education, therapy/counseling, and probation programs are the most effective.

What else might the justice system do? As I’ve written, reducing the permissible blood alcohol content for driving would probably be quite effective; research has shown that cutting the limit to .05 (it is now .08) would probably save thousands of lives per year. If this is too radical a step, some studies have found that a lower BAC limit just for younger drivers would be have benefits. (I’ll take this up some other time when I consider the troubling issue of safety and young drivers.)

Of course, any BAC limit will only be as effective as its enforcement, and as we all know, this is currently spotty at best. According to Paul Zador, Sheila Krawchuk, and B. Moore, the average person who is caught driving drunk has already gotten away with it 87 times.

What to do about this? Ruth A. Shults and colleagues looked at the effectiveness of one aggressive measure: sobriety checkpoints.

Checkpoints work both by getting drunk drivers off the road and by deterring would-be offenders. There is strong evidence (from 23 studies) that they are effective, particularly when backed up by media campaigns. The median selective breath testing program covered by Shults caused a 20 percent reduction in alcohol-related crashes in its community. If those results could be replicated nationwide, thousands of lives would be saved annually.

Three studies have done a benefit-cost analysis on breath testing, comparing savings from crashes avoided to the costs of the programs; the results vary pretty widely, but all show a high level of net benefit.

The Supreme Court has ruled checkpoints constitutional, but then again tax audits and root canals are constitutional too, and that doesn’t make them any more pleasant. Unquestionably, sobriety checkpoints are an intrusion on our privacy, particularly because they give police the opportunity to check for other violations. Is this worth the lives saved?

A policy that is gaining in popularity and has fewer potential privacy issues is ignition interlock devices (IIDs). These are cell-phone sized gadgets into which drivers must blow in order to start their vehicles. A BAC over the pre-programmed limit renders the vehicles inoperable. You have to blow periodically to prevent you from having a sober “friend” start your car for you. The devices cost $70-150 to install and about $60-80 per month to maintain, at the driver’s expense. (For more see this from Mothers Against Drunk Driving.)

Thirteen states currently mandate IIDs for first-time offenders at the .08 level. Nationwide, one in 10 convicted drunk drivers currently has an IID.

Given that 65 percent of DUI violators who don’t get IIDs continue to drive after their licenses are suspended, IIDs sound like a pretty reasonable idea — provided they work.

The good news is that studies show that offenders who have IIDs really do stop driving drunk; one randomized trial showed their re-arrest rate is only about a third of those who don’t get the devices (see this from Charlene Willis, Sean Lybrand, and Nicholas Bellamy). The bad news is that when the IIDs are removed, drivers quickly revert to their old ways and start driving drunk again as if they never had the device.

Does this mean we might consider mandating IID use permanently, as least for the most serious repeat offenders? Or, taking this to its logical conclusion, might we someday make IIDs mandatory for all?

This might strike many of you as a level of surveillance that could only be devised by the mind of Dear Leader, but after all, zero information besides your BAC is collected, and even that is not distributed to others. Indeed, robot enforcement of this type – red light cameras, about which I’ll write another time, is another – may actually be far less intrusive than justice meted out by humans, and has the potential to be less arbitrary as well. Besides, many of us have voluntarily given Google or Yahoo a vastly better opportunity to pry into our most closely guarded secrets than a humble IID could ever aspire to.

In any event, long before we get to IIDs for everybody we will probably have implemented the solution that will virtually banish DUI with zero privacy intrusion while allowing us to drink and drive as much as we want. Oh, by the way, it will work wonders for our congestion problem and do a great deal to eliminate unfair disparities in mobility for our disadvantaged citizens. This policy gets very little play in our public discourse, but hopefully in the next post I’ll be able to enlist your support.

Eric J. Taylor

Enjoyable post. One addition: though the US Supreme Court has ruled sobriety checkpoints constitutional, many state courts have ruled that they violate state constitutional rights.

Zachary Bloom

As has been the case for decades, too many people attempt to approach this problem being overly intelectual and trying to figure out "why people drink and drive," as if it's an obscure genetic disposition hidden within thousands of years of evolution. People drink and drive because people drive... everywhere.

Well, one company I've seen is solving this problem the intelligent way: Zingo -

Watch this interview:

By providing a value added service they have solved the problem logically and in a way that allows the customer to still have full control of and access to their vehicle. After all, for most commuters, a car is like a changing room, closet, and dinning room all in one. Who wouldn't want to have the option of hiring a one time chauffeur at the end of a long evening of merriment?

Mechanical devices and calling cabs and all these silly overpriced options don't make any sense because they leave you in the exact same predicament as before you got in your car to go drinking... if I don't drive my car at all, it's expensive. If I drive my car there and drink too much, then I have to leave my car at my destination. The problem isn't solved unless you account for people's dynamic behavior and infinite possible reasons for how and why one is left at the end of the night with a 20 mile drive back to the suburbs and, often, an even more toasted companion in the front seat.

Politicians are worried about overstepping social boundaries, while Entrepreneurs will solve these problems.

More like this, please.



Some drunk driving enthusiasts also like marijuana, which would be much safer for us and them than their using alcohol. The black market price of marijuana is a very effective way of forcing reckless drunks who also like cannabis to use mostly alcohol

Vivek Nemana

Why do the IIDs cost $60 monthly to operate? An economic deterrent if I've ever seen one, and something that will definitely be an obstacle in spreading their use. Is there anyway that better IID technology could make it all a one-time fee, and maybe the occasional repair cost? It's a great device, but it needs to be cheaper.


IID's for everybody? Are you suggesting we all be charged the $70-150 to install and about $60-80 per month to maintain, at the driver's expense regardless of whether or not we drink or have ever had a DUI arrest? What an insane waste of money and unnecessary infringement on individual liberty!
How about requiring the installation of Intoxilyzer 5000 (frequently used police station breathalyzers) type vending machines at all bars enabling drinkers who fear a DUI arrest to pay a nominal sum to check their BAC? I bet most drinkers, and even tavern owners, would not object to such a law that would benefit nearly everybody. Of course this would not curtail the irresponsible drinkers from driving while over the limit but it would certainly eliminate a vast # of law abiding citizens from driving under the influence who do not want to risk the perils --ethichal, legal or safety-- of driving while intoxicated. .05 & .08 BAC limits are way too low for most drinkers, even responsible ones, to know they are "impaired."



This might sound barbaric, simplistic and backward, but what about just increasing the penalty/cost for driving drunk?

First offense - stiff fine and short term loss ability to drive.

Second offense - stiffer fine and jail/other incarceration coupled with extended loss of ability to drive upon release. This will probably be enough to cause loss of employment, which should be an attention getter for most thinking about reoffending.

Third offense - Mega fine, lot's of time away from society and permanent loss of driver's liscense.

If at any time drunk driving is a contributing factor to property loss, injury, or worse to others - stoning in the square.

Punish those (raise the cost for the behavior) that are a problem, and leave the rest of us alone.

I really don't want to get in their head and understand what makes them do it. I just want them off the road where they can't hurt someone else (or themselves). I'm more interested in changing their behavior than in understanding them.



I think one point we are missing (bravo though on the zingo point) is that there is a sense of entitlement among people that perpetually drive over the limit. In their mind, they are fine. It isn't until an accident occurs that there is an issue.

And note, I didn't say they deem it an issue when they get caught at a check point or pulled over, because then it wasn't a problem, they just got caught. Nothing would have happened if they hadn't gotten caught, they would have gotten home just fine (or so they think until the inevitable accident).

I have to disagree with someone being a better driver high on marijuana than drunk on alcohol. Reaction times are slowed in both cases and therefore, both should be treated equally.


Better solution: draconian punishment for DUI, say 10 years in prison. You can guarantee that they won't be on the road drunk again for a decade, and the deterrent effect would be immense.


Vivek - I think that IID's for just $60 a month isn't that bad! I work for a company that sells them (Smart Start of California) and I've noticed that paying a monthly fee for an IID is much better than some of the other costs an offender would have to pay if they refuse to get one in their car (alcohol education classes, insurance, monthly bus fare, perhaps?) Financially, I think an IID works. The reason it's monthly is that you need to take it in to an autoshop once a month to get it calibrated and to collect results to help the court monitor if a driver has been drinking and violating their probation.


I'm really hoping you're going to start posting about auto-autos. I'd love if my car would drive itself to work and i could nap in the back. Arstechnica did a good series of articles about that a while back, but I'd love to hear more.

Mr. Shiny & New

@Samlbm (#9): The problem isn't the cost of the IID for convicted offenders, it's the notion that they could be mandatory for ALL cars, regardless of whether or not the driver of the car has ever driven drunk.


I like the Zingo and IID "solutions," as they seem to be less related to the emotionalism that draconian punishments seem to be based on. But also wonder if the larger problem isn't "drunk drving" but simply "bad driving" something that being drunk certainly encourages. The advent of the cell phone has further deteriorated the quality of my fellow commuter's driving and at some point (if we haven't already reached this already), other types of distracted or affected driving will replace alcohol as the greatest cause of accidents. At this point we need to start bringing a holistic perspective to the problem of traffic accidents and quit falling for the idea that if everyone just didn't drive drunk, we'd have no more traffic fatalities. That just isn't true.

As for the dangers of drunk driving, we know that they are real, but if the average person who get's caught has already done it 87 times, that's evidence of something besides enforcement. The ugly truth is that many people drive impaired from time to time during their entire lives and never get caught nor cause an accident. Its not something polite people say outloud but then again, I've never been overly polite. Are they just lucky? Perhaps. Then again, they might not think they are driving well when they have had alcohol, they might think they exact opposite and have learned to compensate. I'm not recommending this, just observing. Habits of defensive driving can adhere to people with a wide variety of driving ability. My guess is that if we brought back serious and rigorous driving programs in the high schools, and had tougher overall driving testing, the amount of deaths (both alcohol related and otherwise) would decline. Should we pay for that? We already pay for it, obviously, with the loss and suffering of many Americans who have to bury their loved ones because of something that happened on our roads. Less accidents overall should be the goal and if we have reached the point where only by the loss of freedoms by non-offenders can we reign in the repeat drunk drivers than we have gone about as far as we can and we need to go in another direction. More driver education in general, please.


anne b

All DUI drivers should have their licence and car confiscated upon arrest and until they go before the court. If a first offence a month without the car may make them think twice the second time, if a second time then a year might do the trick but if not and there is a third time then a 5 year loss of both licence and car hopefully would do it and this goes for anyone and everyone. Politicians, police, firemen Dr.s etc. and this should be a Federal law so it is consistent throughout the country and can be tracked through the system from State to State, we have the technology to do it so lets use it.

Ben D

How about a device that uses the car's computer to monitor the driver's actions and flag things like drifting, slow reaction time, etc. It would work for all kinds of impaired driving (DUI, DWI, texting, cellphone, eating, etc), and would require no active participation from the driver. Of course privacy advocates wouldn't like it, but it's no worse than the IID in that respect.


Like A, I believe the problem isn't drunk driving, but bad and dangerous driving, which is sometimes caused by alcohol, sometimes not. Get rid of all drunk driving laws, and simply radically increase the penalities/enforcement for all dangerous driving.


I've heard some concerns that breath tests are not entirely accurate. Maybe this has already been solved, but if it hasn't DUI checkpoints could cause all kinds of problems with a lot of false arrests.


More ideas along the lines of Zingo, please.

We ought to pay more attention to the root causes of drunk driving: that is, people drive their cars somewhere to get drunk, and then they can't get their cars back home without driving.

Better transportation options (public transit) could obviously help.

Wider availability of alcohol might also help. For instance, here in WA, the only place to buy liquor at night -- or on Sunday -- is in a bar, because state liquor stores have limited hours and there are no private liquor sales. If corner stores could sell liquor, perhaps that would reduce the number of people driving to and from bars.


Jeremy Clarkson, host of BBC's Top Gear, suggested this solution, of which I wholeheartedly approve:

"If you have been out and you've had some drinks, you are allowed to drive home, but only if you place a green flashing light on the roof of your car.

And here's the clever bit. If you are driving with the light flashing, then you are limited to a top speed of 10mph.

Think about it. Normal sober people will see you weaving down the road toward them, they will clock the light and they will know, whether they are on foot or in a car, that you are drunk and that they should give you a wide berth.

Because you are going so slowly, they will have plenty of time to make the necessary adjustments, and what's more, even if you do hit a bus shelter or a tree, you will cause very little damage.

Thanks, then, to the green-light system, you would get home without being raped or murdered; what's more, the next day, your car would be outside your house, and not 15 miles away outside a country pub.

There is, however, one further feature in my plan. If you are caught drink-driving with no light on the roof, you will be shot, in the head and on the spot by a police executioner. If the government is going to introduce a fair and sensible system for getting you home safely, in your car, the least you can do is play ball. So it's instant death for people who don't."


John Squire

Unbelievable. Increasing penalties for DUI punishes one behavior that increases driving risk, despite the fact that there are many other behaviors with similar, and perhaps greater risks.

The ONLY moral answer is to treat all risky behavior the same, and the only pracical way law enforcement can do that is post hoc - huge penalities for actual accidents caused by ANY risky behavior - DUI, txting, inadequate sleep, poorly maintained vehicles, etc. I'd suggest death penalty for vehicular homicide caused by driver-controlled risky behavior, felony jail for any personal injury to others (including in the driver's vehicle) and misdemeanor penalties for property-only damage.


Add to my list above the permanenet forfeiture of the vehicle involved, and perhaps all vehicles registered to the offender, and I think we are getting close.