[flickr]photo:5923522776[/flickr]
This photo of a car elevated in a brick wall North Avenue and Kedzie Avenue by Katherine Hodges is not related to the story below.
A comment was left on EveryBlock, in response to a crash at Lincoln and Fullerton, “What a shock, alcohol was involved” (here’s a newspaper’s report). I presume that many other people think alcohol is typically a cause or factor in automobile crashes. I looked at the data to know if it’s true.
Crashes
From 2007 to 2010, there were 394,651 reported crashes. Of those, responding police officers marked on the crash reports (SR-1050) that being under the influence of alcohol or drugs was a primary or second cause in only 3,647 crashes, or 0.924%. However, this does not tell the full story. That “cause code” (#8) is to be used when an arrest is made. When an arrest is not made, officers are to use “had been drinking” (#19), from which the data shows 1,030 crashes. Adding them together, you have 4,677 crashes. In other words, alcohol is a contributing primary or secondary cause in 1.185% of crashes.
What is the most frequent cause? The top five causes (combining primary and secondary), excluding “NA” (the most frequent), are:
- Unable to Determine – 168,551 crashes
- Following Too Closely – 46,271
- Failing to Yield Right of Way – 46,112
- Failing to Reduce Speed to Avoid Crash – 36,991
- Driving Skills/Knowledge/Experience – 29,325
Under influence of alcohol or drugs is #16, and Had been drinking is #24.
Again, we’re not done with the story as people can still be drinking and driving without it being listed as a “primary” or “secondary” cause of the crash. I’m just less likely to be sure that it was a factor in the crash (and I’m assuming that all police officers responding to crashes have been trained well, equally, and are applying their training accurately and consistently – not likely the case).
People
In the data of 394,651 crashes, there were 1,041,378 drivers. The top “driver’s apparent condition” attributes are:
- Normal – 80,9283 drivers
- Other/unknown – 20,9015
- Impaired with alcohol – 12,445
- Had been drinking – 3,277 (this attribute has the same name as the one applied to crashes but this one applies to one of the drivers in the crash and may not related to the Crash Cause)
- Asleep/fainted – 2,310
Impaired with drugs is #8, attributed to 1,348 drivers.
Anti-drunk driving campaigns
If the findings tempt you to feel that we should be spending less attention on safety campaigns to reduce the frequency of driving while drunk, hold on. I’m not ready to make the judgment about the level of attention being paid to drunk driving. I don’t know the trend of drunk driving. Has it gone down? Are the strategies we have in place reducing the proliferance of drunk driving? If that’s true, we should probably keep them going.
But it’s clear that there are more kinds of crash causes than are being addressed. Look at the fifth most frequently cited cause, Driving Skills/Knowledge/Experience. Perhaps that tells us we should expend more resources on shoring up high school driver’s ed programs, instituting mobility education, or having driver’s ed students be in a probationary period longer (the Chicago Tribune had an excellent investigation into this in 2006 that helped lead to a graduated license program in Illinois).
The next step
It should be a shock that alcohol was involved: its presence or attributable cause in car crashes is rare. But this does not discount that crashes having alcohol present may be more severe than crashes without alcohol. That’s the next step in this data analysis. I will also be looking into alcohol presence in crashes with people who were walking or biking, or riding or driving buses.
If you are reading “Traffic” with me, then you probably understand that a driver’s own behavior is only one (measly) factor in the quality of their driving. Other factors include humans’ inability to process what we’re seeing at the speeds we’re seeing them, or think we’re seeing them (meaning we pass when we shouldn’t); the design of roadways; and a lack of cooperation when changing lanes. And that’s just in the first 90 pages of the book.
Where can different road designs play into reducing the number of crashes? And how we can address speeding (or that humans do a terrible job of estimating their own and others’ speeds)?
Note: I know I write a lot about driving on Grid Chicago, and that our mission is to promote active and sustainable transportation. The overwhelming presence of cars on the road, and the injuries this leads to, is a major barrier to progressing our mission. As long as people using active and sustainable transportation continue to receive preventable injuries in crashes with automobiles, I will keep writing about this.
If you would like to visualize this information, I’d be happy to help you and publish the results. Contact me steven @ gridchicago.com.
Addenda
(1) From the Chicago Sun-Times: “The [Illinois] department [of Transportation] released data showing a steady decline in such deaths in Illinois from 2006 through 2010. In 2006, there were 446 deaths related to impaired driving on Illinois roadways. By 2010, there were 298 deaths in crashes involving alcohol.”
I’ll look into the rate for the City of Chicago to see what happened.
(2) Jason noted in the comments the extreme hit and run rate in Chicago. Gabe Klein noted this in one or two interviews (and this isn’t the first time we’ve written about it). What’s the rate? From 2007 to 2010, 28.93% of crashes were reported by the police as “hit and run”. Here is a breakdown for each of those years:
- 2007 – 32,656 (27.18% of 120,163 crashes)
- 2008 – 29,994 (26.85% of 111,701)
- 2009 – 26,140 (31.89% of 81,982)
- 2010 – 25,370 (31.40% of 80,805)
Updated December 21 to add data about hit and run crashes; December 24 to add Sun-Times story about decreasing deaths statewide.
These are known causes, but Chicago also has an extraordinarily high number of hit-and-runs. It requires a presupposition, but I would venture to guess that the true proportion of DWI is in reality much higher.
Give me a few hours while I calculate that metric.
I think one of the most shocking statistics in the CDOT Pedestrian Crash report from this year is that 40% of all crashes causing the death of a pedestrian in Chicago ends with the driver speeding away from the scene, versus 20% nationally. Seriously, what does this say about Chicagoans? Here’s some of the report:
-Hit and run crashes account for 40% of fatal crashes in Chicago versus 20% nationally
-hit and run crashes account for 33% of overall pedestrian crashes in Chicago
-Pedestrian injuries and fatalities in hit and run crashes average out to two per day.
I updated the post to include the overall hit and run rate. It dipped from 2007 to 2008, but has increased each year from 2008 to 2009 to 2010.
Two friends of mine were victims of a hit and run crash several years ago. They were walking across the street legally – with the light, in a crosswalk – around 9 p.m. in clear visibility. A turning driver hit them and kept going. One of them had minor injuries and was treated and released from the ER. The other spent a night in the hospital for observation. The driver was never caught.
The increasing hit and run rate is very disturbing.
I was witness to the “aftermath” of a hit and run in Pilsen, at 18th and Halsted. I arrived about 5 minutes after a woman was hit by a driver who didn’t stop. I saw her body lying in the street, covered in a white blanket.
I spent the next several days imagining if I would have been able to do anything (like be a witness to the crash, or follow the driver) if I had gotten there earlier.
Two friends of mine were victims of a hit and run crash several years ago. They were walking across the street legally – with the light, in a crosswalk – around 9 p.m. in clear visibility. A turning driver hit them and kept going. One of them had minor injuries and was treated and released from the ER. The other spent a night in the hospital for observation. The driver was never caught.
The increasing hit and run rate is very disturbing.
If you correlate time of day with cause of accident, it skews the picture a bit. In the 12 years that my husband has been a Chicago police officer, he’s worked an 11 p.m. to 7 a.m. shift and a 5 p.m. to 2 a.m. shift. He’s responded to hundreds of crashes in those years. After midnight, the percentage of crashes he’s handled that were related to falling asleep at the wheel, being drunk and/or being drugged has been much higher than what he might see at 6 p.m. or 9 p.m. It gets more extreme from 2 a.m. to 5 a.m. The same districts have few daytime crashes involving drowsy, drunk or drugged drivers.
Speaking of DUIs, here’s one he responded to last night.
He’s told me about a number of incidents where he encountered a vehicle stopped at a stoplight or even in an intersection through multiple light cycles between midnight and 4 a.m. When he and his partner approached the vehicle, they found the driver asleep at the wheel – AND drunk, and the vehicle was in gear with the sleeping guy’s foot on the brake. In one case, it was a whole van full of guys and they were all drunk and asleep. It’s a little extra challenge to either wake up a front seat passenger and enlist their help, or reach into the vehicle, to put it into park without waking and startling the driver. Fortunately, most of the incidents were not crashes, but they could have been bad ones.
There are lots of other times when he sees a driver whose driving shows strong signs of impairment, and he stops them before they have an accident. Many of those drivers are drowsy, drunk and/or drugged. And there are certainly others who manage to escape detection. Those are the “almosts” you don’t hear about, because crash statistics don’t capture those numbers.
Sometimes a drowsy driver may fail to reduce speed or fail to yield, resulting in a crash. Or a drowsy and slightly intoxicated driver (less than 0.08 BAC) may make one or more errors that cause a crash. Unless it’s a very serious crash that requires the Major Accidents Unit, how a traffic report is written may involve only one officer’s professional judgment and observation of details. There may be multiple factors involved, and weighing them is affected by judgment and experience. One officer may view the circumstances of a crash slightly differently than another, have better luck in getting details from the driver and/or witnesses, or notice details another might miss. Accident reports are not created equal.
The moral of the story for safe cycling and walking – if you’re out late at night, especially after midnight, assume that the percentage of impaired drivers may be significantly higher than in daytime or early evening hours and use extra caution.
That DUI response is the one I link to at the top, via EveryBlock.
Your comment gives me a lot of ideas for different ways to look at, and analyze, the data I have. There is definitely enough information to see the distribution of alcohol or drug impairment over the day.
This is the first time I’ve heard a story about a driver being asleep, stopped, with the vehicle in gear, but the sleeping driver’s foot is on the brake. That sounds like a scary situation to be in. Maybe those blocks that are placed under plane wheels can be placed at the car’s wheels to prevent the car from moving forward (if those blocks would indeed prevent such movement).
I’d be curious to see how the reported stats look when analyzed by time of day.
Hadn’t thought of wheel chocks in the sleeping driver context. They might help if the driver merely took his foot off the brake, but if he was startled enough to hit the gas, the results might be, um, a little too interesting.
I think you’d probably have to know a cop who works night (especially in a rough neighborhood) to hear stories like that one. It’s not the kind of thing that gets media coverage.
BTW, in the first comment on EveryBlock about the CPD SUV blocking traffic at the scene, the officer in that vehicle was Scott.
I’d be curious to see how the reported stats look when analyzed by time of day.
Hadn’t thought of wheel chocks in the sleeping driver context. They might help if the driver merely took his foot off the brake, but if he was startled enough to hit the gas, the results might be, um, a little too interesting.
I think you’d probably have to know a cop who works night (especially in a rough neighborhood) to hear stories like that one. It’s not the kind of thing that gets media coverage.
BTW, in the first comment on EveryBlock about the CPD SUV blocking traffic at the scene, the officer in that vehicle was Scott.
Interesting story and comments.
Any questions about the data?