Half-Full/Half-Empty?

[Red Arrow]

Got this from a friend.  Interesting spin for those of us who put too much faith in statistics.

Going out before the holidays? Be safe always!
Please, take care of yourself. A recent joint study conducted by the Department of Health and the Department of Motor Vehicles indicates that 23% of traffic accidents are alcohol related. This means that the remaining 77% are caused by the a$$holes who drink bottled water, Starbuck's, soda, juice, energy drinks and sh!t like that.
Therefore, beware of those who do not drink alcohol. They cause three times as many accidents. This message is sent to you by someone who is a drinker and worries about your safety.

[nathanm]

Got this from a friend.  Interesting spin for those of us who put too much faith in statisticsdon't understand statistics.

FTFY 

[Red Arrow]

FTFY 

Nice try.   Now we know your problem.

[nathanm]

Now we know your problem.

What, that I'm humorless? That should have become apparent a loooong time ago.

[Red Arrow]

What, that I'm humorless? That should have become apparent a loooong time ago.

You may be humorless but that's irrelevant to the current topic.

We might be playing a semantics game but let's try to get past that.

I think that you think that if everyone had access to the same data and were properly able to understand it that everyone would come to (nearly) the same conclusion.  In a perfect world that may be at least closer to the truth than present.  We don't live in a perfect world. 

Assuming the 23% and 77% numbers are true (I didn't check them),  it would be difficult to deny that approximately 3/4 of the accidents were caused by drivers who had not been drinking alcohol.  Moving beyond that statistic, we know that those 23% were probably mostly (maybe not all) preventable by not drinking and driving and that the overall accident rate could be lowered but it takes more data to know that.  Not taking the steps to find the extra data is where the faith in statistics that I was writing about comes from.  "These numbers don't lie."

[Conan71]

How many are caused by cell phone users or women putting on make up in the car?

[nathanm]

Assuming the 23% and 77% numbers are true (I didn't check them),  it would be difficult to deny that approximately 3/4 of the accidents were caused by drivers who had not been drinking alcohol.  Moving beyond that statistic, we know that those 23% were probably mostly (maybe not all) preventable by not drinking and driving and that the overall accident rate could be lowered but it takes more data to know that.  Not taking the steps to find the extra data is where the faith in statistics that I was writing about comes from.  "These numbers don't lie."

That's all well and good, but without knowing the prevalence of other specific causes, you can't make a supportable statement as to how dangerous driving drunk is relative to driving with other impairments or no impairments at all. To pick an extreme example, it could be that there are hundreds of other causes, all responsible for less than 1% of all crashes, but in aggregate responsible for 77% of them.

[Ed W]

A cursory search reveals that NHTSA figures for 2005 include drunk drivers involved in 39% of fatalities and 7% of all crashes.

http://www-nrd.nhtsa.dot.gov/Pubs/810616.pdf

Crash rates have declined since 2008, if I recall right.  In general, they hovered at the 40K/year mark for many years, but declined to about 32K/year.

In 1998, speed was a factor in 30% of all fatalities, second to alcohol which was at 39% that year.

http://www.smartmotorist.com/traffic-and-safety-guideline/excessive-speed-is-a-factor-in-one-third-of-all-fatal-crashes.html

Carry on.

[AquaMan]

Stats are just a base to work from. A tool. Without them you merely have supposition. But they are too often used alone to prove cause and effect. That is where they fail.

For instance, there is controversy about putting camera's at intersections to record and cite those who run red lights as that is a major cause of accidents and deaths from T-bone accidents. The controversy is two fold. Government intrusion and the potential for increased rear end accidents as people stop quickly rather than running the red lights.

Though it may be true that red light runners are involved with most accidents at intersections, the real reason for the accident is speed and distraction. By driving too fast the reaction time necessary to stop is decreased beyond the driver/vehicle ability to stop in time. Same thing holds for rear enders. That comes from speeding and tailgating.  Add to that the increasing usage of cell phones, texting and laptops by the driver and THEN you understand why the red lights are being run.

The difference is critical. When you cite people for running red lights rather than the real cause, you can't expect much success.

[nathanm]

Without statistics, you couldn't say that the crash rate goes up significantly at most intersections after red light cameras are installed. You are right that you can't stop there, though. Further investigation reveals that the number of fatalities and severe injuries decreases significantly despite the increase in crashes because more of the crashes are lower speed rear-end crashes instead of t-bones.

Edited to add: I should mention that vehicle crash statistics are notoriously bad because they are often collected with an agenda. For example, "alcohol related" does not mean that a driver was drinking or had any specific blood alcohol content. It means that somebody involved, whether it be a passenger or even a pedestrian hit by flying debris had been drinking. Similarly, "speed related" means that one of the vehicles involved was either traveling too fast for conditions or was exceeding the speed limit. In neither case does it matter whether the specified factor was anywhere in the causative chain.

[Red Arrow]

Without statistics, you couldn't say that the crash rate goes up significantly at most intersections after red light cameras are installed. You are right that you can't stop there, though. Further investigation reveals that the number of fatalities and severe injuries decreases significantly despite the increase in crashes because more of the crashes are lower speed rear-end crashes instead of t-bones.

Edited to add: I should mention that vehicle crash statistics are notoriously bad because they are often collected with an agenda. For example, "alcohol related" does not mean that a driver was drinking or had any specific blood alcohol content. It means that somebody involved, whether it be a passenger or even a pedestrian hit by flying debris had been drinking. Similarly, "speed related" means that one of the vehicles involved was either traveling too fast for conditions or was exceeding the speed limit. In neither case does it matter whether the specified factor was anywhere in the causative chain.

Cars are a LOT safer than in years gone by.  I recently found some pictures of automobile accidents that my dad took in the 1950s.  You'd be surprised how much those old car disintegrated on impact.

Data with an agenda is a problem too.

[Hoss]

Cars are a LOT safer than in years gone by.  I recently found some pictures of automobile accidents that my dad took in the 1950s.  You'd be surprised how much those old car disintegrated on impact.

Data with an agenda is a problem too.

I call that the 'Rasmussen' effect.

[Teatownclown]

Have a great 2012! This vid resembles 2011...

[nathanm]

Cars are a LOT safer than in years gone by.  I recently found some pictures of automobile accidents that my dad took in the 1950s.  You'd be surprised how much those old car disintegrated on impact.

[AquaMan]

One of my uncle's had a salvage, tow truck operation along the infamous Hwy 33 north of here back in the 50's-60's. We became well aware of how badly designed many of those old cars were as we played in them for years. Another misleading stat about drunk drivers involved with fatal accidents is that the drunks are often the ones who survive them. My uncle said that because of their relaxed state and diminished reflexes drunks would bounce around and in effect wrap themselves around the vehicle as it disintegrated. It was the instinctive tension of the sober driver whose muscles contracted when bracing for impact who ended up with broken bones instead of bruises. Not sure how true that was but it was his observation.