32

Well-renowned security expert Bruce Schneier claims that

In the months after 9/11, so many people chose to drive instead of fly that the resulting deaths dwarfed the deaths from the terrorist attack itself, because cars are much more dangerous than airplanes.

—Our Newfound Fear of Risk

Is this backed up by hard data anywhere?

Sklivvz
  • 78,578
  • 29
  • 321
  • 428
  • 2
    http://en.wikipedia.org/wiki/Micromort It's much more risky to travel by car than by plane. 1 micromort is 340 miles with car, and 1000 miles with plane. I can maybe dig up statistics on dying per trip too, plane vs car. – Wertilq Sep 03 '13 at 16:58
  • @Wertilq That Wikipedia page lists 230 miles in a car per micromort. – Ladadadada Sep 03 '13 at 17:16
  • Hm. How uncharacteristic for Bruce to not back up a substantial claim like this. I’ve sent him a tweet asking for clarifications. – Konrad Rudolph Sep 04 '13 at 07:14
  • 5
    He clarified now that it should have been years instead of months (https://www.schneier.com/blog/archives/2013/09/excess_automobi.html). He also added some sources for the claim in this later post. – Mad Scientist Sep 09 '13 at 12:58

1 Answers1

37

No.

According to a 2005 report from National Highway Traffic Safety Administration (NHTSA):

Table 1

The number of billions of miles driven and the number of fatalities on US roads increased slowly [data from Table 1 and Table 2]:

  • 2000: 41,945 fatalities over 2,747 billion miles.
  • 2001: 42,196 fatalities over 2,797 billion miles.
  • 2002: 43,005 fatalities over 2,856 billion miles.

Even if the entire increase in fatalities could be attributed to increased driving due to the fear/unavailability of flying, and the entire increase in 2001 was September or later, that is "only" an increase of 1,311 fatalities over a period of 15 months.

Compare that to the number of casualties in the 9/11 attacks which is generally placed just shy of 3,000.

Examining the graph, the trend for increasing fatalities extends further back than 2001. Even as fatality rates per mile travelled has improved, the distances travelled has increased.

To convince myself there wasn't a spike in the distance travelled data that i wasn't seeing in the column of numbers, I plotted it:

Graph of distance travelled

I didn't perform any advanced statistical analysis on this - it was sufficient to persuade me that there was no noticeable spike after 2001.

Community
  • 1
Oddthinking
  • 140,378
  • 46
  • 548
  • 638
  • Bruce may have been referring to the entire world rather than just the U.S. Even extrapolating those numbers out to the population of the entire world (which is probably not a valid thing to do) I get a figure per month that is still smaller than the death toll of 9/11. – Ladadadada Sep 03 '13 at 17:24
  • I also looked at the [US Air Carrier Statistics](http://apps.bts.gov/xml/air_traffic/src/index.xml) and saw that the distinct drop in passenger miles travelled was back to 1999 levels by March 2002 (and beating previous records by July 2003.) – Oddthinking Sep 03 '13 at 17:50
  • While I think this answer is accurate, it is *possible* that, for other reasons, 2001 would have otherwise seen a *reduction* of motorist deaths, if it weren't for the 9/11 incident... – Flimzy Sep 03 '13 at 18:34
  • 5
    Its probably impossible to measure how many would have not driven had 9/11 not happened, so there's no real baseline. The original claim is getting into the question of driver motive which is pretty murky. – Mark Rogers Sep 03 '13 at 18:54
  • I think it is reasonable to use 2000 (and the first 9 months of 2001) as proxies, combined with a general awareness that there were no other significant events (new technologies, new taxes, advertising campaigns, etc.) unrelated to the attacks that may have reduced travel (unveiling of the Segway in December notwithstanding). – Oddthinking Sep 04 '13 at 00:38
  • 3
    Another potential problem with this data: Bruce mentions that the increase happened in the *months* after 9/11, so maybe this data representation isn’t fine-grained enough. I personally don’t think so though. – Konrad Rudolph Sep 04 '13 at 07:17
  • 1
    @Konrad: If there was a spike in early 2002 that was subsumed by an equivalent dip in late 2002, your suggestion would make sense, but seems highly unlikely. – Oddthinking Sep 04 '13 at 09:15
  • 1
    Another look at this could be ignoring the death data and just look at the miles travelled. There is no claim that 9/11 made roads *less* safe, only that many more people used a car - which I find bogus: not many people would decide to travel a long distance in order to avoid a plane. If they are really scared, they don't travel at all. – Sklivvz Sep 04 '13 at 09:21
  • @Sancho: Yes, that is why I added the second graph, which tries to address that. Back of the envelope calculation: Using the US Air carrier Statistics, there was an immediate drop of very approximately 30 million passenger miles per month travelled (measured with a thumb held to a graph) recovering after 6 months to a drop of about 2 million passenger miles per month. Even if they all turned into driving miles, that would be an increase of less than 0.5 fatalities per month. It is noise compared to the 2.5 trillion miles travelled per annum. – Oddthinking Sep 04 '13 at 14:57
  • 6
    Bruce has now written a blog entry about this statement: https://www.schneier.com/blog/archives/2013/09/excess_automobi.html – Perseids Sep 09 '13 at 12:33
  • 4
    Great to see Schneier defending his position with references. I note that no less than Gigerenzer has drawn the opposite conclusion from me, which drastically reduces my faith in my answer. I haven't had a chance to go through the details though. I encourage someone to put up a counter-argument based on Schneier's references, ripping my argument to shreds. Seriously, (politely) show me I am wrong, and I will delete this answer both cheerfully and contritely (if that is a possible combination). – Oddthinking Sep 09 '13 at 13:22