Wednesday, October 19, 2005

Roadkill myths III: Unmasking a mythmaker

For the next installment in my series on road accidents in Israel (I, II), I had planned to discuss the claim that they are "a leading cause of death in Israel." That will have to wait, however, since a rare opportunity has presented itself: The chance to deflate one of Israel's leading mythmakers on the subject of traffic safety.

Dr. Elihu Richter of the Hebrew University is an "injury epidemiologist", who researches alleged harm from environmental pollution, cellphone use and other population-wide causes. He is, though, something of an alarmist, capable of finding injury risks regardless of the evidence - or lack thereof.

On the subject of road accidents, he is particularly fanatic, asserting (without evidence) that draconian speed enforcement could eliminate a high proportion of Israel's road deaths. I hope to address the question of speed and speed limits in a future installment in this series. For now, I want to show how Richter, by selective and inappropriate use of data, creates a misleading impression of Israel's road safety record.

Before I continue, I should note that despite my vehement disagreement with Richter on the subject of road safety, I am happy to give the man credit for defending Israel's March 2002 Defensive Shield anti-terrorism operations from one-sided assault by campaigners for "public health" and "human rights" in the European Journal of Public Health. As important as it is to improve our accident record, defending Israel from its enemies remains far more important. Kol hakavod.

Richter's latest missive

Back to our subject. Richter periodically spills his wrath on Israel's supposed poor road safety record in opinion pieces he writes for the Jerusalem Post. His latest missive appeared earlier this week.

Readers of my previous two installments know that in fact Israel's road safety record is not bad at all, and it has improved substantially over recent years, even compared with other Western countries at far higher levels of economic development. Let's see how Richter uses selective statistics to give the opposite impression.

Writes Richter: Israel, road death tolls have actually risen over the past 15 years

Richter can only be referring here to raw fatality numbers. Have they in fact risen over the past 15 years? Richter's basis for comparison is presumably 1990, which saw the lowest number of fatalities in recent years: 427 (according to Israel's Central Bureau of Statistics). Since the last full year of data is currently 2004, with 476 fatalities, there was indeed an increase of 11.5% over that time.

But what if Richter had chosen his endpoint slightly differently? The previous year - 1989 - saw 475 fatalities, virtually identical with 2004's total. The year before that, 1988, recorded 511, some 7.4% higher than 2004. Overall, yearly fatality totals are volatile. Indeed, since 1986 Israel has seen year-on-year increases in total fatalities as high as 18% and decreases as high as 14%. By comparing two arbitrary years you can prove nearly anything.

Regardless, as readers of this series already know, the raw total of fatalities is not the significant statistic here. Since 1990, total fatalities have not changed much, but Israel's population has increased by 45%. That is a significant safety improvement by any standard. More impressively, total distance driven in vehicle kilometers has more than doubled, indicating an even steeper decline in road safety per kilometer. Against this backdrop, to claim that "road death tolls have actually risen over the past 15 years" is misleading, to put it mildly.

Richter continues:
mainly from speed creep and urban sprawl.

This is nonsense. I argued earlier that total fatalities should be broken down into fatalities per kilometer (i.e., how safe the roads are) and total kilometers driven (i.e., exposure to the risks of driving). Fatalities per kilometer have been steadily falling, while kilometers driven have been steadily rising. Total fatalities can be reduced only by reducing either fatalities per kilometer or kilometers driven.

By blaming "speed creep" for the stubborn fatality rate, Richter implies that the roads have become less safe due to speeding drivers. If this were true, we would have seen a rise in the fatality rate per kilometer. The reverse is true; fatalities per kilometer continue to drop, nearly every year.

If speeds are increasing, then, it is only because the roads are becoming safe enough to support higher travel speeds. Good drivers drive more slowly on twisty, narrow roads than they do on straight, clear roads. When they speed up because the road conditions are better, they aren't being reckless. They're being sensible.

Speeds in Israel are rising because the roads are becoming safer, as the statistics clearly show. That's a good thing, not a problem.

By mentioning "urban sprawl", Richter refers to the second of the two statistics I mentioned, suggesting that the total distance driven is increasing faster than necessary. If Israelis drove much more than their counterparts in other countries, he might have a point. But we've already seen that the reverse is true.

Israelis drive much less than do Brits, Frenchmen or Australians. With increased economic development, God willing, Israelis will only continue to drive further each year; we are a long way from catching up with most other Western countries. If Richter thinks that increasingly-affluent Israelis can be persuaded to leave their cars at home and take the train en masse, he is woefully naive.

Richter again:
During these years, tough nationwide speed-camera enforcement reduced deaths by nearly 50% in Australia, and by 40% in the UK and France.

Again, by discussing speed enforcement Richter implicitly focuses on the per-kilometer fatality rate. Look again at the charts of per-kilometer fatalities by country: The fatality rate has fallen substantially in virtually every country reporting it, with Israel showing a greater improvement than most. To attribute that drop to a single policy (speed-camera enforcement) in any country is highly questionable, but all the more so when the improvement has taken place across the board, even in countries which have not adopted that policy!

Of course, Richter ignores the most salient difference between Israel on the one hand, and Australia, the UK or France on the other. From 1990 to 2003, Israel's population grew by some 44%. Over the same period of time, Australia grew by 15%, France by 6.2%, and the UK by just 4.3% (see here and here).

Is it any wonder that, while road deaths fell by 40-50% in those countries, they remained level in Israel? With its burgeoning population, Israel had to improve by over 40% just to stay in the same place! To cite these raw total figures while ignoring differences in population growth is to do a disservice to Israel's safety record. If Richter took his readers seriously, he wouldn't engage in such manipulations of the statistics.

Final thoughts

As he has done before, and as he did before the road even opened, Richter blames the new Trans-Israel Highway, Israel's most advanced roadway, for contributing to the road fatality rate. But here, Richter's own arguments work against him: Since the road was opened, total annual road fatalities in Israel have fallen every year, sometimes significantly! Hardly any fatal accidents have taken place on the new road itself. If anything, the Trans-Israel Highway should be credited with improving road safety in Israel, by providing a safe new motorway and channeling traffic away from unsafe old roads. But for Richter, no road is a safe road.

Unlike Richter, I hope Transport Minister Meir Shetreet agrees to raise the national speed limit next month, so that we can continue to drive at safe speeds on Israel's highways without being branded lawbreakers. Surely drivers would be better off keeping their eyes on the road, rather than focusing on their speedometers and scanning the shoulders for hidden cops.

Several times, Richter refers to the possibility of "zero road fatalities" in Israel. Since we are only human, accidents will continue to happen whatever we do. Reducing fatalities to zero would require either reducing the per-kilometer accident rate to 1/400th of its current value, or reducing the number of kilometers driven by 400 times, or some combination of the two (such as lowering each to one-twentieth of its current value). Such changes are inconceivable, and unprecedented in any developed country. If we really want to improve road safety in Israel, we can start by setting realistic goals.

To close on a positive note: The statistics bureau announced last week that, according to preliminary data as of the end of September 2005, road fatalities in 2005 are down 13% over the same period last year. If this trend continues, 2005 may see fewer road deaths, God willing, than even 1990. Odd that Richter didn't bother to point that out.


David Boxenhorn said...

Excellent post!

jlmkobi said...

good post
i always said (to my friends that listened) that richter was blowing too much steam.
so when ever he would raise the volume i would assume his numbers were suspect.
i remember when he started a campaign against lowering the driving the age to 17. intuitively (from my experience as a teen in the US) that teens that drive before going off to the army will have more experience for those weekend evenings out. especially where alchohol is only permitted at over 18.
intuitively (again no numbers from my side) i think two factor in road accidents - especially the severity - is
a. the involvement of trucks and large vehicles (so called professional drivers)
b. narrow lane width.
to shortly explain. -
a. we just know too many acquaintances that died when hit by a truck or bus (including pedestrians)
b. the narrow lanes leave less time to 'correct mistakes'
yasher koach

Zman Biur said...

Thanks for the comments!