January 14, 2004
SPEEDY STATISTICS
Speed limits in Australia are low. Urban limits have been reduced over the past few years to 50 km/h (30 mph) in many cities. Yet, writes Peter Martin in the Sydney Morning Herald, “this year's holiday road toll was the worst in eight years.”
This doesn’t stop Martin from concluding that higher speed limits cause more deaths. He doesn’t address the lower limits now enforced thoughout Australia, which should, according to him, have reduced traffic deaths to historical lows; instead he relies on old, possibly unmoored data:
During the energy crisis of 1974 the Carter administration succeeded in enforcing a low nationwide US speed limit of just 55 miles per hour (88kmh). Road deaths slid 15 per cent.
From 1987 each state again became free to choose to lift the limit on its rural interstate roads. Within a year most states had lifted their limit to 65mph. But seven left the limit unchanged.
In a paper soon to be published in the Journal of Political Economy, the economists Orley Ashenfelter from Princeton University and Michael Greenstone from the University of Chicago examine what happened in those states that lifted their limits. Their findings are surprising.
First, the actual increase in speed in those states was quite low, an average of only 2mph (3.2kmh) on the roads affected. The professors say that is because a lot of drivers on those roads were already speeding.
Second, the small increase appears to have pushed up deaths per mile on those roads by an astounding 36 per cent.
I’m not a former Treasury official like Martin, but when I see a claim that an average statewide increase in speed by only 2 mph blew out the number of deaths by 36%, I sense some deep selectivity and flawed analysis. These results aren’t “surprising”; they’re unbelievable. This is an issue for genuine experts to pull apart. With that in mind, over to you, readers.
UPDATE. In other automotive news, it turns out that cars can cause five-year jail terms even when they aren't moving.
UPDATE II. I missed a howler in Martin's piece, but readers didn't: Carter was elected in 1976, raising certain doubts about his administration's involvement in the 1974 oil crisis.
Posted by Tim Blair at January 14, 2004 12:50 AMReduce the speed limit to zero. As has often been pointed out, it's not speed that kills - it's relative speed.
Posted by: Theodopoulos Pherecydes at January 14, 2004 at 01:01 AM
First, the 1987 change in the federal speed limit isn't the meaningful one. That allowed a lot of rural highways to go to 65mph. In 1995, however, all states were free to set all of their own speed limits -- and now most suburban and many urban highways at 65+ and most rural highways are between 70 and 75mph.
Now, with a "massive" speed increase on the close urban highways, where most of the driving occurs anyway, highway deaths-per-mile have dropped quite a bit.
More data in a bit.
Oh, and that shmuck should realize Nixon -- conservative boogeyman himself -- set the 55mph limit, not Carter who was elected in 1976 and served from 1977 to 1981.
Hmmm... maybe it's plausible. Could be that the higher speed limit attracted more drivers to a faster route. More cars = more crashes. The post doesn't mention if traffic increased.
(And yes, I know the average speed only went up 2 mph, seemingly negating my "faster route" theory. But the legal speed limit increase may have attracted drivers who would not have used the route at scofflaw speeds.)
I'm also curious what cell-phone sales were like at the time. I'd say at least 4 out of 5 close shaves I've had the past five years were with cellfools.
Posted by: Dave S. at January 14, 2004 at 01:21 AMStephen Moore of The Cato Institute debunked claims about the 55mph spped limit here:
In 1995 the Republican Congress repealed the 55-mile-per-hour federal speed limit law. At the time, the highway safety lobby and consumer advocacy groups made apocalyptic predictions about 6,400 increased deaths and a million additional injuries if posted speed limits were raised. Ralph Nader even said that "history will never forgive Congress for this assault on the sanctity of human life."
But almost all measures of highway safety show improvement, not more deaths and injuries since 1995. Despite the fact that 33 states raised their speed limits immediately after the repeal of the mandatory federal speed limit, the National Highway Traffic Safety Administration reported last October that "the traffic death rate dropped to a record low level in 1997." Moreover, the average fatality rate even fell in the states that raised their speed limits.
Posted by: Dan at January 14, 2004 at 01:26 AM
Data mining can "prove" most any argument that you want to advance. The old saw "Statistics are used by rascals to influence fools." still holds true.
The 55 mph speed limit was probably the most widely broken law since Prohibition, especially in the West where driving across a state like Texas or Montana at 55 mph would be a career.
I think people are driving about the same speed now on the interstates that they were back then. Most people have a comfort level of around 70 mph and won't drive much faster than that. They also won't drive much slower.
Posted by: Randal Robinson at January 14, 2004 at 01:51 AMCHWO 740 in Toronto Canada (``The last radio station you'll ever need'') urges its drivers to speed up; the police have elderly traps out for them. CHWO elderly driver PSA (32kb)
Posted by: Ron Hardin at January 14, 2004 at 01:51 AMRandal - I don't know about the comfort level being 70. I drive from East of San Francisco to my job in Sacramento -- and if you aren't driving 80mph, you're a road hazard. :)
low nationwide US speed limit of just 55 miles per hour (88kmh). Road deaths slid 15 per cent.
Most of these deaths were from old age....as drivers tried to get places in states like Texas and Montana.
I'm and Aussie, so I don't know the minutae on US speed limit laws and when they were were changed and by who, but in 1994, Mum and Dad drove us across the US for six months in a Combi van. We did 60mph for a lot of the journey, which wasn't enough to forever consign us to the right lane and much overtaking by semi-trailers that made our little van shudder.
Posted by: Gareth at January 14, 2004 at 02:14 AMIt may be difficult to find meaningful data. One could hypothesize that the advent of mass cell phone usage in motor vehicles introduces another significant variable to the equation, likely reducing the true r-squared on the variable "speed". Last I checked, most jurisdictions were not collecting and/or collating meaningful data in their police reports regarding cell phone usage. Maybe somebody can fact-check me on this.
Posted by: Tongue Boy at January 14, 2004 at 02:22 AMRe: Randall's comment.
Comfort levels seem to vary from place to place and by time of day. I live in a rural area not far from a major metropolis and it seems like 70-75 is the norm on the interstate. However, the norm increases to 75-80 as I approach the major metro area during rush hour. Vice versa on the way out of town. However, this is only true in long-daylight times of the year. The rural norm going home even during rush hour seems to be 65-70 when it gets dark early. Also, the adage about Sunday drivers seems to be true; while the roads may be empty, the average speed in the metro area seems to drop sharply as bluehairs dominate the roadways.
I need a beer.
Posted by: Tongue Boy at January 14, 2004 at 02:31 AMOn my regular Texas-to-New Mexico route, any speed that doesn't involve relativistic time effects seems too slow.
Posted by: Bruce at January 14, 2004 at 02:45 AMOh, what crap. The death rate increased by 36%? From 8 to 13? How many actual accidents were there? Howabout in the 55 mph year there were five deaths of people driving alone in their cars, i.e., five relevant accidents. In the following year, after the limit had been raised, there was one accident, involving 8 people in a van, all of whom perished.
Buy this man a bicycle and turn him into a hood ornament.
Posted by: grayp at January 14, 2004 at 03:11 AMI'm a traffic engineer and have always sworn by the 85th percentile rule, i.e. post the limit at the speed which makes 85% of the traffic legal. This reduces speed differentials and collision exposure. Political constraints on govt controlled roads make this approach practically a non-starter (for any road engineer who wants to keep his job).
There's a whack of data out there to debunk the speed kills garbage. The drop in fatalities following the imposition of the 55 mph speed limit was linked to the economic recession (fewer vehicle miles travelled) There is nothing more misleading than claiming total collisions increased without also referencing the trend in total vehicle travel.
Good clearing house of sound bites here.
Posted by: Jay at January 14, 2004 at 03:26 AMEveryone's overlooking some salient facts here. First and most important is the steep increase in vehicular safety. Cars today have safety features unknown when Richard Exxon was president. We have airbags, seatbelt pretensioners, enhanced crumple zones and better interior designs that lessen fatalities. Second is suicide/homicide by vehicle. That doesn't need description, but cops tell me they believe the number of single-vehicle deaths are caused deliberately. Third is the big improvements in highway safety dengineering. Here in the States we've seen changes in guard rails medians, bridge supports, lighting and so forth that have heled save lives.And lastly but most important is that investigators haven't yet figured out how to apportion driver stupidity in accident reports. If they could, we'd probably get a better understanding of what an "accident" is and could lower the death rate by creating moron lanes.
Posted by: Gary at January 14, 2004 at 03:59 AMNice site, congratulations!
Viagra Online
Pharmacy
Order Viagra online, from a real pharmacy.
Tim/Andrea, try the MT-Blacklist plugin to kill comment spam.
Posted by: Robert at January 14, 2004 at 05:29 AMWhat Gary said, plus with evil cell phones to call for ambulances and helicopter flights to hospitals and improved trauma care, fewer people die now than would have in '7x, anyway, even with identical injuries.
Even if Speed Killed in '74, it's a lot less likely to now.
(And doesn't Martin know that Carter was elected in 1976, and thus incapable of signing laws in 1974? If I had Carter's magical pre-emptive bill-passing powers, I could pave the way now for my own election in 2008!)
Posted by: Sigivald at January 14, 2004 at 06:54 AMfunny, no one has mentioned that there are far fewer (at least here in the states) folks driving completely shitfaced than there used to be; i think that's got quite a bit to do with it as well.
Posted by: Mr. Bingley at January 14, 2004 at 07:00 AMi seem to remember that the study does to show speed increases led to more fatalities included a horrific crash in Texas that killed quite a few people. it skewed the data quite a bit because there are actually very few road fatalities per 1,000 miles traveled. the reasoning was, Texas raised the limit, Texas lost a whole bunch of people this year, speed kills. the problem was that the crash occurred in a 55 mile per hour zone. i think it was actually fog related, and occurred even slower than that.
the point is that traffic fatalities are relatively rare, in terms of number of miles driven. the reason so many people get killed is that there are so many miles driven each year. even a very low percentage translates into a pretty big death toll when applied to an enourmous number.
In explaining fatalities from speed obscures the major reason: dangerous driving and, occasionally driver error.
On speed,most drivers drive according to raod conditions.In peak hour, as a blunt example, drivers proceed below posted speed: 40-60kph on roads posted at 70.
Many fatalities occur on country roads, due to linked reasons:a large number of country roads are in need of an upgrade but in the meantime, safe driving on those roads requires familiarisation and even then some drivers drive dangerously on dangerous sections.And additional hazards can affect safety, like dense fog which can blanket even very good highways reducing visibility to the end of the car bonnet.
Speed is a straw man used precisley to buttress what is nothing but an excerise in money grabbing by govts in the form of speeding tickets.
Posted by: d at January 14, 2004 at 08:13 AMJust had a 19 year old kill four people, two in his car and two in the car he hit head on. The speed limit is 65 mph. He was going 94 mph. If the speed limit had been 55 mph he would still have been going 94 mph.
Posted by: Fred Boness at January 14, 2004 at 08:29 AMI’m an ex-Treasury official (NSW) too and that's the biggest load of shit I've ever heard. Most road accident deaths in Australia are from cars running off the road because the driver falls asleep. Many are pedestrians who step in front of cars (no fault of the driver). In both cases speed has little to do with it.
I wonder which Treasury Peter Martin worked for. If, as I suspect, it was the Commonwealth, then who cares what he thinks? They don't regulate traffic and speed laws, the states do.
Over Christmas I got booked in Melbourne for doing 64 kph in a 60 kph zone; A$120 fine. I’m sure the person I would have inevitably killed by my reckless speeding is very pleased.
Posted by: Matt Crowe at January 14, 2004 at 08:29 AMYou'd better ask the spiv( the Bracks Govt) responsible for the fine, who the saved `victim' was, the govt after all resembles a pack of pyschics,ie.- stupidity plus mystic explanations for reality, so they should be able to say who the womabt is.
Posted by: d at January 14, 2004 at 09:10 AM(This comment refers to the state of Queensland, Australia, where the open road speed limit is generally 100 kph, in some areas 110 kph. Roadside billboards have sprung up everywhere prominently displaying the message "Every K over is a killer", and much prominence is given in newspapers to reports of the number of motorists "caught speeding" over a holiday weekend or other time span.)
Let's start by dropping the blanket use of phrases such as "caught speeding", and "every K over is a Killer". In some circumstances they're meaningful, in others grossly misleading and irrelevant. In urban areas, particularly near schools, hospitals, aged care facilities etc where stopping distances are critical, go for it, step up visible presence and enforcement and save lives and reduce injuries. Key words: "visible presence". Hidden cameras might be great earners, but useless at prevention.
On rural roads a problem is fast developing, with at least two causes. One is entirely inappropriate , too low, yes LOW, speed limits on many kilometres of road. A comprehensive review of open road speed limits is urgently needed.
The second is potentially more dangerous and could be summed up as "every K under is a killer". The stringently enforced requirement that in overtaking another vehicle the posted speed limit must not be exceeded (we're talking open road here) is plain stupid and downright dangerous. Ask the many experienced drivers (company reps, couriers, truckies,those who drive for pleasure - yes they do exist) who do upwards of 50 thousand kilometres annually, their observations regarding bottlenecks being created by this requirement. Bottlenecks equals frustration equals stupid risk taking equals tragedy. And "speeding" is probably blamed. Inappropriately low speed would be nearer the truth.
It's a very complex, many faceted problem, and certainly won't be solved with higher fines, more points, double penalties on holiday weekends, and still more cameras. Or with sanctimonious bleatings like "if you don't want to be penalised, simply don't speed". This is a real problem faced by real people in the real world out there, particularly in a big country like Australia where distances to be covered are huge. Drivers want smoother and more hassle free flow of traffic on our roads which equals less frustration and loss of concentration which equals happier and safer drivers resulting in fewer incidents, injuries and fatalities.
Let's have frank and open discussion, which includes input from many drivers at the coalface, traffic police, ambulance drivers, all those who've "been there, done that". It just might solve the problem. Then you can proceed to better roads, driver training etc.
Well, no one thinks it will be easy. But it must be better than the present approach.
Posted by: Galen at January 14, 2004 at 09:33 AMJay beat me too it, but if there was a drop in deaths with the reduction to 55mph then it was probably due to the fact that so many cars were lining up for gas rather than actually driving anywhere.
No-one's commented on the three ecoterrorists from the Earth Liberation Movement who trashed cars. Who will join me in my Automobile Liberation Movement? Liberate all drivers from speed limits, radar traps, revenue-raising fines, NOW! Death to all traffic cops; stupid pedestrians; tree-hugging, vegan-loving cyclists; and wandering wildlife! The road is NOT there to share!
Posted by: freddyboy at January 14, 2004 at 09:54 AM>I’m an ex-Treasury official (NSW) too and that's >the biggest load of shit I've ever heard. Most >road accident deaths in Australia are from cars >running off the road because the driver falls >asleep. Many are pedestrians who step in front >of cars (no fault of the driver). In both cases >speed has little to do with it.
Nonsense. In NSW 36% of fatal accidents involve excessive speed. Only 15% involve fatigue - people falling asleep at the wheel would be a smaller subset of this figure. Pedestrians account for 15% of the road toll, again, people "stepping in front of cars" would be a tiny proportion of this figure.
You can educate youself here:
http://www.rta.nsw.gov.au/roadsafety/downloads/dynamic/monthly-accident-data.pdf
Posted by: Spud at January 14, 2004 at 10:01 AMSpud, pedestrians walking out in front of cars is a very ferequent irritation, even when driving slowly as in the case of one who launched themselves from the footpath a couple of feet in front of my car.What is not amusing is the sheer number of them who,in doing so, look the other way as if the driver is reponsible.Stupidity, its name is pedestrian.
Posted by: d at January 14, 2004 at 10:47 AMRobert: I have MT-Blacklist installed, but turned off as it slowed posting uploads to about the speed of the average glacier. However, that was some time ago; they have done some upgrades since. If Tim wants, I'll upgrade & reactivate it.
Posted by: Andrea Harris at January 14, 2004 at 11:03 AM>>"I’m not a former Treasury official like Martin, but when I see a claim that an average statewide increase in speed by only 2 mph blew out the number of deaths by 36%, I sense some deep selectivity and flawed analysis.
Yeah, Tim, I'm sure I should trust your 'gut feeling' over a peer reviewed paper in the JPE, produced by the University of Chicago Press. I'll decide when I see the paper.
Posted by: Jason Soon at January 14, 2004 at 11:11 AMMy mama always told me to ignore peer pressure.
PS: MT-Blacklist updated; comments and posting seems to work as normal.
Posted by: Andrea Harris at January 14, 2004 at 11:27 AMSpud thinks that one can educate oneself at the RTA website. The RTA provides some useful information, but one needs to be aware of some of the questionable aspects of the data they provide - including the way in which crashes are attributed to different causes. The attribution of deaths to speed being particularly fraught. Spud might also like to consider the possibility that there may be an inverse relationship between speed and crashes caused by fatigue. Alas, this possibility is also largely ignored in road safety research.
Posted by: Tom at January 14, 2004 at 11:33 AMMost driving deaths are due to people eating LUNCH while driving, often drinking EXPENSIVE WINE at the same time. How many people have DIED because of your drive-thru LUNCHES of CHICKEN MCNUGGETS, Tim?
Posted by: Jerry at January 14, 2004 at 12:23 PMSpeed limits in Australia are low. Urban limits have been reduced over the past few years to 50 km/h (30 mph) in many cities. Yet, writes Peter Martin in the Sydney Morning Herald, “this year's holiday road toll was the worst in eight years.”
Tim you are being very selective in your use of statistics!
1) You look at just one year's figures to support your shakey conclusion that lower speed limits are useless. Anyone with a basic grasp of statistics would warn you to look at the figures over a few years, not just from one year: thus from 95/96 onwards to 2002/03 the fatality figures for the xmas holiday period have been: 71, 86, 82, 73, 75, 75, 58, 67. In other words, there has been a slight downward trend during the period when lower urban limits have been in operation. You remind me of greenie doomsayers looking at a 4 day heatwave and saying "See we told you global warming is real!"
You also make no allowance for the increased population of Australia in the last 7 years nor the increased number of vehicles on the road!
It's also worth looking at the road toll stats for the whole year rather than just the xmas holiday period:
from 1995: 2017, 1970, 1767, 1755, 1764, 1817, 1737, 1715
A decreasing trend. But per vehicle-kilometer travelled the trend is even more markedly downwards.
see: http://www.atsb.gov.au/road/stats/pdf/crash_rates_2003.pdf
further
2) if you look at http://www.atsb.gov.au/road/pdf/xmas_fatal.pdf you'll see that more fatal crashes at any time of year appear in areas with speed limits of 100kmh or greater. During the xmas holiday period almost 60% of fatalities occur in such zones. It's not logical to blame the low limit in urban zones for the high number of fatalities we recently experienced.
If you really wanted to see what effect low urban limits on fatalities you would check the stats for fatalities in those urban areas with the low speed limits over a number of years. (Not all urban areas have changed to lower speed limits as you know.)
And by the way the report also points out that, on average there are less fatalities per day over the xmas holiday period than throughout the rest of the year. This may be due to the tougher penalties in force during the holiday season, and the public campaigns that are run then.
Cheers
Tom
Speed limits on the highways in the Northern Territory are unlimited, teenagers are able to get thier L's at 16, and once on thier P's, only stay on them for a year.
And what state has one of the lowest road tolls?
Posted by: Vikki at January 14, 2004 at 01:09 PMFYI, you can view the Ashenfelter/Greenstone paper here.
Posted by: Alex Robson at January 14, 2004 at 01:12 PMMuch of the debate here seems to be focussing on the issue of exactly how speed limits alter driver behavior and fatalities. For the purposes of this discussion, the critical part of the Ashenfelter/Greenstone paper is on page 21 of the working paper version:
"Thus far the analysis has attempted to obtain the causal effects of the adoption of the 65 mph speed limit on fatalities and travel times. But these estimates are unlikely to be the structural relationships between speed limits and these two outcomes. We suspect that the adoption of the higher speed limit may be accompanied by other changes that are unobservable to us but are intended to affect these outcomes. For example, states may accompany the introduction of higher speed limits with alterations in state trooper behavior that limit the increase in speeds and/or fatalities. Further, the higher speed limit may induce changes in the variance of speed that could have an independent effect on fatality rates (Lave 1985). The available data does not permit an investigation of whether such relationships underlie the above estimates, but such an investigation is unnecessary for our purposes. The important issue is that the adoption of the higher speed limit provided the median voter/driver a trade-off between increased fatalities and reduced travel times, whatever the precise mechanism."
In other words, they explicitly point out that they aren't concerned with exactly why fatalties changed when speed limits changed. They are trying to measure the value of a statistical life, and they do not answer (and, for the purposes of their paper, they don't need to answer) the questions of why and how speed limits change behaviour. For framing policy, I would have thought that this was the most important issue.
So on the basis of the Ashenfelter/Greenstone paper, I would find it difficult to accept the claim in the headline of Martin's article that "comparative data from the United States shows a speed limit increase does push up the road death toll". Perhaps that claim is true, perhaps it's not. But the Ashenfelter/Greenstone paper isn't much help in verifying this claim.
"The RTA provides some useful information, but one needs to be aware of some of the questionable aspects of the data they provide - including the way in which crashes are attributed to different causes."
-------------
Um, Tom, I'd be backing the data compiled by police accident investigators, coronial inquiries and the RTA (which is what those tables represent) over anecdotal best guesses posted on a discussion board any day of the week.
You originally asserted most road deaths in Australia were caused by people falling asleep at the wheel or pedestrians blindly walking in front of cars. The best available data shows that isn't even remotely true. Please feel free to cite some other source for either of these contentions, aside from an ill-defined gut feeling on your part.
Posted by: Spud at January 14, 2004 at 03:28 PM
>Yeah, Tim, I'm sure I should trust your 'gut
>feeling' over a peer reviewed paper in the JPE,
>produced by the University of Chicago Press. I'll
>decide when I see the paper.
Yeah, Jason, peer reviewed studies on questions of public policy published by academic houses never contain selective data or researcher bias.
Forgot something-
>I'll decide when I see the paper.
Decide what? It was "peer reviewed" and "produced by the University of Chicago Press." Why is Tim's reading and reaction worthy of dismissal on those grounds, but yours is not?
Tim (academic issue - word limit waiver),
Can't comment on the Princeton paper, unless one gets the base line figures, one can't compute the various ratios of variation.
But here are a few caveats before petrol-heads dismiss Martin's point.
First, the actual increase in speed in those states was quite low, an average of only 2mph (3.2kmh) on the roads affected. The professors say that is because a lot of drivers on those roads were already speeding.
This stat refers to average speeds. However, accidents occur, almost by definition, at the extremes of a distribution. If the speed limit on interstates was increased by 10 mph, from 55 mph to 65 mph, then this is an 18% increase in speed limit - fairly sizeable. In some cases, the speed limit on interstates went up to 75 mph, which is a 36% increase in speed limit. That's a big move, and it is likely that actual speeds went up a lot more in those states.
Also, when speed limits go up, it's possible that a larger fraction of extreme drivers will try to push the envelope beyond their car's, and their own, effective competency limits.
A recent study in the USverifies this phenomenon:
In Maryland, where the interstate speed limit is 65 mph, the mean speed was 66 mph. About 1 percent of drivers exceeded 80 mph. By contrast, in Colorado, where the interstate speed limit is 75 mph, the mean speed was 76 mph. About a quarter of drivers regularly went over 80 mph...."These are the fastest speeds we've ever observed," Richard Retting, the institute's senior transportation engineer, said in the report.
Martin's article refers to the powerful, non-linear, effect a small increase in speeds has on the rate of road fatatlities:
Second, the small increase appears to have pushed up deaths per mile on those roads by an astounding 36 per cent.
This is not such a couter-intuitive result, if you bear in mind the fact of non-linear effects, ie curves that radically change gradient after a threshold point is reached. Imagine if gun licences were originally restricted to persons without violent record, and then were allowed to persons with such a record. This might result in only 10% more people getting gun licences, but gun-related homicides might go up by, say 50%.
The US study just quoted confirms that increases in the speed limit can trigger increases in the road fatality rate, although these increases appear to be proportionate.
The study focuses on 22 states that raised their interstate limits to 70 or 75 mph almost immediately after the repeal of the federal cap and tries to isolate the effects. Those states are compared with trends in 12 states that kept their limits at 65.
The study found 1,880 more deaths on the interstates in those 22 states from 1996 to 1999
Martin's claim of a much larger increase in deaths per mile travelled could only come about if the total number of miles travelled was significantly reduced during this period - which is unlikely.
Still almost 500 people per year killed is ~ 1% increase in the road death rate. Even if Martin's figures are dodgy, that is still a lot of lives saved, well worth the inconvenience of slower, more boring, journeys. Posted by: Jack Strocchi at January 14, 2004 at 06:40 PM
The overall rate of US road fatalities is declining, although not as fast as other OECD states.
Once World Leader in Traffic Safety, U.S. Drops toNo.9
As of last year, the death rate in the United States had fallen to 1.51 deaths per 100 million miles traveled from 1.58 in 1998.
Since 1970, the United States traffic death rate has fallen from nearly 4.8 deaths per 100 million miles traveled. By 2000, the rate in Britain had fallen to 1.2 deaths per 100 million miles from 6.1 in 1970. The new figure is the lowest traffic death rate compiled by the Organization of Economic Cooperation and Development, which collects a variety of statistics from industrialized countries.
Australia's death rate has fallen from 7.13 in 1971 — the country did not estimate distances traveled the previous year — to 1.45 in 2001.
The main reason for the US becoming an outlier appears to be seat belt usage, drunkedness and heavy-vehicle ownerhsip:
Most traffic safety experts agree that the seat belt remains the world's most effective safety device. The nation's usage rate has risen considerably over the past couple of decades, to nearly 80 percent today. But top safety regulators in Canada and Australia say their use of seat belts is about 10 percentage points higher.Posted by: Jack Strocchi at January 14, 2004 at 06:50 PM
It might be more useful if we looked at figures for "killed or severely injured".
Are more peopel surviving crashes which once would have killed them?
Still almost 500 people per year killed is ~ 1% increase in the road death rate.
IN short - statistical noise.
Even if Martin's figures are dodgy, that is still a lot of lives saved, well worth the inconvenience of slower, more boring, journeys.
In other words, screw the fact that the data is bad, and that it will impose societal costs, I support lower speed limits.
Posted by: R C Dean at January 14, 2004 at 10:10 PMHey Mr "screw the data" RC Dean,
If you followed my comment's links you would see that the ~1% pa death rate figure is cumulative 1996-99
- not random, but all in the same direction: upwards.
This is not "noise" it, is a trend.
Capice?
Jack Strocchi
The majority of people that drive 10-20 k over the speed limit don't have accidents. So "speed kills" is a simplistic fall back for those that feel the need to save others from themselves.
Posted by: Gary at January 14, 2004 at 11:57 PMGeez you people need to take your country back. The country that builds the Monaro and Falcon with a 60 mph speed limit?
When I meet someone here who expresses a desire to vote democrat, I simply remind them who forced that 55 mph on us. And yes, I know who signed the bill given him by the Imperial post watergate congress.
I guess we should never underestimate the willingness of some people to regulate the behavior of other people for their own good, data and/or facts notwithststanding.
At least in the US, to get good data you would certainly have to hold a great many things constant: accidents per mile driven; deaths per mile driven; average mass of vehicle or something to measure change in the M x V equation; age and experience of drivers; alcohol consumption; enforcement techniques and application; seatbelt usage; and so on and so on.
Jack Strocchi's post on the decline in fatalities actually aggregates (not really so usefully for his purposes) the above inputs and undermines any argument he could make about increased speed limits increasing fatalities. The doggone rate is going down. To show the the rate of decline is being slowed by higher speed limits would require more data than is used. (And better statistical techniques, such as structural equation modeling.)
Posted by: JorgXMcKie at January 15, 2004 at 02:16 AMJack,
The Insurance Industry has always bitched about higher speed limits. They continue to spit out reports showing absolute increases in collisions or fatality rates in states that raise speed limits but ignore the effects of increased travel in those states. I notice in the article you cited (Atlanta-Journal Constitution) that the 55 mph myth is still being repeated, despite the drop in total fatalities being linked more to the drop in travel than any reduction in speeds.
If total number of murders in a city increases by 50% and the population by 200% over the same time period after a policy change (say gun control), how absurd would it be to harp on the absolute increase in murders when the actual risk of being a murder victim has fallen considerably?
Get real!
In the U.S. it's the insurance industry that's gung-ho about lower speed limits. They dress it up in terms of safety, but that's not their reason. They support lower speed limits because it means more tickets being issued, which means they can raise rates on more drivers.
Re: Maryland and Colorado speed limits and number of drivers speeding. There are two factors that affect the distribution of driver speed: road conditions (by this I mean design, weather, and traffic volume) and the likelihood of getting a ticket if one goes at a given speed.
Maryland's freeways are vastly different from Colorado's in terms of traffic volume and even design. Although Colorado is a mountainous state, of the three major freeways in the state only half of one (interstate 70) goes through the mountains. The others are in the plains.
It's not surprising that people drive faster on flat limited access highways that have limited traffic and where there's huge distances between cities. When I'm in the Colorado plains I usually set my cruise on 82 mph.
I don't buy the idea at all that if a speed limit is raised that the top end speed goes up all that much. If someone's used to doing 90 in a 65 zone and the limit gets raised to 75, he's not going to go 100--he'd already be going 100 otherwise.
Frankly, from my experience, it's speed differential that causes the most problems. That's why the 90 mph guy is a hazard, because he's going a lot faster than others. That's also why the 55 mph limit was such a problem, because some people will go 55 while driving through Montana and others were going 85.
Finally, the 55 mph limit, which was in place for far longer than necessary, caused a general disrespect for speed limits that percolated down to city streets.
Posted by: Jeff at January 15, 2004 at 03:43 AMThis takes the cake: the British govt's.dictat that bluebottles inspecting grave accidents find links to speeding though speeding be not a factor. In other words, cook the evidence and the reasons for the accident so long as speeding is specified as the cause.A policy of officially sanctioned lying, it might be said, which has other not neglible ramifications such as : false explanations leads to false measures in order to counter to reduce further the probability of grave accidents.
And that also sums up the position of the spivs in Spring St.
Posted by: d at January 15, 2004 at 09:40 AMJorgXMcKie
I am not an expert at statistics, just a one year course at uni. But surely one does not need "structural equation modelling" to tease out the death-defying effects of maintaining lower speed limits. That was exactly what the first report showed, by control against the states that raised speed limits.
Posted by: Jack Strocchi at January 15, 2004 at 02:24 PMSpud said that he'd "be backing the data compiled by police accident investigators, coronial inquiries and the RTA (which is what those tables represent) over anecdotal best guesses posted on a discussion board any day of the week."
I used to work in road safety and know that the police attribution of many accidents is often quick, dirty, and entails an in-built bias to nominate speed in the absence of any other easily identifiable factor.
As for the other comments that Spud ascribes to me about pedestrians and falling asleep, I did not make them - there appear to be more than one "Tom" contributing to this topic.