This work is licensed under a Creative Commons License.
But, ummm... they keep hinting about this important lesson but what ARE we supposed to learn from this story exactly? What has Melony Springfield learned?
Dangerously inexperienced in night driving and dazzled by the headlights, she hadn't seen a bend in the road and her hotted-up Mazda 121 hatchback ploughed into a tree.
So what she did wrong was driving a "hotted up" car. How very evil of her. By the way, how exactly do you go about hotting up a Mazda 121? Did she have some really racy stickers on the side? Hooo, that is dangerous. Possibly her tires were a few millimeters different to the manufacturer's specifications. Yeah, that'll make you crash every time.
She says that she used to drive over the speed limit regularly but on the morning that she had an accident she was driving at the speed limit of 80kph. Is this the lesson? When you drive over the speed limit you don't have an accident but when you drive at the speed limit you do... interesting. Are we supposed to believe that the accident (which happened at the speed limit) was a punishment for all those other times she was naughty?
These days, Ms Springfield is a compensation worker in the city. She is telling her story to help educate P-platers. "I want people to understand how dangerous speeding is," she said.
Well she would know how dangerous speeding is, seeing how she crashed while NOT speeding. This is a bit like going to an Alcoholics Anonymous meeting and recruiting someone to tell the world about the great dangers of smoking.
Here's a novel idea, let's put our collective brains into gear for a short session. Driving any sort of car (hot, cold or lukewarm) into a tree is bad for your health and for your car. So forget about this whole speeding issue and forget about what type of car she was driving and who was driving it and lets get to the basic issues here -- cars, night visibility and trees.
I seem to remember there used to be a lot of those steel safety barriers (with the M shape profile) on corners, especially bends on country roads. You don't see as many these days because they have been replaced by the cheesegrater steel cables, or in this case by nothing at all. Saves a bit of money, doesn't save too many lives. Simple government equation, nothing else to it.
Now a divided road with a bit of grass and shrubbery in the middle is a lot more comfortable to drive on, the headlights aren't right in your face and with a good choice of greenery in the middle (thick grasses, small shrubs and similar) a car that ends up in the middle section will slow down before it hits anything major and the driver wakes up pretty quick if they drift into that stuff. Then there is the "rumble strip" that they put here and there but not very many places. You can't tell me in a blue fit that rumble strip cost much to install, why isn't it everywhere? How about a decent width of shoulder on the road? How about some curb and guttering? Plenty of NSW state roads have absolutely nothing at the edge of the road, the tar just stops at a ragged edge. It's all over Western Sydney and it's pathetic.
The NSW state govt finally got its arse kicked into fixing Windsor Road and parts of it are really nice to drive on. A lot of roads further West are still terrible. I notice the the Telegraph carefully fail to mention which road or which bend was involved just to discourage anyone from thinking that improving the roads might be worth a try.
Talking of terrible things that need improving, how about public transport at 5am? How about public transport in Sydney at any time? If we had a working (and affordable) public transport system then hey presto, not as many people would need to be on the roads while they are half asleep going to and from work.
I was not in the wrong with my accident however a full licensed driver had their high beam on coming the opposite way to me and not being able to see the road I had hit a tree at 80km/h. I feel that if they are going to now enforce tougher laws on P Plate drivers that they should learn to enforce the laws already in place for full licensed drivers. P Platers are not the only drivers that break the law and I wonder why the government believes that making new laws are going to reduce the amount of people young and old dying on our roads. Is it because they need to be able to sleep better at night or is it because they are actually concerned? If they were so concerned they should get more patrol cars on our roads to enforce the road rules already there in order to protect everybody.
Taking the trouble to look through the published RTA statistics, I found a special report here about accidents involving young drivers. First thing I notice is that there is no author. Not even anyone willing to admit that they reviewed or authorized the report, not a soul, blank, anon, gutless. Next thing I notice is that there are no references, nothing to trace back where the information is coming from, nor any method by which a reader might verify the data. Now I'll quote some examples of completely fallacious statistical reasoning -- coming (presumably) from someone who does statistics for a living (but who won't put their name to their work).
Crashes and passengersYoung drivers have a greater risk of involvement in a fatal crash if they have two or more passengers. Of the 89 young drivers involved in fatal crashes from 2001 to 2003, 36 per cent were driving with two or more passengers.This contrasts with 15 per cent for drivers aged 26 and over.
Time of crash
Numerous international and national studies, along with an analysis of NSW crash data, have confirmed the increased risk for young drivers at night. The involvement of young drivers in late night and early morning fatal crashes is significantly higher than that of older drivers as outlined in figure 3. Of the 96 P plate drivers aged under 26 who were involved in fatal crashes from 2001 to 2003, 33% crashed between 10pm and 5am.This compared with 14% of drivers aged 26 years and over.
Pretty obvious flaw in the logic, and sadly a common mistake. You guessed it, Bayesian probability strikes again and (just like always) these guys screw it up. Spend any time on a Friday or Saturday night and what you are guaranteed to see is a lot of younger drivers, usually with a lot of passengers. That's because young people do go out at night and they have the energy and the inclination to stay out late. Let them work a soul destroying daily grind for ten years and they will decide it's easier just to stay home.
More young people tend to work the late and early shifts doing the jobs that no one else wants to do (like changing tapes on computer backups at 02:00 or typing in the names on cheques presented to the bank at 01:00). Older, more educated workers have the ability to negotiate for the better jobs and work more enjoyable hours.
The concept of Bayesian inference is very simple, but still seems difficult for many people to grasp. In basic terms, to understand the statistical linkage between a potential cause and an observable effect, you require knowledge of the samples where the effect is observed and you also require knowledge of the background population from which you are selecting the samples. Thus, just because there are plenty of young drivers out on the road on a Friday night, we would logically expect to see more accidents amongst this age group at this time, even when young drivers are no more dangerous than older drivers.
Exactly the same situation holds for passengers -- young drivers are more likely to carry more passengers because they have friends with no license, friends with a license but no car, and they just have more friends (plus more available time to be out and about with those friends). Thus, the background population is uneven, thus it is completely logical that the sample population would also be uneven. Follow the links above, or buy a statistics book, it will all be explained better than I can, with worked examples.
Here's another statement from the same report:
A 17-year-old driver with a P1 licence is about four times more likely to be involved in a fatal crash than a driver aged 26 or older.
This is a direct probability statement, but there are no other statistics in the report to support it, no reference and no explanation of the methodology used to get this result. Detailed statistical reports are available for 2004, 2005, 2006 and other years. These reports do not provide sufficient information to be able to confirm the above statement. They do not give a breakdown of both licence type and age and they do not isolate the 17 year-old drivers from other groups (the age category is always 17-20 years). We can take a rough stab and presume that everyone in the 17-20 category is a P1 license holder (a bit rough, but should give something in the right ballpark). We can get figures from Table 16a (basing this on the age of the driver rather than the victims). We have the young drivers, the 26+ drivers (those old enough to almost 100% be off their P plates) and the 60+ drivers (the oldies, for comparison purposes).
| Fraction of Fatal Accidents with XX Year Old Driver Involved | |||
|---|---|---|---|
| 17-20 Year Old | 26+ Year Old | 60+ Year Old | |
| 2004 | 88 / 700 = 12.6% | 518 / 700 = 74.0% | 103 / 700 = 14.7% |
| 2005 | 62 / 665 = 9.3% | 504 / 665 = 75.8% | 101 / 665 = 15.2% |
| 2006 | 66 / 646 = 10.2% | 475 / 646 = 73.5% | 101 / 646 = 15.6% |
| TOTAL | 216 / 2011 = 10.7% | 1497 / 2011 = 74.4% | 305 / 2011 = 15.1% |
That's not enough to determine the probability -- still need that all important background population level. We need to know what proportion of all driving is done by 17-20 year old drivers. Let's presume that all license holders do equal amounts of driving (highly unlikely but we don't have anything better) and use Table 33.
| Fraction of XX Year Old License Holders | |||
|---|---|---|---|
| 17-20 Year Old | 26+ Year Old | 60+ Year Old | |
| 2004 | 268614 / 4345070 = 6.2% | 3675617 / 4345070 = 84.6% | 818922 / 4345070 = 18.8% |
| 2005 | 274625 / 4396993 = 6.2% | 3711947 / 4396993 = 84.4% | 836409 / 4396993 = 19.0% |
| 2006 | 280120 / 4474183 = 6.3% | 3773815 / 4474183 = 84.3% | 867841 / 4474183 = 19.4% |
| TOTAL | 823359 / 13216246 = 6.2% | 11161379 / 13216246 = 84.5% | 2523172 / 13216246 = 19.1% |
Some simple ratio calculation gives as a comparative "dangerousness" factor which is the relative change in probability as compared to the average probability of a fatal accident (we don't actually know what the average probability is because we don't know the total amount of driving, but the probability of a fatal accident is fairly small).
| Driver Fatal-Accident "Dangerousness" Factor (ratio to average, presuming all license holders do equal driving) |
|||
|---|---|---|---|
| 17-20 Year Old | 26+ Year Old | 60+ Year Old | |
| 2004 | 203% | 87% | 78% |
| 2005 | 150% | 90% | 80% |
| 2006 | 162% | 87% | 80% |
| TOTAL | 173% | 88% | 79% |
Putting pairs of "dangerousness" factors together, we can estimate that the 17-20 year old drivers are approximately twice as likely to have a fatal accident as the 26+ year old drivers. This is a long way different to the "four times" figure cited above.
Where are the errors in this estimate? The first error is that no public data is provided for individual age years (e.g. 17, 18, 19, 20) so we only have the lumped data to work with. The big error is trying to estimate how much driving is done by any particular age group of license holders. Certainly, older people tend to drive less and there are many more old people who hold a license that they rarely or never use. Thus the "dangerousness" factor for all the old people above can be pushed up a notch to counter for that.
I would guess that most of the young people who aren't interested in driving, also aren't going to be bothered getting a license. Those who do go to the trouble of getting a license will also be getting out on the road much more actively, so all of the "dangerousness" factors for the young people above are overestimated (but without a census of road users, it isn't possible to get better results).
All told, when you put it together, the young people might be a bit more dangerous than other drivers but the whole "P-place menace" bullshit is blown way out of proportion.
Just for completeness, heres the same results generated using Table 16b (which is injuries) and gives similar results (except that old people seem to be more fragile or when they do crash, it is more likely to be fatal, presumably the opposite happens with your people, they are less fragile and more likely to survive). Also, the figures for injuries are larger so they average out better over a year.
| Driver Injurious-Accident "Dangerousness" Factor (ratio to average, presuming all license holders do equal driving) |
|||
|---|---|---|---|
| 17-20 Year Old | 26+ Year Old | 60+ Year Old | |
| 2004 | 195% | 80% | 51% |
| 2005 | 194% | 80% | 53% |
| 2006 | 192% | 81% | 55% |
| TOTAL | 194% | 80% | 53% |
It's an observable fact that the overwhelming majority of students hate statistics and overwhelming majority of journalists love to quote statistics (almost always wrongly). There's probably a connection between these. No, I don't have an actual study for this, but readers of my news commentary will see plenty of examples of statistical abuse by journalists.
While I have the RTA stats files open, I'll have a little peek at some of the major factors of accidents. First an explanation of how they come up with their speeding and fatigue measurements.
Criteria for determining speeding and fatigue involvementSpeeding
The identification of speeding (excessive speed for the prevailing conditions) as a contributing factor in road crashes cannot always be determined directly from police reports of those crashes. Certain circumstances, however, suggest the involvement of speeding. The Roads and Traffic Authority has therefore drawn up criteria for determining whether or not a crash is to be considered as having involved speeding as a contributing factor. Speeding is considered to have been a contributing factor to a road crash if that crash involved at least one speeding motor vehicle. A motor vehicle is assessed as having been speeding if it satisfies the conditions described below under (a) or (b) or both.
(a) The vehicle's controller (driver or rider) was charged with a speeding offence; or
the vehicle was described by police as traveling at excessive speed; or
the stated speed of the vehicle was in excess of the speed limit.(b) The vehicle was performing a manoeuvre characteristic of excessive speed, that is:
while on a curve the vehicle jack-knifed, skidded, slid or the controller lost control; or
the vehicle ran off the road while negotiating a bend or turning a corner and the controller was not distracted by something or disadvantaged by drowsiness or sudden illness and was not swerving to avoid another vehicle, animal or object and the vehicle did not suffer equipment failure.Fatigue
The identification of fatigue as a contributing factor in road crashes similarly cannot always be determined directly from police reports of those crashes and the following criteria are used to assess its involvement. Fatigue is considered to have been involved as a contributing factor to a road crash if that crash involved at least one fatigued motor vehicle controller. A motor vehicle controller is assessed as having been fatigued if the conditions described under (c) or (d) are satisfied together or separately.
(c) The vehicle's controller was described by police as being asleep, drowsy or fatigued.
(d) The vehicle performed a manoeuvre which suggested loss of concentration of the controller due to fatigue, that is the vehicle traveled onto the incorrect side of a straight road and was involved in a head-on collision (and was not overtaking another vehicle and no other relevant factor was identified); or
the vehicle ran off a straight road or off the road to the outside of a curve and the vehicle was not directly identified as traveling at excessive speed and there was no other relevant factor identified for the manoeuvre.
One well-known thing about stats is that the collection and sorting process has a great effect on the overall results -- this must be kept in mind at all times. The RTA collection process does not distinguish (a) from (b), nor does it distinguish (c) from (d). The result is that for any accident where no one has any real idea of what went wrong, the accident will be classified as either speeding or fatigue based on the crash situation. For example, a driver who has a surprising brain hemorrhage and blacks out then drifts into an oncoming truck and is killed will be put down as "fatigue" in class (d) unless the coroner can actually find the hemorrhage in what is left of the driver's brain (and even presuming the coroner does make the effort to closely sift through).
Similarly, suppose a particular corner has a rough edge, no curb and no safety barrier. If 99 cars go round this particular corner at 60kph and keep clear of the edge with no problem then 1 car goes round the same corner at the same speed but a wheel comes in contact with the rough edge and the driver skids and loses control of the vehicle then it will be calssified as (b) and speeding will be the cause of the accident (regardless of the fact that the driver was doing the speed limit).
In summary, all undiagnosed medical conditions and undiscovered distractions turn into "fatigue", all unnoticed road conditions turn into speeding.
Words you won't find in any of the RTA statistics reports: "barrier" (as in safety barrier), "bitumen", "curb", "dirt", "edge", "gravel", "gutter", "kerb", "pothole", "sealed", "tar" (as in road surface), "unsealed", "visibility". It seems that we don't need any statistics that might be able to relate road maintenance and road construction quality to accidents on that road. Such statistics might put our government in the position of taking blame unto themselves rather than finding a suitable marginalized group and making scapegoats out of them (hey, blame the P-Platers, easier and cheaper than fixing any roads).
This work is licensed under a Creative Commons License.