On 05 Oct 2017 15:22:14 +0100 (BST),
ijac...@chiark.greenend.org.uk
(Ian Jackson) wrote:
>In article <or5e3m$120a$
1...@gioia.aioe.org>,
>Martin Brown <'''
newspam'''@nezumi.demon.co.uk> wrote:
>>On 05/10/2017 12:31, Ian Jackson wrote:
>>> Mandatory seat belts did nothing for the safety of vehicle occupants,
>>> and increased risk for those outside vehicles.
>>Tell that to the guy who was garrotted by his laminated windscreen in a
>>high speed crash on the A19. He wasn't wearing a seatbelt and paid for
>>it with his life (pre airbag era admittedly).
>
>This kind of argument from anecdote is of course incredibly
>unreliable. (The proportion of helmet-wearing cyclists who think a
>helmet saved their life is so large if they were all right, each
>unhelmeted cyclist would have been killed many times over. Perhaps
>there is a similar statistic for seat belts.) Did that crash occur
>before the mandatory seat belt law ? If not, then it is not an
>argument for the seat belt law of course!
>
>If we actually want to know whether a new "safety" law is a good idea,
>or whether it is effective, we need to look at the whole population,
>not at individual cases.
http://www.roadsafetyobservatory.com/HowEffective/vehicles/seat-belts
offers a comprehensive list of studies more recent than the much
touted Isles report (which pre-dated the legislation and the
conclusion of which was predictive (and now more up to date
information is available seemingly wrong) rather than objective.
"A meta-analysis of 29 studies of seat belt use was published in 2009
(Elvik et al 2009). .(Elvik, R., Hoye, A., Va, T. & Sorensen, M.,
2009. The Handbook of Road Safety Measures, Second Edition.) . It
found that seatbelts were effective at preventing injury, and were
more effective at preventing more severe injuries. (TRL Client Report
CPR1818 p4.1.1) "Elvik & Vaa conducted a meta-analysis of 29 studies
on the effectiveness of seat belts from different countries, including
USA, Australia and several European countries (Elvik & Vaa, 2009). Due
to the very large sample sizes resulting from the combination of
several studies the best estimates are considered by TRL as robust
numbers with acceptable confidence intervals. The results are also
frequently cited by other experts, for example in (FIA, 2009) and
(Road Safety Observatory, 2013)"
>The statistics (and not just those in the Isles Report) show that the
>effect of seat belt law is: for every person in a vehicle who is saved
>by the seatbelt that they are now wearing because of the law, there is
>another person who dies in an extra crash (or a worse crash). Our
>best guess is that these extra crashes occur because the driver of the
>vehicle felt safer because they were now wearing a seat belt.
"A study from 1985 was published by the Department of Transport and
was commissioned to examine the statistical evidence on the effect of
the seat belt law on road casualties. (Durbin and Harvey1985)
Based on the model, the authors predicted that the seat belt law saved
the lives of 241 drivers in 1983 and 270 in 1984. Similar estimates
were made for front seat passengers, where an estimated 147 lives were
saved in 1983 and 181 in 1984.
The authors commented on a 7.8% increase in pedestrian deaths compared
with what the model would have predicted. However, they suggested that
this was due to the annual reductions in the number of pedestrian
deaths not being as large as in recent years, rather than being due to
the seat belt law.
They noted that 1984 had the lowest number of pedestrian deaths than
any other year in their data and that the number of pedestrians killed
or seriously injured showed no apparent increase.
The model also found that there was a 4.8% increase in the number of
cyclists killed or seriously injured following the introduction of the
law. However, this was not a statistically significant finding and may
have been due to chance. The comparatively low numbers of cyclists
killed or seriously injured in each month means that small differences
in the numbers can lead to large percentage changes. Similarly, the
model could not be modified to incorporate data on weather, which
influences cycle use.
"A study into the long term effects of seat belt wearing was published
in 1989. This was intended as a follow up to the study by Rutherford
et al and investigated the validity of some questions which that study
did not address. This new study included data over a six year period
between 1980 and 1985. (Tunbridge 1989)"
"The study found many of the same associations as the previous study
by Rutherford. There was a 20% reduction in the number of drivers
admitted as in-patients in 1983-85: there were 3,804 injuries compared
with 4,768 in 1980-82. There was a larger fall in the number of front
seat passengers admitted as an in-patient:1,599 injuries in 1983-85
compared with 2,396 injuries in 1980-82. This was a reduction of 33%.
Both of these falls were statistically significant.
There was a reduction of all serious injuries in 1983-5 compared with
1980-82. There were 787 drivers admitted to hospital with serious
injuries in 1980-82 compared with 633 in 1985-86, a 20% reduction. A
similar fall was seen for severe injuries to front seat passengers
with 16% fewer being admitted to hospital: a fall of 356 to 298
between the two time periods.
The study found significant reductions in the already relatively low
number of serious injuries to the head, such as skull or facial
fractures. There was an increase in sprained necks and fractured
sterna following the introduction of the law, although compared to the
number of injuries prevented the increase was extremely small.
The authors also investigated whether there had been any changes to
the number of casualties amongst non-car occupants. This was performed
as a follow up to the previous publication by Durbin andHarvey, which
had suggested that there were modest non-significant increases in the
number of vulnerable road users following the seat belt laws, but that
further research was required based on hospital data.
Based on analysis of the new linked data set, the authors found that
there was a 2% reduction in the number of pedestrian casualties; there
were 5,478 in 1980-82 and 5,357 in 1983-85. This fall was not
statistically significant. There was a slight increase in pedal cycle
casualties from 807 to 837 between the two time periods, although this
was also not statistically significant. The authors also looked at
pedestrian and cyclist casualties from accidents involving cars, and
again found no significant change in the numbers between the two time
periods."