FAO Simon Mason

wrote:
In article ,
Nick wrote:
I'm not a statistician but I do understand a little of scientific method
and a reasonable amount of probability theory.

Do you? Then I am surprised that you have just posted this.

My statement is that the claim that the cause was the 20 MPH zones
(with the clear implication that it was the limit) is the propaganda.
It is.
No it is evidence. In the general everyday world of information we base
decisions upon it is quite strong evidence it is only in a
scientific/academic way that it is lacking.

I thought that you claimed to have some understanding of scientific
method?

The evidence is the data. The claim is, at best, a conclusion. All
scientists are taught to be very careful of the distinction.


Our Data is the report, the conclusion of the report is part of our
evidence. Its not an absolute proof, but then what is?

I won't hazard a guess as to what science you have in mind but in every
day life we are in the business of making decisions based upon limited
information.

Furthermore, scientists are taught that, if two factors differ between
two experiments, never to claim that one of those factors is the cause
of any differences.

That is an ivory tower idealisation, factors always differ, scientists
make value judgements as to what is acceptable. Drug trials (even with
their multi million dollar budgets) generally have a lot of different
factors but scientist still draw conclusions about drug efficacy.


Lastly, the way that that conclusion was drawn was precisely the
converse of what all statisticians are taught is acceptable analysis.
It is vanishingly unlikely to be a mistake, given its extreme nature.

I don't understand what you mean by this comment.

Anyway AIUI the report was by a traffic engineer not a statistician.


Regards,
Nick Maclaren.

On Wed, 20 May 2009 19:41:46 +0100, "Just zis Guy, you know?"
wrote:

On Wed, 20 May 2009 10:11:57 +0100, "Dave Larrington"
wrote:

AFAIK, however, the claim that this incident was responsible for the
introduction of the 70 mph limit is an urban legend.

Probably is, as I heard the same UL about Astons being tested on the
M1 around Northampton.

Guy


You're "joining in " a lot "Lou" - trying to ingratiate yourself.

Tell us how the microsoft nym-shift kill file avoider works.

--

Passenger casualty rates by mode Per billion passenger kilometers:
Killed or seriously injured:
Pedal Cyclists : 533 Pedestrians : 384
All casualties:
Pedal Cyclists : 3739 Pedestrians : 1795

Which do you think is the most dangerous?

spindrift wrote:
http://www.neath-porttalbot.gov.uk/d...482&pr_id=3229

The 20 mph zones introduced in Sandfields by Neath Port Talbot Council
continue to make the streets safer.

How do they know that the zones make a difference?

Recent figures published show the number of personal injury accidents
continues to drop since the zone and street calming measures were
brought in more than six years ago.

So are they saying that the same zones are becoming more effective every
year (in which case how come?) or admitting that actually other factors
are coming into play, or even using the police STATS19 injury data which
no longer correlates with hospital data (i.e. the number of accidents
are not actually dropping).

Before the zone was introduced, there were up to 30 personal injury
accidents each year in Sandfields.

"up to 30"? The same trick that the big furniture stores use when
announcing sales reductions - "up to 50% off"!

Last year, that number was down to just 13.

What was it the year before that? 30? 20? 10?

End quote.

The reductions in accidents were measured entirely within the zone.

Were they compared to the reductions outside the zones? Were the
reductions outside the zones better or worse? Perhaps the zone resulted
in worse improvements than would have happened without them.

You didn't bother to read the links, did you?

The links don't give the information necessary to draw the conclusions
that have been presented.

--
Matt B

In article ,
Nick wrote:

I won't hazard a guess as to what science you have in mind but in every
day life we are in the business of making decisions based upon limited
information.

That is PRECISELY the domain of statistics, don't you know?

Furthermore, there is a major difference between making a crude guess
in a hurry (which is often justifiable) and arranging that a particular
conclusion is drawn. And it is the LATTER I am accusing them of.

That is an ivory tower idealisation, factors always differ, scientists
make value judgements as to what is acceptable. Drug trials (even with
their multi million dollar budgets) generally have a lot of different
factors but scientist still draw conclusions about drug efficacy.

As I said, I am a rusty statistician, and I have been closely linked
to medical reserahc all my life. I can assure you that I know how
such things are analysed, and it is supposed to be as I say. When they
do it the way you do, and are found out, there is an outcry (very
reasonably) and often people lose their jobs.

Lastly, the way that that conclusion was drawn was precisely the
converse of what all statisticians are taught is acceptable analysis.
It is vanishingly unlikely to be a mistake, given its extreme nature.

I don't understand what you mean by this comment.

Clearly not.

Anyway AIUI the report was by a traffic engineer not a statistician.

Then you don't understand much about how such data are collected and
presented. Hull will have had a statistician involved in the whole
process.


Regards,
Nick Maclaren.

wrote:
In article ,
Nick wrote:
I won't hazard a guess as to what science you have in mind but in every
day life we are in the business of making decisions based upon limited
information.

That is PRECISELY the domain of statistics, don't you know?

Furthermore, there is a major difference between making a crude guess
in a hurry (which is often justifiable) and arranging that a particular
conclusion is drawn. And it is the LATTER I am accusing them of.


We are always in a hurry we are always making a guess, this may vary
from project to project but the truth is resources are finite. I suspect
that the producers of the Hull report did so on a very limited budget.

You may be accusing them of arranging for a particular conclusion to be
drawn but I don't see that you have backed that accusation up with
anything substantive.


That is an ivory tower idealisation, factors always differ, scientists
make value judgements as to what is acceptable. Drug trials (even with
their multi million dollar budgets) generally have a lot of different
factors but scientist still draw conclusions about drug efficacy.

As I said, I am a rusty statistician, and I have been closely linked
to medical reserahc all my life. I can assure you that I know how
such things are analysed, and it is supposed to be as I say. When they
do it the way you do, and are found out, there is an outcry (very
reasonably) and often people lose their jobs.


I'm a consumer of statistics not a producer so I don't "do it" any way.
But for clarity which part of my statement above, or "doing it the way I
do it", was incorrect.

Stop me if I'm wrong but don't drugs companies test on cultures and
animals in order to reach conclusions about how the drug would work on
humans. When they do test on humans is it not the case that they can't
ensure a truly representative sample. In particular they haven't tested
on me before the doctor decides the drug is suitable for me? But this is
the real world, compromises have to be made.

Is it not also the case that the testing process and statistics produced
are incredibly expensive.

Somehow I just don't think Hull had a Viagra sized budget available.

As a consumer of statistics I work with what I'm given. I certainly
wouldn't throw information like this out of the pram because it wasn't
as good as it could have been.

Lastly, the way that that conclusion was drawn was precisely the
converse of what all statisticians are taught is acceptable analysis.
It is vanishingly unlikely to be a mistake, given its extreme nature.

I don't understand what you mean by this comment.

Clearly not.


I was hoping you might help me out?

Anyway AIUI the report was by a traffic engineer not a statistician.

Then you don't understand much about how such data are collected and
presented. Hull will have had a statistician involved in the whole
process.


And it was presented by a Traffic engineer are you claiming that
executive report writers always gain approval from statisticians?





Regards,
Nick Maclaren.

Nick wrote:

wrote:

[ ... ]

... you don't understand much about how such data are collected and
presented. Hull will have had a statistician involved in the whole
process.

And it was presented by a Traffic engineer are you claiming that
executive report writers always gain approval from statisticians?

Before any political action* is taken, it is important that the politicians
are sure that the conclusions are valid. Taking the bland reassurance of the
traffic engineers (a profession currently passing through a very
anti-motorist phase) that the conclusions are valid is not the way to be sure
of it.

[* "political action", well you know what I mean, even if those two words
don't sit together easily.]

In article ,
JNugent wrote:

Before any political action* is taken, it is important that the politicians
are sure that the conclusions are valid. Taking the bland reassurance of the
traffic engineers (a profession currently passing through a very
anti-motorist phase) that the conclusions are valid is not the way to be sure
of it.

That is correct, except for one aspect. The failure to do that is why
we now have this dogma that cycle helmets save lives, that the best
way of encouraging people to cycle rather than drive is lines around
the gutter and the right to use the pavement, and so on.

The aspect that is wrong is that traffic engineers are currently passing
through a very anti-motorist phase. They aren't, at least not like
that. They are very anti-PRIVATE-motorist in cities, but that seems
to be about all. Even there, they favour buses and taxis over bicycles
and often over pedestrians. There is no sign that they want to encourage
anything except motorised road transport outside cities or between
outside and inside cities.

The ridiculous thing is that this even applies to commuting into the
smaller cities (like Cambridge), where the proposed solution is mainly
park-and-ride (even for 3-5 mile trips!) and bus routes along certain
profitable highways. That's generally true elsewhere, too.


Regards,
Nick Maclaren.

Just zis Guy, you know? wrote:
On Sun, 17 May 2009 17:24:34 +0100, "Simon Mason"
wrote:

In my experience that it not true. My wife has 9 points and now sticks to the
limit all the time as she will lose her licence if she continues to be caught
by cameras. The cameras have forced her to stick to the limit and so are
doing their job.

Those with multiple speeding convictions are more likely, mileage
adjusted, to be involved in collisions.

Is that assertion based upon research using recent data which takes into
account the millions of technical camera "convictions" of "normal"
drivers, or is it based on old data (if it is indeed based upon any
data) which is more likely to mainly contain drivers who police officers
actually thought were driving dangerously when they were stopped?

So it does appear that
cameras are indeed targeting the dangerous drivers, albeit in a rather
crude way.

That would depend upon those other factors.

--
Matt B