Dear Doug,
I read your message and your attachment.
There are so many flaws in your analysis
that it is difficult to know where to
begin.
Since you find road accident statistics
so compelling I shall confine myself to
demolishing all you have to say there...
There is a wonderful book by Darrell Huff
"How to lie with statistics". You seem
to have studied this carefully.
Still, since you want us all to tell
lies about the time, I suppose it is in
order to tell lies about statistics.
First, two matters where you are right:
1. The UK had year-round BST from
1968 to 1971.
2. During this period there was a
reduction in road accidents.
That is called an ASSOCIATION. You
cannot infer CAUSATION. There could
be any number of reasons for this
reduction:
a) The introduction of drink-drive
legislation in 1967.
b) The increase in seat-belt usage,
not yet compulsory but more and
more cars were being fitted with
seat belts.
c) Perhaps the winters were unusually
mild.
Those are the kinds of questions you
should immediately ask before making
ANY inference.
Next, you absolutely MUST investigate
what happened AFTER we abandoned using
year-round BST. Did road accidents
increase again?
WHY HAVEN'T YOU TOLD US?
Answer: either because you didn't look
or because you didn't like what you
found when you did look.
OK, well look at the attached file:
RoadDeathsGB.jpg
You will see that following a post-war
low around 1950, road deaths increased
relentlessly until a peak in 1965.
They then started a steady decline
which continued throughout the period
1968 to 1971 and you won't notice so
much as a kink in the line from 1965
to 1971. The slope in the three years
up to 1968 exactly matches the slope
from 1968 to 1971. BUT:
AS SOON AS WE STOPPED TELLING LIES
ABOUT THE TIME IN WINTER, THE RATE
OF FALL SIGNIFICANTLY ACCELERATED.
Using YOUR logic, we can see that during
the winters that we were telling lies,
we actually HELD BACK the reduction in
road deaths.
YOUR proposal was killing people!
Funny how the time liars don't tell
us that!
Now this is still flawed inference
and I wouldn't make this assertion
myself.
The truth is that it is very hard to
design an experiment to verify one
way or another. You cannot re-run
the same years with different time
rules.
There is, though, a much better way
to investigate. What you need to
do is to identify two U.S. towns in
the same state that are not far
apart but which are in different
time zones. They should be about
the same size and have the same
mix of population and industry.
The more easterly of the two towns
will have have its clocks one hour
ahead of the more westerly. By
your hypothesis, there should be
significantly fewer road accidents
in the more easterly of the two.
OK, go find some research data.
We can't do this in the U.K. but
you can look at Dover and Calais.
These towns are fairly close
and the clocks in Calais are
one hour ahead of the clocks
in Dover.
By your hypothesis, there should be
significantly fewer road accidents
in Calais.
OK, go find some research data.
Very best wishes
Frank
---------------------------------------------------
https://lists.uni-koeln.de/mailman/listinfo/sundial
