| Friday,
August 20, 2010 There are a couple of items brought to my
attention by mail that seem very much worth looking into:
Cyberattacks via automobile worms might cripple
traffic flow...
Jerry,
Stumbled across this quite by chance this morning.
Plausible.
http://earlywarn.blogspot.com/
2010/08/city-crippler-car-worms.html
-John G. Hackett
It's worse than it appears on first look. All our modern "smart cars" are
vulnerable to wireless attacks that could stick the accelerator at full on,
or lock the brakes, or disable all the cars on the road at rush hour.
I don't know how vulnerable our military vehicles are to this sort of thing.
I wonder how long it will be before the first anti-auto electronic terrorist
attacks?
====================
The other was from Mike Zawitowski, which reminds me of
something we have written about before.
<http://www.sciencenews.org/view/
feature/id/57091/title/Odds_are,_its_wrong>
When scientists make mistakes using
statistics
And is quite frightening. It says among other things
There is increasing concern,” declared
epidemiologist John Ioannidis in a highly cited 2005 paper in PLoS
Medicine, “that in modern research, false findings may be the
majority or even the vast majority of published research claims.”
Ioannidis claimed to prove that more than
half of published findings are false, but his analysis came under fire for
statistical shortcomings of its own. “It may be true, but he didn’t prove
it,” says biostatistician Steven Goodman of the Johns Hopkins University
School of Public Health. On the other hand, says Goodman, the basic
message stands. “There are more false claims made in the medical
literature than anybody appreciates,” he says. “There’s no question about
that.”
Nobody contends that all of science is wrong,
or that it hasn’t compiled an impressive array of truths about the natural
world. Still, any single scientific study alone is quite likely to be
incorrect, thanks largely to the fact that the standard statistical system
for drawing conclusions is, in essence, illogical. “A lot of scientists
don’t understand statistics,” says Goodman. “And they don’t understand
statistics because the statistics don’t make sense.”
The misuse -- indeed culpable misunderstanding -- of statistical analysis
applies even more to the "social sciences" which are shot through with
"significant" findings that can lead to far reaching consequences, but on
analysis are not repeatable or reliable and are more likely to have arisen
by chance than by any actual real world mechanisms. One of the big problems
with analyzing the "HIV = AIDS" hypothesis came from the nature of the
evidence.
When I was in graduate school in psychology we were required to take a
statistics course given by the Psychology Department. It was the terror of
many students, particularly the clinical psychology students. It used a
textbook written by the professor who taught it, and was in fact a far
better and more meaningful course than those taught in most university
social science departments. It was also fairly trivial, largely consisting
of learning cookbook techniques, including tricks that could be used on a
Monroe Calculator so that you could get a number of required calculations
done with one entry. (Ah, if we had only had small computers in those days!)
There were lectures on "the null hypothesis" and such like, but there was
almost no discussion as to how probability works (or indeed if it works at
all). Those who wondered about all this were told they could go read L.
Savage's Foundations of Statistics. Few got past the first chapter. (You can
find substantial parts of the book at Google Books if you want a sample.)
In my case I wanted to study with Paul Horst, and he required his
students to go to the Math department to take advanced math courses, which
turned out to be the best thing that ever happened to me since I was
required to learn all the mathematical tools of Operations Research
(including of course probability theory and linear algebra), but that's
another story. But even in the math department the probability courses were
more concerned with proving theorems than in understanding the relationship
between data and reality. Fortunately I was eventually able to get
through Savage, and that led to Bayes, and eventually to a realization that
it's very difficult to prove general rules when you don't have a full
understanding of the mechanisms involved. That is: scientific method
consists of formulating falsifiable hypotheses. The problem is, unless you
know (or are testing) the exact mechanisms involved, how do you know that an
hypothesis has been falsified? The data are often inexact and are sometimes
based on inferences from other inexact data. Clearly an example here is
Global Warming: what data do we need to falsify the hypothesis? But that is
probably simple compared to some social science theories.
Now of course much "scientific" theory is never subjected to the test of
falsification to begin with; now suppose that a good half of the "findings"
of science probably are false.
Clearly the better we understand a phenomenon, the more narrow a
hypothesis we can make, and thus the more refined our definition of data.
Also clearly the better the quality of our data, the more closely we can
test the falsity of our hypotheses.
I'll stop rambling now. The important point here is that a lot of
"science" isn't very good, because the hypotheses being tested don't make
sense, and attempting to resolve the problems in experimental design when
the data are fuzzy will make you head ache. If you want to know more on
that, start with the Wiki entry on liklihood
http://en.wikipedia.org/wiki/Likelihood_principle and read it (just
assume the math is correct; you don't really need to be able to follow the
equations) and see what happens when you're trying to measure simple voltage
fluctuations all of which are below 100 volts with a 100 volt meter...
=================
We have become a nation of taboos. Dr. Laura is leaving her program
because she pronounced the dread N-word on the air. Not as an epithet
directed toward some human being, which would at least be rude and which has
implications, but simply pronouncing it. Jennifer Anniston used the word
"retard" on the air, and now she must be hounded into submission. You can't
say "stupid" even in casual conversation. I suppose no one should ever read
Charles Erskine Scott Woods Heavenly Discourse because one of the chapters
concerns what happens when God becomes disgusted and decrees that the stupid
shall not enter Heaven. So must that book be banned. Of course it was banned
in its day for being blasphemous, and the very people who would ban it now
stood up in its defense.
Saying those awful words is hurtful, and mean, and anyone who says
'retard' must be destroyed, and we know all this because we have the
statistics on the number of people who are harmed by people saying things
like that, and it's all science, and after all there are no stupid people,
and no one is retarded. This is America where all the children are above
average. And everyone is entitled to be protected from being offended. It's
offensive for people even to know words like retard and stupid.
"Oyez! Oyez! Oyez! All persons having business before the Honorable, the
Supreme Court of the United States, are admonished to draw near and give
their attention, for the Court is now sitting. God save the United States
and this Honorable Court!" How long will the Marshall be allowed to continue
that insidious practice?
==============
DH asks "So what is a
'disorder' or 'syndrome'?
That's easy - it's anything
that an insurance company will pay a "licensed professional" to "treat".
Any wonder that they breed
between releases of the DSM?
David Smith
==============
Friday TOP
Current Mail
|
