>>> "J. Barkley Rosser, Jr." <[EMAIL PROTECTED]> 05/11/99 04:43PM >>>
Charles,
1) I believe that I have been using the term "mistake'
not "accident" with respect to the Chinese embassy
bombing. Certainly that is what I have meant, even if
I have said otherwise.
CB: I'll bite. What is the difference between a mistake and an accident ?
((((((((((((
I have repeatedly said, that,
mistake or accident, the bombing of the Chinese
embassy was not justified and should be condemned,
along with the bombing of other targets in Yugoslavia,
which has resulted in many more innocent civilian deaths
than did the one of the Chinese embassy, although that one
is certainly going to lead to much more negative consequences
for the US.
CB: Agree
(((((((((((((
2) I am only going to respond to your remarks about
econometrics on pen-l because I am already overquota on
lbo-talk and the econometrics discussion has been here.
Properly done (according to the textbooks) an econometric
test should be of a hypothesis derived from some economic
theory (however bourgeois, alienated, misguided, or ideologically
twisted). Thus, when Card and Krueger set out to look at the
relationship between increases in the minimum wage and the
impact on employment, they were looking at something that,
although subject to a certain amount of randomness, nevertheless
supposedly had a causal relationship. When they produced
serious evidence that the expected relationship was not there,
it created quite a stir.
CB: Not to be picky, but isn't there a difference between causation and and
statistical correlation ? Seems they would be setting out to show a statistical
correlation that would be some empirical evidence of a causal connection. Anyway, I
read on.
((((((((((((((((
Everything has noise, even if governed by probability
distributions a la quantum mechanics.
CB: As Engels said, the laws of nature assert themselves amidst a welter of random
events.
((((((((((((((((
The embassy bombing
looks pretty noisy to me, whether mistake, accident, plot by
Serbian spy, plot by super-sneaky Clintonites too smart for
their own good, or plot by secret cabal of ultra-militarists, or....
CB: Just to be clear, what would be an example of this not being part of the noise ?
If they were bombing the Chinese embassy ON PURPOSE, INTENTIONALLY, not neglegently or
recklessly, they might use the excuse that it was an accident or mistake to coverup or
for plausible denial. But as you say, the U.S., is culpable whether it was on purpose
or reckless.
((((((((((((((((
BTW, I agree with whoever it was (Peter Dorman?) who
said that a lot more effort should be put into getting better
data than on developing ever fancier ways of torturing the data
that we have. One of the strengths of the Card-Krueger study
was that they developed their own data set.
I also agree with Peter (I think) that a serious look should
be made at the McCloskey critique of the standard practice
of econometrics, along with the related Leamer "let's take the
'con' out of econometrics" stuff.
Finally, I would note that econometrics has been used to
expose bogus baloney from racist ideologues. The example
that comes to mind most sharply is the expose by Arthur
Goldberger and Charles Manski of the fallaciousness of the
Murray-Hernnstein _The Bell Curve_.
Chas.: I'm not opposed to the use of statistics, including econometrics. But if you
were using probability theory on this bombing, it would be a pretty good correlation,
as some others have pointed out, no ? I'd have to think about how you set up the
correlation, I guess.
On econometrics in general, I don't have enough experience to really enter the
discussion on this thread, however, I find credible the criticisms from those who do
(probably including you) who say that it has in fact been abused by many , and there
must be an ongoing struggle by progressive economists to clean it up.
Charles Brown
-----Original Message-----
From: Charles Brown <[EMAIL PROTECTED]>
To: [EMAIL PROTECTED] <[EMAIL PROTECTED]>
Date: Tuesday, May 11, 1999 4:27 PM
Subject: [PEN-L:6684] Mistakes, randomness, accidents and econometrics
>
>
>>>> "J. Barkley Rosser, Jr." <[EMAIL PROTECTED]> 05/11/99 01:48PM >>
> BTW, for those who come up with half-baked probability
>calculations and then declare that this could not have been
>a mistake, I would remind you that the chances of the Three
>Mile Island accident happening were supposedly something
>like one-in-ten billion.
>
>((((((((((((((((((
>
>Random Barkley,
>
>Maybe it was a butterfly in China flapping its wings that caused this
mistake, this accidental bombing.
>
>Funny how econometrics is used to find order and correlation in plottings
of points that seem random. But when it comes to explaining events that look
causally connected, you tend to see it as random , an accident.
>
>As Engels said, the laws of nature assert themselves through a series of
accidents.
>
>Charles Brown
>
>
