----- Original Message -----
From: "Erik Reuter" <[EMAIL PROTECTED]>
To: "Brin-L" <[EMAIL PROTECTED]>
Sent: Wednesday, December 26, 2001 7:29 PM
Subject: Re: Tragedy in Israel


> On Wed, Dec 26, 2001 at 06:53:28PM -0600, Dan Minette wrote:
>
> > First of all, the suicide bombings are factually the same things as
> > the terrorist attacks.
>
> Fuzzy thinking. You need to be more careful. First of all, you again
> messed up the wording. The article said "suicide attacks". Obviously,
> a suicide attack is a generality of which a terrorist attack is one
> overlapping, but not identical, set.  It should be obvious that there
> can be suicide attacks that are not terrorist attacks. They are
> certainly not "factually the same thing".

Sure, there have been, in history, suicidal attacks that are not terrorist
attacks. The kamakazi attacks in WWII would be an example of this. But, I
was talking about actual attacks by Palestinians on Israel, in particular
the recent ones. Would you care to name one suicide attack from this infatal
that was consistant with the Geneva convention?

>
> > Why is the evidence conmpletely worthless?  If the same question is
> > asked both times, why shouldn't the shift be worthwhile.
>
> First of all, I didn't read your original use of the data as being used
> to prove a shift, but rather an absolute.

Well, it was given as a general indication of opinion. Lack of stated
methodology is a minus, but the fact that the poll was sponsered by someone
who has a track history (AP) is a plus.  It is probable that they know
something about methodology and would not lend their name to a poll that
didn't follow any decent methodology.

>Regardless, was the same question asked both times, with the exact same
phrasing, and using
> identical polling techniques?
>

Errors can be introduced by small changes in wording.  One always wants
identical wording.  But, the changes that result from small changes in
wording usually are in the range of 10% or so.

> > I know, as a engineer/scientist working with imperfect data is an
> > absolute requirement.
>
> So is accurately and precisely explaining the methodology used, and
> citing references for more detailed information. And peer review.
>

If you publish, yes that is worthwhile.  If I can supply methodogy for the
poll, then it would decrease the probable error.  That's valid. It would
increase the versimilitude.  But, even without these, a poll is better than
nothing.  What else would you base your views of the Palestinian's attitude?

> You work in a technical field. Do you really make important decisions
> based on data of ambiguous methodology, without references or peer
> review?

Yes, of course I do.  An example was when it was 10 PM, the tools are going
out at 7AM, and it will take 6 hours of work to prepare the tools.  No
matter what you do, you can't push the deadline back.  So, you give it your
best shot. You take field reports, inconsistant lab results, and make the
very best decision you can.

There is a school of thought to delay all decisions until the information is
perfect.  Most people I know who are in that school of thought, really do
little, but are spectacular Monday morning quarterbacks.  There is one
exception who is a friend of mine.  We had to pry his work out of his hands
and tell him it really was good before it was used, though.  He wanted to be
absolutely sure.

You know I tend to provide as much data as I can to back up my claims.
Indeed, I would wager that if we went back and looked at arguements made
over the last few years, I would supply more data than the average poster.
I don't think you tend to provide more data than I do, but I'd be willing to
be proven wrong.

Going back to the main point, this is one of the biggest differences between
applied science/engineering and science.  In science, one does have peer
review, documentation of methodology, etc.  In applied science and
engineering, one usually just makes the best decision possible.

>
> > AP does have a reputation for polling.  They have a track record.
> > Even if there are biases in the polls, one should expect +/-10%.
>
> How do you determine this? I can pull numbers out my ass, too. I expect
> +/- 50% due to biases in this situation.
>

>From looking at the variation of answers to polls with slightly different
wording.  AP has a reputation and a history of polling.  I've followed it,
and have seen their typical errors in Presidential polls and multiplied it
by about 2.  What is your basis for the possibility that the  real number is
15% to 115%?

BTW, I found a detailed poll that was done last June at:

http://www.jmcc.org/polls/2001/no41.htm

It gave the following answers to a question about the suicide bombongs:

What about the suicide bombing operations? Do you see them as a suitable
response in the current political conditions or do you oppose them and see
them harmful to national interests?

I see them as a suitable response in the current political conditions  68.6
I oppose it and see it as harmful to Palestinian national interests 23.1
I don't know 6.2
No answer 2.1



Its worth noting that Hammas has only 18.6% support in this earlier pollThe
change in numbers for Hammas are consistant with news reports from people in
the region.  The question was asked

> > Plus, the second poll was from a totally different organization, which
> > is Palestinian in origin.  So, we have a Western news agency poll and
> > a Palestinian poll agree, and its still worthless?
>
> Without knowing the methodology, it is worthless as far as I'm
> concerned. Maybe it is the same pollsters in both cases? Not too
> farfetched, given the sketchy details available.
>

I can see taking data with a grain of salt.  But, calling it worthless?  I
would like to suggest that this frees one to hold whatever opinion one wants
to.

I have now given methodology for an earlier poll.  The questions don't look
all that weird.  I suppose that you can insist that I get the origional
questions and translate them myself, but I think that this does quilify as
evidence of Palestinian attitude.

Let me ask you, have you got contrary evidence?

Dan M.

Reply via email to