[TowerTalk] Porqupines

[Jim Lux]

Empiricism is good, and Norman has given a good example of anecdotal data.
However, it doesn't really address Kelly's question

In order to say that the porcupine "works", you also have to eliminate that
it was something else you did or that happened that resulted in the lack of
strikes. There are several alternative explanations which Norman hasn't
addressed (but actually, he might be able to):

1) Something else in the installation process was actually responsible for
the reduction in lightning. This is what someone else mentioned about "you
also need to install good grounding, etc.
2) What you're really reporting might be a lack of lightning induced damage,
not a lack of lightning hits.  Unless there's some mechanism for actually
counting strikes (which is what the FAA study did), you're inferring that
there's less lightning.
3) There could be some freak weather condition so that there just didn't
happen to be the same lightning frequency.  This one crops up all the time
with tornado protection schemes.  While tornadoes are pretty common,
overall, they're very, very unlikely in any given spot.  Lightning is the
same.   In a well designed experiment (which rarely occurs with anecdotal
reports), you'd do something like have half the boats with no porcupine (but
everything else the same), and change them over periodically (to eliminate
the odds that some captains like to sail into storms and others don't, for
instance).  Norman mentioned that other boats did get hit, so he's reporting
good data there.

I think the real problem is #1 or #2 haven't been excluded.  Norman's data
looks pretty good for #3 (assuming that the boats were relatively similar in
size and fishing area).

The porcupine thing is one of those areas where they're making an
extraordinary claim so they need extraordinary proof.  There's no "good"
physics reason for them to work (as has been described in the literature,
and to some extent here on the list). That in itself isn't enough to say
they "can't work" (lots of things that physics doesn't explain at first),
but it also means that empirical data and experiment design needs to be that
much more rigorous.  It's been said that truly great physics doesn't come
from experiments confirming what you already think is the case, but from the
experiment that provides results like "Why did it do that?", and then you
have to go looking for the reason why.


Once you've got the experimental data, you're really only half way there.
As they say: correlation doesn't imply cause.  You need to have a plausible
mechanism for the effect.  There's a famous example of a study done at
Kaiser which found that women working on computer terminals had a higher
than normal incidence of miscarriage. This was widely reported as, the
radiation from the CRTs causing the miscarriage. The problem is that there
isn't any significant radiation from the CRTs, and that prompted a further
study. It turns out that the miscarriages were more likely the result of
stress.  The operators had grueling performance targets to meet, terrible
ergonomics, etc., all of which are known to result in health problems.

 I still see the "CRT radiation causes miscarriages" reported in various
places though. For instance, a few years back it showed up in an article
about workplace safety in a magazine aimed at self employed women.  Since
the corrective study discrediting that finding had been done probably 15
years before, I wrote and asked why they had written the incorrect data in
their article.  Turns out the author of the article (who had no workplace
safety or epidemiology or technical background at all) had read about it in
"Self" magazine a few years back.  The author of the Self article probably
got it from some news story, etc.,  with a chain all the way back to the
very first alarmist report (which had been subsequently contradicted in the
usual scientific method peer reviewed way, but hadn't got nearly as much
press, being fundamentally a boring "nothing found" report).

And this is why the credentials (such as they are) and background of the
author is important, particularly in the popular press.   Some authors are
much, much better at research than others. Not because they necessarily know
everything about the topic at hand, BUT, because hopefully, you can assess
whether they are familiar with how to find  the material available, and will
trace at least some back to primary sources, rather than relying on columns
in magazines which rely on newspaper articles, etc.


----- Original Message -----
From: "Norman Hockler" <norsan at bright.net>
To: "Kelly Taylor" <ve4xt at mts.net>; <TowerTalk at contesting.com>
Sent: Sunday, March 20, 2005 5:10 AM
Subject: Re: [TowerTalk] Porqupines


> The best qualifications you can get...empirical data.  I installed 3 on a
> 110 foot topsail schooner that had been hit 3 times.
> It was never hit again ( lost track of it 10 years later).
>
> Installed them on about 30 shrimp boats.  None were hit in the 5 years I
> owned the marine radio service company.
> Other boats were hit during the same period.  (the same and other
canneries).
>
> That was good enough for me.
>
> Norm N8NH
>
> At 04:41 PM 3/19/05 -0600, Kelly Taylor wrote:
> >If they work, then that's in conflict with studies by the IEEE, USAF,
NASA
> >and FAA.
> >
> >All of those organizations, in peer-reviewed papers subject to threat of
> >lawsuit, concluded there was no value to the porcupines.
> >
> >I'm curious, what qualifications do you bring to the table that the U.S.
> >military lacks?
> >
> >73, kelly
> >ve4xt