[net.origins] Making the model fit the data?

beth@sphinx.UChicago.UUCP (Beth Christy) (09/03/85)

               [Massage this line with your replacement]

From: csdf@mit-vax.UUCP (Charles Forsythe), Message-ID: <717@mit-vax.UUCP>:
>It's
>funny, when a model doesn't work, science tends to look at the model and
>fix it. That way, accurate models are maintained.

I don't know how religiously (:-) you believe this statement, but I sort
of feel called to point out that science, like religion, is carried on by
people, and people aren't always that willing to toss out their beliefs
(and especially not their preconceptions) just because they don't match
the available data.  Even scientists occasionally ignore or fudge data
that doesn't fit (although it may often be entirely unintentional).  If
you consider psychology and research into child development to be at all
scientific, read Carol Gilligan's _In_A_Different_Voice_ for *truly
shocking* examples of "scientific" researchers FLAT OUT IGNORING over 50%
of the available data, even during the formative stages of their theories.
(Even if you don't consider psychology scientific, I highly recommend the
book.)

It's true that during working hours scientists are much more open to
alternate models than most folks, and to the degree that they're able to
maintain that, accuracy is more likely in science than in other areas.
But I just wanted to remind folks that they're not perfect at it.

-- 

--JB       (Beth Christy, U. of Chicago, ..!ihnp4!gargoyle!sphinx!beth)

		"Oh yeah, P.S.,
		 I...I feel...feel like...I am
		 in a burning building
		 And I gotta go."            (Laurie Anderson)

gjerawlins@watdaisy.UUCP (Gregory J.E. Rawlins) (09/05/85)

In article <1088@sphinx.UChicago.UUCP> beth@sphinx.UChicago.UUCP (Beth Christy) writes:
>[.......]
>I don't know how religiously (:-) you believe this statement, but I sort
>of feel called to point out that science, like religion, is carried on by
>people, and people aren't always that willing to toss out their beliefs
>(and especially not their preconceptions) just because they don't match
>the available data.  Even scientists occasionally ignore or fudge data
>that doesn't fit (although it may often be entirely unintentional).  If
>you consider psychology and research into child development to be at all
>scientific, read Carol Gilligan's _In_A_Different_Voice_ for *truly
>shocking* examples of "scientific" researchers FLAT OUT IGNORING over 50%
>of the available data, even during the formative stages of their theories.
>(Even if you don't consider psychology scientific, I highly recommend the
>book.)
>[.....]
>--JB       (Beth Christy, U. of Chicago, ..!ihnp4!gargoyle!sphinx!beth)

    Although i hate giving ammunition to creationists <half smile>
it's important to remember that scientists are in fact human (just
like creationists - draw your own conclusions). Since i don't really
consider psychologists scientists (simply because there are too many
variables in human interactions to control so it makes the concept
of a repeatable experiment vacuous) here is a book that shows
that (hard) scientists sometimes even _fabricate_ results out of 
whole cloth:
    Betrayers of the Truth - William Broad & Nicholas Wade
    Simon & Schuster 1982 ISBN 0-671-447469-6
Most of the cases documented tend to be in medicine then biology
then chemistry then physics and astronomy only a few are actual
fabrications and most of those are in medicine, but it does make
the point that scientists are fallible. It would be easy for a
creationist to use the examples in this book and "extrapolate"
that we may as well discard evolution since it's based on experiments
conducted by people with (on the whole) a vested interest in
seeing the "expected" result. This is a non-trivial problem
to which i have no convincing answer. Help anyone?
	greg.
-- 
Gregory J. E. Rawlins, CS Dept., U. Waterloo, Waterloo, Ontario N2L3G1, Canada
CSNET:gjerawlins%watdaisy@waterloo.csnet                   BELL:1-519-884-3852
ARPA :gjerawlins%watdaisy%waterloo.csnet@csnet-relay.arpa  BELL:1-519-886-4671
UUCP :...!{allegra|clyde|linus|utzoo|inhp4|decvax}!watmath!watdaisy!gjerawlins

csdf@mit-vax.UUCP (Charles Forsythe) (09/06/85)

Beth Christy writes:
>From: csdf@mit-vax.UUCP (Charles Forsythe), Message-ID: <717@mit-vax.UUCP>:
>>It's
>>funny, when a model doesn't work, science tends to look at the model and
>>fix it. That way, accurate models are maintained.
>
>I don't know how religiously (:-) you believe this statement, but I sort
>of feel called to point out that science, like religion, is carried on by
>people, and people aren't always that willing to toss out their beliefs
>(and especially not their preconceptions) just because they don't match
>the available data.

I'm not going to disagree with that, but there are some distinctions to
be made. I will, once again, bring up the "ether" theory.

It physicists didn't abandon that because of stubborness, they would
have to come with some way of explaining contradictory data coming from
special relativity. It was too much trouble, so they accepted the
wave-particle theory and it's worked REALLY WELL for a long time.

Meanwhile, theoretical physics can be a bit hairy, because some things,
such as black holes, are difficult to study.

The theory of creation from big-bang through evolution has worked very
well, but is fuzzy enough to come under scrutiny from certain
intellectuals such as Ted Holden.

It seems to me though, that if the theory of evolution et al were really
that big a failure such creationism would do ANY good to adopt (ie
explain anything new) it would be more widespread. 

A creationist lost a lecture at MIT two years ago. The creationist
said,"I am only going to argue from science!" so the evolutionist
said,"I am only going to argue from religion!" and he won. No
creationist could ever win tenure at MIT. Ha.
-- 
Charles Forsythe
CSDF@MIT-VAX
"We pray to Fred for the Hopelessly Normal
	Have they not suffered enough?"

from _The_Nth_Psalm_ in _The_Book_of_Fred_