In a message dated 1/7/98 10:56:20 AM, you wrote:
Robin Hanson writes:
>Curt Adams writes:
>>Well, my bachelor's is in economics, and I would indeed say that social
>>science is short in empiricism. The harder sciences are full of
>>empirical tests so convincing that they shove unpopular theories
>>down opponent's throats, but such things are rare in social science.
>
>1. More total research effort has gone into the "harder" sciences.
>2. Social processes may just be more complex and harder to understand.
>3. People have much stronger preconceptions about social things.
I agree with all 3 claims. All the "harder" sciences permit extensive
designed experiments except astronomy. Astronomy gets by as it has
such an enormous library of things to observe and its objects of study
don't interact like people. Really convincing evidence usually requires
carefully designed and thorough experiments. One of my favorite
examples was the guy who proved pellagra wasn't transmissible by
eating and snorting virtually every obtainable bodily secretion from
afflicted individuals, combined with treatment by dietary therapy.
You just can't do that kind of thing in economics, and it's no fault
of the economists.
>I think the current distribution of effort into various sorts of empirics
>and theory in economics is roughly about right.
Can't judge this; not enough evidence.
>2. Undergraduate courses in computer science and especially economics
>typically show lots of distain for their students. Profs don't think
>these students can handle enough math to show them the good stuff, and
>they prefer to take as grad students those who did undergrad work in
>math or some difficult science. I think this is relevant for Curt's
>experience as an UG econ major.
True enough; my professors generally groveled in apology for such a
monstrous demand as taking a derivative, at the University of Chicago,
even. Nonetheless, it's the shortage of empiricism which disappointed
me, not the shortage of math. A lot of biology is based on relatively
simple math; generally I'd say the math in econ was more complex than
that in bio. But since the simpler mathematical models of biology
often carry substantial experimental support, I feel like I have
something to take home at the end of the day. A lot of the math in
econ is like Arrow's great demonstration that if people have rational
evaluations of gambling outcomes, then their behavior can be precisely
described with cardinal (numerical) utility functions. But people
aren't rational with their evaluation of gambling outcomes, so where
does that leave us?
My attitude towards math in science is governed by the "extraordinary
claims require extraordinary evidence" principle. The fancier the
math, the more claimed (usually) and the more empirical support
I expect.
Received on Wed Jan 7 18:28:58 1998
This archive was generated by hypermail 2.1.8 : Tue Mar 07 2006 - 14:45:29 PST