Re: Skewering stilted language and theory: F. Crews

["Devine, James" <jdevine@xxxxxxx>]

Title: RE: [PEN-L] Skewering stilted language and theory: F. Crews

DD wrote>
I'm assuming you mean "medical research" here; I'm
entirely unsure how you'd define the concept of a
double blind in social sciences research, most of which
is not experimental.<

FWIW, academic psychology involves a lot of experiments, as does so-called "behavioral economics."(Two of the practitioners of the latter recently received the "Nobel" prize in econ.) The guy whose office is next to mine does this kind of research. He seems very conscious of all of the niceties of experimental controls and the like.

BTW, his main conclusion is that "homo economicus" is nonsense, that people seem to have an innate sense of justice and the like (where it could be based on either nature or nurture or both). This is backed up by a lot of similar research.

>And even in the medical context, I think that the
demand that psychoanalysis use double blind tests would
be silly.  It's one thing to give someone a placebo
pill, but how in the heck do you carry on a placebo
version of a talking cure?  The only ways of designing
such an experiment that I can think of would involve
systematically lying to a patient which, aside from
being unethical in the extreme, would not even really
satisfy the double blind criterion; the "placebo"
psychologist would certainly know that he was faking
it.<

FWIW2, an article in SCIENTIFIC AMERICAN reported a few years back that a version of this kind of test was used to test Prozac and other SSRIs against "cognitive therapy" (a non-Freudian version of the talking cure). It turned out that the latter worked as well as the former, while the combination of the two was even better (if I remember the conclusions correctly).

>And then even if you somehow solve this problem (or
more likely ignore it), you're faced with the task of
trying to carry out statistical analysis of your
results in a context where it is not clear at all that
any of the fundamental assumptions necessary for such
analysis are satisfied; there is no statistics of
individual cases, however much the Austrian economists
wished that there were.

>Now you might want to define "scientific method" as
being identical with the use of statistical methods on
the results of double-blind experiments, but then it
seems pretty clear that if you do this, the charge that
psychoanalysis is "unscientific" loses all of its
rhetorical force; you've simply reduced the scope of
the term "science" until Freud falls outside it, not
pushed him out the door.  I'm not a Freudian, but nor
do I like attempts to create a "quick way" with
theories that people don't like in this manner.<

this was the point of my missive. It's hard to criticize Freud without applying the same standards to other social scientific research -- and it turns out that most other social studies don't do very well either.

BTW, in reading a critique of the "revisionist" history of the origins of the Cold War (which says that the US played a big, if not _the_ big, role in starting the CW), the critique invokes Popper, saying in effect that "these revisionists don't state their views as falsifiable hypotheses but instead argue for specific conclusions." (quel horreur!) This ignores the fact that the revisionists were trying to falsify the orthodox theory (the Reds did it) and that using Popper in this way is a way to avoid such falsification. I think Popper was used that way a lot. It's probably unfair to Popper.

Jim