Re: boundary permeability (was Behaviorism)

Posted by Eric Charles on May 06, 2010; 2:22am
URL: http://friam.383.s1.nabble.com/Behaviorism-tp5003979p5012211.html

Glen sayeth" In other words, I (currently) don't care whether
you increase
the rate of coherent sentences in my dad _alone_,
regardless of whether it's
30-60% or 1-100%.  What I care about is
whether you increase the rate of
coherent sentences in a statistically
significant portion of the population
of patients by X-Y%."



You are correct about dancing the line
between ideographic and generalizable work. Behavior analysis doesn't really
lend itself to the sort of analysis you are looking for. Many people slammed
Skinner for the lack of statistics in several of his books. He replied that
statistical significance was just a poor man's replication. If he can put 50
rats, 50 chickens, 50 monkeys, and 50 children under the same schedule of
reinforcement, and get the same learning curves for each and every one of them,
what does he need statistics for (he argued).

If I was faced with a
patient in the nursing home who had a problem behavior, I would do a functional
analysis, which is a set of procedures to determine what was reinforcing the
particular behavior in question. Then I would design a procedure tailored to
that specific case. Then, if I did my initial analysis well, procedure based on
that analysis would be successful. 

People aren't as unique as we like
to think, so over the course of a career, I am likely to employ the same
results several times - but in another sense the treatment is always customized
to the individual situation. 

Note, this is partially because clinical
cases are "found". In a Skinner box I "create" and the creation is so
ridiculously replicable that statistics are silly (though many people DO use
them, as the norms of the field often demand it). In the clinical case I am
making an educated and scientifically informed guess as to the developmental
history of the behavior in question - if my guess is correct, the treatment
will work. 

The one place where I know that statistics of the type you
are looking for exist is in areas like clinical treatment of depression. I know
that behavioral therapy (broadly construed) performs as well or better than
cognitive oriented therapies in most studies. That is, if you take a bunch of
depressed kids and put them in behavioral therapy, you get fewer depressed kids
afterwards. Of course, that is mixing and matching theoretical approaches is
potentially icky ways. I don't know the exact stats, but I know they exist. If
such stats would answer your question, I will dredged some up.

As for
the Carr study, I am unfamiliar with it, but you state that 38% of the people
didn't need shocks any more at all, I will presume that at least that many were
responding positive but needed the shocks to stay that way. If I'm right that's
close to 70% of an otherwise un-help-able group being helped. Potentially, that
is a really good result, although I'll bet they could have achieved the results
through more socially acceptable(but still behavior analytic) means.


Eric

============================================================
FRIAM Applied Complexity Group listserv
Meets Fridays 9a-11:30 at cafe at St. John's College
lectures, archives, unsubscribe, maps at http://www.friam.org