On Fri, 1 Sep 2000, bill fancher wrote:
> on 8/31/00 9:44 PM, Miles Jackson at cqmv at pdx.edu wrote:
>
> > This is a pretty generous summary of the rise and fall of behaviorism
> > in psychology as a scientific discipline.
>
> Compressed perhaps, but I would disagree that it's overly generous.
>
> > In fact, Skinner's ideas
> > fell out of favor in the 1960s and 1970s among psychological researchers
> > because a growing corpus of cognitive research clearly disconfirmed
> > basic behaviorist tenets.
>
> The "growing corpus" was the result of research dollars being funneled into
> cognitive psychology. Give me control over the cash and I'll produce a
> growing corpus of research in any field you name.
>
> > One example: Festinger's research on
> > cognitive dissonance clearly demonstrated that people are often
> > motivated more by the need for cognitive consistency than by
> > positive reinforcement.
>
> I assume you are talking about "COGNITIVE CONSEQUENCES OF FORCED COMPLIANCE"
> by Leon Festinger & James M. Carlsmith
> <http://www.yorku.ca/dept/psych/classics/Festinger/index.htm>
>
> Festinger's results don't seem to me to demonstrate anything clearly.
>
> To sum up the experiment. Subjects were asked to do a boring task. Some
> subjects were then paid to lie to someone by saying the task was
> interesting. The amount offered was either $1 or $20. All subjects were
> asked to rate the task and the experiment afterwards.
>
> Responses of 3 of the people who were offered $1 (13%) were not reported
> because either they "told the truth" or were going to "tell the truth".
> Presumably they "reduced their cognitive dissonance" this way and hence were
> unsuitable subjects.
>
> People who lied for $1 afterwards rated the task as more enjoyable than the
> ones who lied for $20.
>
> Isn't it just possible that the people in the $1 group whose responses were
> counted REALLY DID find the test less boring and hence didn't feel (as much)
> that they were lying and hence didn't "tell the truth", and hence weren't
> booted out of the results? Meanwhile the people who REALLY DID find the task
> boring were insufficiently motivated to lie by the smaller reward, and hence
> "told the truth" and hence were not counted. And wouldn't that skew the
> results in just the right way as to "confirm" the "theory"?
>
> Unfortunately, the responses that were not used in the calculations were not
> reported at all, so we may never know for sure. But the above consideration
> indicates that there was selection on the $1 group that did not occur for
> the $20 group.
>
This is a perfect demonstration of cognitive dissonance: when research findings are not consistent with our preconceptions, we go over the findings with a fine tooth comb to find some detail that might possibly explain the results so that we don't have to change our preconceptions. Random assignment is going to minimize the between group attrition you're worried about here. And even if you can convince yourself that the possible confound of attrition is a fatal flaw in this study, get out your reading glasses--you have a few hundred other research findings in social psychology with similar research results to find fatal flaws in. Let's see-- which is more plausible: hundreds of studies that converge on the conclusion that humans are motivated by cognitive consistency, or the idea that all these studies are flawed in some way or other?
> How do you account for the fact the behavioral curricula outperformed
> cognitive curricula in virtualy every measure reported?
>
> How do you account for the fact that cognitive curricula were recommended
> despite their poor performance?
>
> How do you account for the fact that the cognitive curricula were actually
> harmful to the students and yet were still adopted?
>
I'm not sure who you are arguing with here. Yes, behavioral interventions are often effective; as I stated earlier, it is clear that operant conditioning can influence behavior. But to claim, as Skinner did, that behavior is completely determined by reinforcement contingencies is patently false.
Miles