[From Dick robertson] (951003.1840CDT)
I got a post from Brian D'Agostino that my post, "Self Research, Brian's
Research and PCT method," had not got on to the net, so I am re-posting it
herewith. (Some of it might be superfluous in the wake of Tom Bourbon's
carefull analysis of Brian's thesis as just relayed by Rick Marken.)
Raking up an old issue or project sometimes leads to new movement. I had
long ago given up on gaining any further mileage with the "Self as Control
System" project David Goldstein and I did these many years ago already.
But a recent remark by [Brian D'Agostino (950908)]
I am now planning follow-up psychological interviews to begin testing the
model sketched in my article.... (I am familiar with Dick
Robertson's technique of disturbing the subject's self image, but
that may not be socially appropriate for my purposes. Can you
imagine me calling George Bush a wimp in an interview to test my
theory that he controls a macho image of himself?
set me thinking about what that study was about and how it might be
expanded. And yes, I can imagine Brian calling George Bush a wimp in an
interview to test his (Brian's) theory that he (George) controls a macho
image of himself; however, that would not be the way to do it and that
wasn't what I did with the "Testing Self as control-system" study. Let me
give the background and then say how I think it might be done.
What Dave Goldstein & I did - in the "breakthrough study" - was encourage
people to describe themselves and then pick the first self-attribution they
gave and contradict it, with, "Oh, no you're not like that."
I called it the "breakthrough study" because it was the fourth attempt.
The first three were attempts to blend PCT with traditional methodology.
We started with a study not very different from the one by Frey and
Stahlberg (1986) (Pers. & Soc Psych Bul. 12(4), 434-41) that Phil Runkel
analysed so painstakingly in Casting Nets. That is, we got people to check
off self descriptive adjectives, then a week later gave them back composed
personality-descriptions with some deliberate errors to constitute
(hopefully) disturbances to the perceptions they were controlling about
their own seld-image. The first round came out, statistically, in support
of the hypotheses [that people would make corrections in inaccurate
descriptions of their personalities.] That is, some people did and some
didn't, but the group's mean sum of correction scores - for tampered items-
was greater than the mean sum of correction scores for non-tampered items.
This is exactly how Frey and Stahlberg did it, and practically everyone
else who researches in this area. If the results come out the way you
hope, you publish, otherwise you throw them away and do something else.
We felt insecure about it. A leading statistician (sorry I forget his
name) warned psychologists back about 25 years ago that the habit of
publishing only positive results would gradually fill the various fields of
psychological resarch with false facts. (I can't see that he was ever
taken seriously.) So, we replicated our study.
The next time it came out exactly the opposite way. The third time we
complicated our analysis by an ad hoc separation of subjects into strong-
ego ones who did correct false attributions and weak-ego ones who did not.
This way of increasing fine slicing to get a significant finding, is also
pretty standard psychological research strategy. But, we realized that we
were "cooking the books."
I sat around discussing these misgivings with the students in my PCT
seminar. Gradually the talk drifted away on to more personal thoughts.
One of them said, "I'm a rather shy person." I got an inspiration [is that
a self-perception being controlled by a control system?] I said, "Oh, no,
you're not." Her eyes widened, she blushed (as I recall) and answered in
what sounded like an emotionally charged tone, "I am so." And went on to
offer support for that postulate.
That was the background of the first (so far as I know) use of "the test"
in studies of the "self." We got volunteers from another class, got them
to pick out self-descriptive adjectives and then immediately a stooge-
partner read the top item and said, "Oh, no you're not..." and wrote down
what the s said in reply. Almost every one reacted as did that very first
person. We did it several times after that and it always comes out that
way.
It wasn't until I got the first draft of Casting Nets, and saw how many
different logical possibilities there are for a reader to interpret to
himself the kind of task that is imposed by the Frey and Stahlberg (and our
first) method, that I came to understand why no one should expect anything
different from traditional psychological research methodology than "some do
and some don't" support the hypothesis. And if the group mean of do-ers is
greater than the group mean of the don't-ers at >.05, then you publish. If
not you try like hell to find some way of cooking the books so as not to
waste all the time you have put into it, much less your grant money, if
any.
The next thing I derived from Runkel's logical analysis was that the more
time that elapses between self-description and disturbance, or between
disturbance and reaction the greater the likelihood that the s will be
attending to_controlling_something entirely different when he gets the next
step of the task. We had kind of lucked out in setting up our application
of "the test" to work in the immediate present. Or had we?
If you understand the meaning of what the test is testing I don't see any
other way you could do it. I was uneasy about the deception aspect of that
method, though, and that led to the modification in which we asked people
to describe themselves and then immediately to say how they would reply to
someone who looked at their self-description and said, "Oh, you're not like
that." Well, that came out the same way too. There were a couple of
people who equivocated, but almost all acted as if the disturbance had
occurred in real time.
Now, let me take this back to Brian and George Bush.
You certainly could do our second method, that is just a routine old
questionnaire method. You are asking a person to imagine a disturbance to
a perception of his self-image that he is presumably controlling. If he
doesn't correct the disturbance AND you accept the PCT model, then he isn't
controlling that.
Brian continued,
I am now planning follow-up psychological interviews to begin testing the
model sketched in my article.
I hadn't yet seen Brian's "model sketched in my article" at that time so I
couldn't begin to guess how one would go about testing it with interviews,
but I can infer something from his statement,
My purpose was to use control theory to make sense of what is known about
the psychology of militarism, including my original survey data.
Whenever I see someone stating a purpose I resort to the general PCT model
and infer a RS. I interpret the RS to be: perceiving control theory making
sense of what is known about...etc. Now, if what Rick and Bill P had to
say about Brian's article was perceived as, "you aren't using control
theory to make sense out of your data," (or something to that effect), I
can perceive Brian's series of replies as efforts to correct his perception
and get it to match the RS. (I'd be surprised if the ES has in fact been
reduced yet, unless the imagination arrow has come into play, though I
think I can see some evidence of that, but that is another story.)
If there is anything to the above, then Brian could expect, on the basis of
general PCT theory, that George Bush, or anyone else he interviews, will be
controlling his perception of any attribution that he corrects when
disturbed IN ANY WAY and will not be controlling any attribution the
disturbance of which he does not correct.
As to Brian's statement,
Hawk and dove policy preferences are associated, for males, with "macho"
and "idealistic" self systems respectively
it seems clear, by the use of the word, `males,' that Brian is looking at
groups of people, not a group of individuals. Therefore, by definition we
are not dealing with control theory. I assume you do not postulate a group
of people as a living entitiy that can control its perceptions or anything
else. When fine slicing what goes with what - in groups of people - I go
back to Runkel for his convincing (to me) discussion of the value and the
how-to-do-it of survey research. I also subscribe to his explication of
why that methodology never applies to the method of specimens. Since PCT
is a method of specimens I take it that fine slicing the interrelation of
categories in various groups of people - while it can be valuable research
for appropriate purposes - can not test a model.
Brian, why can't you use both methods, if they suit your purposes, paying
attention not to use the wrong one for the wrong purpose and get
meaningless results, as David and I did in the first three attempts at our
study?
I think I can see now how one could carry the self-as-control-system idea
in several directions from where we began. If you have a person correcting
attributions - whether they be, "Yes, I'm a dove," or "I am a shy person,"
or ... wouldn't a next logical question be, what does this mean for you?
(esp.) What principles do you derive from being that? One procedure would
be to question the logic whereby a principle is implicit in a system-
concept, self, identity, self-image. I mean `question' in both senses: 1)
to dispute that it follows; 2) to request clarification as to how it
follows. From there you can go to what courses of action (sequences,
strategies) are required to enforce, enact, apply, impliment the principle.
That would be a way of testing the HPCT model as a way of understanding
behavior. Does it work backwards? Can PCT be used to test one's own model
of something? I don't know, but first I think it depends on what you mean.
Another way to go with it might be useful in the issues about self-control
that Bruce Abbot launched and various others have commented on. The
example of the person who wants to lose weight and wants to eat more
calories then he burns/day is a real one for me. It has come up a number
of times in my clinical practice. What I found convinced me that Rick's
analysis is right, when you say it is a matter of conflict of systems at
the same level. If there WERE a higher order system to arbitrate the
conflict would resolve. That is where the method of levels applies. If
the person comes to a vacuum, that realization might result in reorganizing
such that sometime afterward there is a higher level. Usually this is
reported AFTER the person mentions that the problem has resolved. But for
some reason I can't fathom, people with the particular issue of obesity
have so far in my experience been highly resistant to the method of levels.
A research on it could follow the self-as-control-system method by
proposing a series of tests, like, [Do you believe that your high blood
pressure results from your weighing over 300 pounds?] (This is real
material from my practice) Yes. [What is your attitude toward that fact?]
I'm scared. [What does your fright make you want to do?] Eat comfort
food. [What kind is that?] Greasy.
Years ago one of my students did an MA thesis that seemed to find a
permanent-like setting of an RS for intake of protein in some society in
Africa where protein was chronically scare. She found that those among
these people who had had very lo protein intake during the first year or so
of life thereafter reacted as if they had been poisoned if ever they got a
large amount of protein in a short time. I wonder if some of the studies
of fat cell determination early in life indicate a similar process. In
that case, the kind of reorganization that would be required would seem to
be at a very fundamental, physiological, level, at least for people who
acquired their obesity early on. The particular client described above did
not seem to me to have a very developed systems concept of herself as a
person. She also did recall that her father was fat through out her
childhood, and that she kept trying to win his favor even though he was
brutally harsh in punishing her and volatlile/unpredictable.
The way the term, `self-control' is usually used is that the person who has
it acts consistently with his stated purposes. This client vacillated.
Sometimes her stated purpose was to lose weight. Other times it was to
comfort herself from the slings and arrows of the world. She didn't have
self-control. But, then, she didn't have much of a self, either.
Best, Dick R.
Best, Dick R