From Tom Bourbon [950801.2246]
I don't have the original subject line from the post to which I am replying.
Rick is forwarding some of the csg-l mail to me. For some reason, my mail
goes out on the net, but nothing -- not even an ack -- comes back to me from
the listserver.
[From Bill Powers (950801.1710 MDT)]
Bruce Abbott (950801.1715 EST) --
This gets more fascinating by the minute.
The average response rate using the 60 gm lever is on average
about 72% of the 45-gm lever rate.Your new data show that what is constant about the behavior is not the
bar-pressing rate, but the work (force times distance) done by the rat
on the bar per unit time. 45/60 = 75%. This argues even more forcibly
that some sort of limit has been reached.
Neat!
This will probably seem off the track of the discussion, but something a
little bit like this possibility occurred to us when my students were doing
their theses on "operant performance." The students wanted to use a set-up
in which an animal could "earn" unlimited access to an always-open bin with
a water bottle and a big pile of food, but at various times the student
could set different requirements for the work the animal must do to get to
the bin -- something that allowed the animal to use continuous behavior,
rather than discrete bar presses. (These students were interested in
gathering data somewhat like those reported by Marwine, Collier and
associates, in the 70s.)
One student (Milton Crawford) used a large box as a home area. A 5-foot long
runway extended out of one side of the home box. The animal was free to
traverse the runway at any time. Whenever it did, it had free access to the
bin than contained the water bottle and the pile of food. The runway could
be either horizontal, or elevated, rising at an angle of 50 degrees. (These
two conditions might be _roughly_ analogous to the different forces required
for a rat to operate the bar in the data you are analyzing.)
When the runway was in the elevated position, compared to the horizontal,
animals would (1) traverse the runway fewer times each day; (2) spend more
time at the bin after traversing the runway; (3) consume more food and water
during each visit to the bin; (4) engage in fewer, but longer, bouts of
eating and drinking while at the bin. My student argued that the animals
were controlling for amount of food and water consumed (amounts were the
same for either positon of the runway), _and_ for effort required to reach
food and water. The apparatus was primitive and we didn't have enough solid
data, on enough animals, to warrant a try at publishing the results.
----------------------------------------------------------------------
Tom Bourbon [950801.1652] --Good stuff on Staddon's previous work. Actually I was corresponding with
him around then.
I recalled that you did. At the time, I thought Staddon had missed a great
opportunity to become a _really famous_ behaviorist. It looks like Bruce
gets to play that role.
I also saw that Bill Powers "unloaded" on Rick Marken, for what
Bill characterized as Rick's negative role in the discussions about
using PCT to model Staddon's work.I could have unloaded on you, too. Not that that makes me right. I just
feel queasy about hurling generalizations around about a whole class of
psychologists, when we're communicating with one guy who is a member of
that class and is obviously not like this evil (and imaginary) image of
the Bad Behaviorist. Take people one at a time, I say. Don't judge them
by their labels, I say, any more than you'd judge them by the color of
their skins.
I'm sorry you thought I did otherwise. That certanly wasn't my intent when
I was expressing what I clearly labeled as my personal opinion that direct
work on identifying controlled variables would be a good thing. As for my
critiques of some recent articles in which psychologists talk S-R talk, I'll
stand by everything I said. I believe we should keep in touch with the
literature, so we can be aware if things do change. (Alas, they have not.)
Nothing in those critiques was aimed at Bruce or at the present work he and
you are doing. I assumed Bruce would know that, just as I would know that a
critique of Lord, Carver, Scheier, or any other of the popular "control
theorists" does not apply to our work.
Good work, Bruce. Now, _run some real data on control by
experimental animals_!I think we're getting a handle on how to set up experimental conditions
that really show control. Obviously we have to make the food sizes
larger and reduce the effort needed to ingest the food each time it's
presented, so it doesn't conflict with lever-pressing. Maybe roll it
down a trough right under the rat's nose? If we could get collection
time down to less than a second, we'd at least be rid of that problem.
A real problem. That's one reason my students went the route of allowing
free access after the required performance. But that doesn't work if you
want to "make" rats press bar on a schefule -- not unless you use apparatus
like that of Marwine, Collier and others.
Later,
Tom