NetNews Usenet Archive 1992 #31/NN_1992_31.iso/spool/bit/listserv/csgl/2031

Text File | 1993-01-02 | 8.0 KB | 154 lines

Comments: Gated by [e-mail]
Path: sparky!uunet!gatech!paladin.american.edu!auvm!UIUC.EDU!G-CZIKO
Message-ID: <[e-mail]>
Newsgroups: bit.listserv.csg-l
Date:         Sat, 2 Jan 1993 20:44:03 -0600
Sender:       "Control Systems Group Network (CSGnet)" <[e-mail]>
Comments:     Warning -- original Sender: tag was [e-mail]
From:         "Gary A. Cziko" <[e-mail]>
Subject:      3 Pencils & 4 Rubber Bands
Lines: 142

[from Gary Cziko 930103.0240 GMT] (I got "93" on the first try!)

I did it!  I just designed the most awesome manual PCT demonstration of all
time.

And all you need to do it is 3 pencils, 4 rubber bands, and two pieces of
paper taped together (on the other side) to make a long sheet about 22
inches high and 8 1/2 inches wide (or just use two attached fan-fold
computer sheets)

Take the 3 pencils and attach them to each other like wrungs on a ladder
using the rubber bands.  Like this (with the vertical lines representing
the pencils and the horizontal ones the rubber bands:

         |------|------|   (not bad for my very first ASCII diagram)
         |      |      |
         |------|------|

(And yes, it IS possible to loop a rubber band around a pencil when the
other end is ALREADY looped around another pencil.  Greg Williams showed me
how to do this at the last CSG meeting in Durago--another reason not to
miss CSG meetings).

Now get your long piece of paper and draw a line horizontally across the
middle (just above or below the seam of the two sheets).  This is the
target line.  Place the paper on a table and tape down the corners so that
is won't slide about.

Take one end pencil and have your subject take the other end pencil.  Put
your pencil point above the target line and the extreme left side of the
paper and have your subject put her pencil point below the target line so
that (a) all pencils are perpendicular to the paper, (b) the middle pencil
point is on the target line and (c) the rubber band connecting the
subject's pencil to the middle ("cursor") pencil is perpendicular (90
degree angle) to the target line.  Tell your subject to maintain these two
perceptual variables (cursor pencil point on target line and rubber-band at
90-degree angle) as you SLOWLY trace out an approximation to a sine curve
above the cursor from one side of the paper to the other.  Make sure that
all three pencil points are making contact with the paper and leaving a
trace (those of you with lots of research money from NASA or NSF may want
to use felt-tip pens instead since they leave a nice trace with little
pressure; in that case the demonstration is called "3 felt-tip pens & two
rubber bands," but all else remains the same).

After you've done this once, DO IT AGAIN, this time making sure that you as
experimenter follow the same line as you did the first time.

You will now have before your very eyes a very remarkable piece of paper.
Above the target line you will see an approximate sine wave drawn twice
(they will look more like one line if you're a really good disturbor like
Rick Marken or Tom Bourbon).  These are records of the two DISTURBANCES.
Below the target line you will see two mirror images of the approximate
sine curve drawn twice.  These are records of the subject DID.  They will
probably be more irregular than the disturbances, but there should be an
obvious similarity between the two response curves.  In the middle you will
have two "cursor" lines, which are records of what the subject SAW during
the two trials (of course, there is no record of the rubber-band angle seen
by the subject, but remember that PCT is still in its infancy).  These two
lines should not have any discernible pattern to them.  In addition, they
will NOT BE SIMILAR TO EACH OTHER (if they are, this is an indication that
you disturbed too fast and the subject lost good control; Greg Williams may
like this, but I won't).

This is very strange indeed since the subject's responses are SIMILAR on
the two trials and yet what she saw (the cursor pencil point) during the
two trials was very DIFFERENT.  How can the subject respond similarly on
two trials when what was seen (the "stimulus") was so different?  If anyone
can come up with an explanation of this which does not look like
closed-loop negative feedback model, please let us here on the CSGnet know
about it.

Your subject may find it difficult at first to control both the position of
the cursor pencil on the target line and the angle of the rubber band.  So
you may want to let her practice first using the eraser ends of the
pencils.  Alternatively, you can practice yourself and let your participant
be the experimenter.

I hope that many of you have now realized that this is a manual (i.e.,
non-computer) approximation of the task and analysis used by Rick Marken
in:

   Marken, R. (1980).  The cause of control movements in a
     tracking task.  _Perceptual and Motor Skills, _51_,
     755-758 (also contained in _Mind Readings_ available
     from The Control Systems Groups, 460 Black Lick Road,
     Gravel Switch, KY 40328.  Price is $18 postpaid).

Marken [Hi, Rick; I know you are reading and loving this.  While I have
your attention and admiration, why don't you be nicer to Ed Ford?  I bet
you're just jealous because he's on PBS-TV and you're not!] showed that in
a similar task using a state-of-the art Apple II computer and game paddle
that the correlation between cursor variations (here, middle pencil
variations) were usually less than .20 while correlations between response
variations (here subject's pencil) were always greater than .99

Now comes the fun part.  If you are a psychology student, show the
demonstration to your local non-PCT psychology professor (if these are hard
to find, please let us know where you are located) and ask him or her to
explain the findings.  He or she will most certainly have to say that the
two sets of cursor variations are similar, even though there are not.  If
he or she doesn't believe they are not similar, show them Marken's paper
with the fancy computer and game paddles and correlations sometimes to FOUR
decimal points (I can never get more than three; that's what being a
computer programmer gets you).  He reports one correlation between cursor
variations of .0032 with a corresponding correlation between response
variations of .997.  (If your non-PCT psychology professor is really sharp
he or she will quickly point out that a correlation of .0032 can be
statistically significant with a large enough sample).  If he or she is not
that sharp (or much sharper), he or she should be quite shaken up.

Now if one psychology student, just one psychology student does this to his
professors, they may think he's really sick and won't give him a passing
grade.  But if two students, two students do it, in harmony, they may think
they're both faggots and they won't pass either of them.  And three
students, can you imagine, three students walking in their professors'
offices and doing the 3 pencils & 4 rubber bands demonstration and walking
out. They may think it's an organization.  And can you imagine fifty
students a day, I said fifty students a day walking in to do the 3 pencils
& 4 rubber bands gig and walking out! Friends, they may think it's a
movement.

And that's what it is.  It's the Three Pencils & Four Rubber Bands
Anti-Input-Output Psychology Movement, and all you got to do to join is to
do the demo the next time you run into a non-PCT psychologist!  With
feeling!

--Gary

P.S. With apologies to Arlo Guthrie.  None to anybody else.

P.P.S  This should also work on a chalkboard for classroom and other public
demonstrations.  In that case it should be called the "3 pieces of chalk
and 4 rubber bands demonstration."
------------------------------------------------------------------
Gary A. Cziko                    Telephone: [phone]
Educational Psychology           FAX: [phone]
University of Illinois           E-mail: [e-mail]
1310 S. Sixth Street             Radio: N9MJZ
210 Education Building
Champaign, Illinois 61820-6990
USA
------------------------------------------------------------------