: Doing versus Seeing

[From Richard Thurman (961008.0720)]

(Bruce Gregory 961007.1100)

It occurred to me while thinking about how we teach problem
solving in physics that we almost always tell students what to
_do_, rather than what to _perceive_. I suspect this is the
reason that problem-solving is so difficult and why we are so
ineffective teaching students how to do it.

I really like this idea. Wouldn't it be interesting to create a
new (well not so new) theory of instruction based upon the
principles of perceptual control. Actually it would not be an
instructional theory at all -- instruction seems to imply that
we just pour information into students heads. Perhaps there is a
way to describe a new learning activity that one could call a
'learning environment' where students are not 'taught' per se but
instead are supplied with the correct tools to learn. It seems
to me that the most valuable tool would be the ones which help
students determine what to _perceive_ in a given problem situation.

(Bill Benzon 961007)

I'm not anywhere near physics so I may be way off base. But how
easy is it to tell students what to perceive?

It seems to me that this could be fairly easy. Perhaps it would
be nothing more than a teacher telling the class, "Notice that
when I try to solve this problem I first look for all the things
that have the same unit of measurement. Then I look for all the
things that use a different unit of measurement and try to see
if I can get them into a common ..."

At first it may seem that the teacher is trying to get them to 'do'
something. In fact it seems we often try to get people to 'do'
things. But upon closer inspection I bet we would see that almost
all of our commands for action really break down into requests for
people to perceive certain events taking place.

Rich

···

------------------------------------------------------------------
Richard Thurman
Armstrong Lab
Aircrew Training Research Division
6001 S. Power Rd. Bldg 558
Mesa, AZ 85206
thurman@alhra.af.mil

[From Richard Thurman (961008.0750)]

(Avery Andrews 961908 Eastern Oz Time)

I think the real problem with
problems is that to solve them you need to build up an army of little
`demons' that work away in parallel, and suggest possible paths to
an answer, which you can try out.

The idea of 'demons' for some reason reminds me of a study carried
out in the mid 80's on the "Space Fortress Game." This video game
was used by several research institutions to examine, "the extent to
which it is possible to improve on unsupervised practice when
subjects must master a rather complex task. " It was a very intense
video game (for the time) and really taxed out the subjects. It had
all sorts of things they had to keep track of, all sorts of timing
problems. Many subjects initially complained that it was impossible.

A group of researchers modified the game so that they could manipulate
the emphasis that each game sub component received. For example, if
they wanted to emphasize the aiming aspect of the game, they presented
an aiming subscore on the screen. Later, when aiming was down pat, they
took that score away and started to emphasize another part of the
game (like flying the space ship). Several different aspects of the
game were manipulated in this way. All this was done while the subject
performed the whole task, its just that different aspects of the game
were given emphasis (through subscores) at different times.

Well needless to say, subjects who used this version of the game
outperformed subjects who were given unstructured practice on the
standard game. In fact their performance curves continued to go up
even after the 'treatment' stopped. They continued to 'learn' at a
faster rate than the 'control group' long after both groups were
given the standard game to play.

Their explanation, "...voluntary control of attentional resources
plays an important role in the development of strategies as well
as the ability to carry out the performance policies which are
dictated by them... The act of shifting strategies is hence also an
act of changing the focus of attention."

It seems to me that teaching problem solving skills could be
facilitated by decomposing the skill into its principle perceptual
elements and creating learning environments which emphasize those
elements and give students 'feedback' as to how well they doing.

By "decomposing the skill into its principle perceptual elements" I
mean nothing more than performing the test for the controlled
variable until one is relatively confident that each controlled
variable has been found. Then one can then start to specify what
elements (perceptions) of the problem need to be emphasized.

Eventually, all perceptions would come to be controlled in parallel.
The 'demons' all get exercised.

Reference:
Gopher, D., Weil, M., & Seigel, D. (1989). Practice under changing
priorities: An apporach to training of complex skills. Acta
Psychologica, 71, 147-177.

···

------------------------------------------------------------------
Richard Thurman
Armstrong Lab
Aircrew Training Research Division
6001 S. Power Rd. Bldg 558
Mesa, AZ 85206
thurman@alhra.af.mil

[From Bruce Gregory (961009.1200 EDT)]

Richard Thurman (961008.0720)

Wouldn't it be interesting to create a
new (well not so new) theory of instruction based upon the
principles of perceptual control. Actually it would not be an
instructional theory at all -- instruction seems to imply that
we just pour information into students heads. Perhaps there is a
way to describe a new learning activity that one could call a
'learning environment' where students are not 'taught' per se but
instead are supplied with the correct tools to learn. It seems
to me that the most valuable tool would be the ones which help
students determine what to _perceive_ in a given problem situation.

I'm trying to do this right now with a group of Master's
students at the School of Education. Most of them have
undergraduate degrees in biology. I am trying to produce an
environment in which they can come to understand floating and
sinking by carrying out measurements of their own design. One
substantial problem is that they rarely have any sort of model
which they are trying to test. In a sense, they do not perceive
the experimental arrangements they set up as exemplifying
fundamental principles whose validity they can test. (Telling
them that they _should_ do this is about as effective as telling
the dogs they should eat more slowly!)

Perhaps it would
be nothing more than a teacher telling the class, "Notice that
when I try to solve this problem I first look for all the things
that have the same unit of measurement. Then I look for all the
things that use a different unit of measurement and try to see
if I can get them into a common ..."

At first it may seem that the teacher is trying to get them to 'do'
something. In fact it seems we often try to get people to 'do'
things. But upon closer inspection I bet we would see that almost
all of our commands for action really break down into requests for
people to perceive certain events taking place.

I arrived at the same conclusion. One problem is that the
processes that make sense to me, may or may not make sense to
them. For example, I can never remember formulas, so I carry out
a dimensional analysis to convince myself that I must multiply a
force by a distance to get an energy... If the students have any
breakthroughs, I will certainly report them.

Bruce