Chronology Current Month Current Thread Current Date
[Year List] [Month List (current year)] [Date Index] [Thread Index] [Thread Prev] [Thread Next] [Date Prev] [Date Next]

Re: [Phys-l] detailed + narrow - sophisticated



If your look at the various methods that are used in PER inspired courses
you do not find that they are either just telling students the answers, or
are teaching a detailed view of a narrow concept.

The questions were crafted by looking at free responses and the alternatives
were picked according to the top misconceptions. Then the FCI correlates
very strongly with the FMCE which asks very different questions, except for
identical NTN3 questions.

Now we know from cognitive psychology that what you perceive is conditioned
by your mental model. This was discovered by PER researchers, but cognitive
psychologists have known this for a long time. So the standard lecture,
recitation, labs simply do not change mental models very well. What
students remember from labs and lectures is what they already knew before.
The inappropriate mental model changes memories so that they will tell you
they remember things that never happened. They will see one result of a
demo, and then a few weeks later tell you the complete opposite was seen.
Watch the Mazur videos for his accounts of the various things he has tested
and observed. How about some comments on his research which he has detailed
in his videos, and a number of papers.

A test bank may be useful for a particular course, but it is useless in
comparing across courses and institutions. Standard conceptual test of
which we not have many more are useful for that purpose. And similar
results are found in other courses to what we have found in mechanics.

McDermott never used the FCI, but she came to many of the same conclusions
using her tests which are very different, and test different facets of
understanding. The rivalry between McDermott and Hestenes is well known,
but they came to very similar conclusions.

If you think MC tests are not good enough, turn the FCI into free response
questions and grade the free responses. Then give the FCI at least 2 weeks
later and see if they correlate. Another way of seeing if the FCI is valid
is to give it, and then later interview students to see what they actually
know. So rather than say it is not good, why not do some research to see
how good or bad it is, and then publish that research? Opinions are
interesting, but only research can validate them. You can do the same thing
with the FMCE.

We now know that there is a fairly high correlation between the Lawson test
and the FCI, but not as high as between the FMCE and the FCI. Again this is
from research.

As to narrow concepts, is being able to figure out the sign of acceleration
from knowing the directions of the velocity and whether you are slowing down
or speeding up a narrow concept? Is knowing that acceleration does not go
to zero at the top of the trajectory of a ball thrown straight up a narrow
concept? Is knowing that acceleration does not change direction (or sign)
when the velocity switches direction a narrow concept? Yet the sign and
direction misconceptions persist even after a student has taken calculus.
These are explicitly tested by the FMCE, but not by the FCI. I would think
these are fairly basic.

John M. Clement
Houston, TX



The main thing about the FCI is that it covers a VERY NARROW bit of the
physics curriculum--actually requires fairly sophisticated understanding
of
some concepts--and can be 'taught to'. It is NOT the benchmark for much
of
anything other than this narrow, sophisticated understanding of Newton's
Laws. Using it to 'prove' one teaching style better than another is,
IMO,
absolutely bogus!

I agree with that in spirit, and agree with N-1 of the words.

I might have used the word "detailed" rather than "sophisticated".

We agree that when students are taught a detailed view of some
narrow concept, that's a recipe for trouble.

I would like to think that a truly _sophisticated_ understanding
of the subject would not be narrow. Narrowness with a pretense
of sophistication is almost the definition of sophomoric.

The idea of a short, multiple-guess test being a measure of
"sophisticated understanding" strikes me as ludicrous. Also,
because of the tendency toward teaching to the test, the idea
of a short, invariant test being a reliable measure of /anything/
beyond rote strikes me as ludicrous.

By way of contrast, if we had a test bank consisting of
thousands of questions, and we constructed each instance
of the test by selecting a sample of the questions, then
rote wouldn't pay, and I might believe some understanding
was involved. But even then, getting from multiple-guess
to "sophisticated" would require quite a leap.