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] Physics First Revisited



My so called invective was aimed at the idea that teaching is only an art.
While it is true that it is usually practiced that way, the science is there
and has been reported in the journals. If you read the journals you will
discover that. There is a depth of evidence from a whole variety of
researchers. This goes far beyond one test, namely the FCI. There are now
many tests the MPEX, FMCE, VOSS, TUG-K... The tests by Heather Brasell have
been leveraged into a test of understanding of kinematic graphs by Bob
Beichner. The point is that all of the evidence seems to point in the same
direction. Just as evidence for evolution now has depth, so the evidence in
favor of IE is quite large. There is now evidence in math that also points
the same way using a recently developed calculus test.

If you believe the FCI gain can be replicated easily try giving it and
calculating gain. Then write a paper showing your results.

Hake has many references, and anyone can easily go to the journals to find
more evidence.

As to long term gain there was one test of long term gain that was reported
in The Physics Teacher a number of years ago. It showed that FCI results
did not go down even 3 years after finishing the course. I will leave it
to the reader to look up the reference. It might be good practice in using
the education literature.

As to self reporting, that is how research is usually done. Physics
experiments are reported by the researchers, and the paper is merely read by
a referee who does not examine the actual data books. So honesty is assured
by consistency with later experiments. And indeed there has been a large
amount of consistency in results of PER experiments. But unlike physical
science, education science has so many variables that it is impossible to
get 100% consistency.

There are a large number of experiments where the researcher did not deliver
the treatment, but rather taught the teachers to deliver it, and then the
researcher gathered the evidence and analyzed it. And of course it is
completely impossible to do blind educational experiments because both the
teachers and students can easily see if the experience is different from
usual. The closest to a blind experiment was done by Merlyn Mehl in S.
Africa where the experimental group was 100% passing on the final exam,
while the control group had the usual 50% failure rate on the same exam.
The 2 groups spoke different languages, and prior to the experiment had
identical 50% passing rates. Mehl used the theory base developed by the
psychologist Reuven Feuerstein. Notice this was not the FCI, but a
conventional physics test.

The FCI may or may not be a good test, but it tracks perfectly with the
FMCE. Also the FMCE scores are identical after a 3 week Christmas vacation
when given as a surprise evaluation on the first day students returned to
class. So I have first hand evidence that these scores hold up for at least
3 weeks, which is more than happens for a conventional test.

So if anyone thinks there is no credible evidence for education as a science
let them read the literature, and come up with some good reasons why. I
would suggest any articles by Anton Lawson which show the efficacy of the
learning cycle. They do not involve the FCI. He publishes at least 2
papers a year, and has been at it since the 70s.

John M. Clement
Houston, TX


Simply reciting the mantra that "there is creible evidence" neither
creates the evidence, nor, to the extent that there is any, make it
creible.

So-called FCI gains are reported by those who give and correct the tests,
the circumstances under which the tests are given are unreported, the test
review procedures (there shouldn't be any) are unreported, the error
rates in correcting the tests are unknown, and, persistence, over time, of
the purported gains, if any, are unknown.