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] Lecture Isn't Effective: More Evidence #2



Looking at Nobel Laureates is fruitless because the statistics are so low,
and you have not way of knowing what actually went into their education.
They can certainly tell what they think went into their education, but that
is like asking a patient how they got a disease. The actually circumstances
may be far from what is reported. And what will be reported by the
Laureates will be what their paradigm of education prompts them to report.
In other words you remember what you believe and not what actually happened.
This is a well known psychological trait.

You can get some clues from looking at Nobel Laureates, but then you have to
do studies on other students to confirm these results. This is similar to
medicine where anecdotal evidence can be used to give you hints as to what
may be going on, but you need strict studies to figure out what actually
happened.

In education you try to come up with things that you can vary and then do
testing to measure. But unfortunately there are things you can not vary and
at a given time you may not have good measurement tools. We do have tools
besides the FCI and FMCE such as the Lawson test, MBT(Hestenes et al), other
concept inventories, VASS (Hestenes et al), MPEX (Redish), TUG-K (Beichner),
McDermott tests ... These other evaluations show similar results in that IE
classes do better when research based material is used. Studio style
classes with research based materials show improved attitudes toward
learning.

Many of the facets of what is done in PER actually go back to other articles
in JRST such as the use of multiple representations, so PER has drawn on
other sources. In particular the TUG-K confirms the experiments by Heather
Brasell in the 80s as I recall, where specific labs involving the use of the
sonic ranger improved kinematic graph understanding. The depth of research
is quite large and it all confirms what we are doing.

There is some research that shows that targeted lectures such as bridging
after the ILDS help students. Or a lecture on how to design experiments is
helpful after appropriate PER style labs brings up scores. Similarly Laws,
Sokoloff, Thornton have designed homeworks that come after Real Time
Physics. These homeworks do improve gain in understanding. So conventional
things are used, but in a particular coherent sequence.

So there is a lot of information beyond the FCI. The Minds on Physics
program at U.Mass Amherst was shown by an independent evaluator to promote
"expert like" problem solving. The Hellers at Minn use a more conventional
approach but with group problem solving and have shown that engineers are
then capable of solving harder problems than usual. As a side effect the
FCI gain was much higher. Mazur showed that his students did better on a
conventional problem solving final exam after he switched to interactive
lectures. But his program was guided by trying to achieve high FCI gain.
So high FCI gain would seem to be a necessary condition for good
understanding and good problem solving. However it is probably not
completely sufficient as there are things that are necessary beyond the FCI.

Read the rest of the literature. There are now many evaluations that go
beyond the FCI. The FCI is the first easiest to evaluate target. Once
someone achieves good gain there, they can go on to other issues. So then
there are other targets that need to be included of which my favorite is
gain on the Lawson test. And one can get decent gain on both. We are
looking at all levels, but nobody can predict Nobel Laureates. That level
of performance usually requires extremely high levels of perserverence and
often a bit of luck. Being single minded and maybe even a bit OCD
(obssessive compulsive disorder) is needed.

How about looking at "The implications of a robust curriculum in
introductory mechanics", Hoellwarth, Moelter, AJP, 79 (5) May 2011. Keep in
mind that they are doing just the first level of PER, and looking at other
effects has not been done in this study.

And why not do some of your own experiments in PER???? Try to show that PER
results are not improving student learning! Or try to verify PER restuls!

John M. Clement
Houston, TX


I am troubled by another problem that I see with PER. There
seems to be no consensus as to the type of student we are
addressing. or
the ultimate goals of the teaching. It is as though the fci
score is the
be all and end all of the physics teaching process. The
academies that
graduated some of our outstanding physicists become irrelevant. Why
aren't we looking at Nobel Laureates and how they were educated?