Re: [Phys-l] Premed Requirements Commentary

["John M Clement" <clement@hal-pc.org>]

Well, according to various researchers you need a certain minimum 
score on these tests to succeed in advanced courses.  Normalized gain 
admittedly does not give that score, but it does provide a relative 
measure of how well you are doing in instruction.  Hake showed that 
there is a weak correlation between the pretest and the final score, 
but the normalized gain correlated strongly with the type of course. 
Now as to where the researchers have gotten their information, I have 
not seen studies that validate their conclusion about the scores.

I am not the only one who thinks this way, I am just willing to be the 
most annoying and insistent.

The average physics course for premed and science majors should be 
able to get at least 60% normalized gain because most of these 
students have high scores on a Piagetian test, which is a good 
predictor of the maximum gain, and also a good indicator of thinking 
skills.

Unfortunately there is no way to go back and retrospectively test the 
hypothesis that low scores mean poor ability in engineering or 
science.  But it should be possible to test it in the future.  At 
present the idea that high scores may make a difference in employment, 
or even a difference in future courses is a belief or hypothesis, but 
then so is the opposite idea.  There have been a few studies of 
subsequent courses, but I don't remember that there were any 
astounding results.  This type of thing is difficult to study, while 
just measuring what happens in your own class is possible.  If we had 
better tests that could be used for comparison, we would use them.  At 
present there is no easily scored test for expert problem solving, but 
it is being worked on.

Incidentally Beth Thacker did find that her elementary ed teachers who 
took her high gain reformed course did better on standard problems 
than the conventionally taught engineers.  Only on the more complex 
problem did they not do as well.  So there is more than FCI scores, 
but that is the only thing we can easily cross compare.

John M. Clement
Houston, TX

John M. Clement
Houston, TX

> JC constantly harps on 'gain'--normalized gain on the FCI is what I 
> assume 
> he's referring to--AS IF this is a proven, meaningful measure of 
> something. 
> Of what?  Does it predict which students will end up being good 
> scientists, 
> good engineers, good doctors, good Indian Chiefs?  Without such a 
> correlation, then the results of one (now ancient) test, a test that 
> CAN be 
> taught to, a test that is limited to a narrow conceptual and topical 
> area, a 
> test that (admittedly) requires a lot of student/instructor work to 
> do well 
> on, is only a tiny piece of data.  As many of us have 
> commented--look to 
> your graduates and their successes or failures to evaluate the 
> educational 
> program.  If you can show that there is a general failure to 
> understand 
> Newton's Laws that is causing bridges to fall, causing people to be 
> misdiagnosed, that is holding back the frontiers of science 
> research, etc., 
> then we can all be much more worried about 'gain' and lack thereof.