Re: [Phys-L] science education articles

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

The flipped classroom has an obvious problem.  The students have to view the
lectures before they come into class, and they have been accustomed to being
spoon fed the lectures.  Then the teacher has to get the students to do work
in class, but here the teacher has to be the guide at the side, something
they are unaccustomed to doing, and usually do not know how to do it well.

Why would the students bother to view the lectures before the class??  If
they won't read the book beforehand, why would they view a lecture
beforehand.  In addition they might look at a different source than the one
assigned.  Mazur solves this problem by giving a reading quiz before each
class, and he has really competitive students.

As to PER what exactly would you quote as being contradictory???  I have
found it to be very consistent.  I really do not expect an answer because
those who diss PER have never answered any of my questions.  No answer, in
my mind, is essentially an affirmation that they do not know what they are
talking about.

Now as to the problem of education research.  It is much harder than medical
research because you are dealing with students who may fight what you are
doing and deliberately work to maximize their gains, not what you want them
to do.  In addition it is impossible to do a double blind study.  Even a
single blind study is nearly impossible because both the recipient of a
treatment and the deliverer of it know it is different.  The closest to a
single blind study was done by Merlyn Mehl who delivered different
treatments to classes that had different languages.  That way there was
minimal opportunity for students to compare what they were getting.

And then there is the study by Feuerstein where he delivered his treatment,
and half the students got a teacher designed enrichment.  His treatment
raised IQ substantially to average normal, while the other students remained
much lower.

The big problem often comes in because results are often not compared by
either an effect size or normalized gain.  Yes, there was improvement, but
is it big enough to be worth shooting for?

John M. Clement
Houston, TX


> This is why forums such as this one are invaluable.
>
> People say things here that they can't or won't publish in 
> the "literature".  Here's an example, or maybe two examples in one:
>
> On 09/23/2015 11:40 AM, rjensen@ualberta.ca wrote:
>
> > people don't tend to publish negative results. Flipped 
> instruction is 
> > a current example. Going to conferences, I find many people 
> have tried 
> > some form of flipped instruction. For most, it fails and/or the 
> > students revolt. And the event is buried, not published.
>
> Note that such publication bias makes it impossible to do a 
> meaningful meta-analysis.
>
> BTW ... I never said the literature was /uniformly/ bad.  I 
> said the signal-to-noise ration was poor.
> There are occasional bright spots.  For example, I disagree 
> with several parts of the following paper, but at least it 
> admits that there are validity problems and biases to worry about:
>   http://mazur.harvard.edu/sentFiles/Mazur_424102.pdf
>
>
> ============================
>
> On 09/23/2015 09:26 AM, Joseph Bellina wrote:
> > > You have often disparaged PER, but let me suggest that your
> > > description below is exactly what PER is about.