I don't feel M&M's are any more trivial to introduce the concepts than
the usual "chips in an urn" or "balls from a jar".  And I don't agree
that using M&Ms is trivial or juvenile.  My class has basic science
grad students, M.D.s, Health professionals of all types and they do
seem to get something from this.

I use M&Ms to discuss the binomial.  You can also mention that there
are "competing risks" but at the moment we are only interested in
"being blue".  We know what the manufacturer claims about the
population of M&Ms, but "I felt shorted the other day because I found
a bag with no blues" and I want to take the manufacturer to task.  On
any given bag, you can reject or fail to reject the claim.  You can
introduce a bunch of different topics, estimation, hypothesis testing,
power, sample size requirements, sampling distributions.  It's simple,
universal and gets their attention.

You can't do these same exercises with "real life examples" and make
it as concrete.  Once they get the basics, you should try to find
meaningful examples, I totally agree with that.

I use Daniel's book and appreciate the examples from the literature,
but you've got to admit that most of these are a little forced for the
introductory stuff.  One of the problems is the lack of common
knowledge...just yesterday, we were working with one that measured
glucose in Syrian hamsters and no one in the class could tell me
whether the difference in means was "clinically meaningful" since they
had no experience with Syrian hamsters.  I suppose that I should be
more careful in assigning these real life examples from the
literature.

Another tool that I like is a deck of cards, all red with the
exception of the Ace of Spades.  It's a good tool for introducing
hypothesis testing and the dangers of accepting thy null hypothesis, I
feel.


=================================================================
Instructions for joining and leaving this list and remarks about
the problem of INAPPROPRIATE MESSAGES are available at
                  http://jse.stat.ncsu.edu/
=================================================================

Reply via email to