Re: Teaching statistics

[Paul Bernhardt]

One of the things I particularly like about Tabachinick and Fidel's 
Using Multivariate Statistics is each statistical procedure has an 
introductory section that discusses assumptions for that procedure and 
the theoretical implications (and practical importance) of violations. 
Almost all discussion in these sections refer to the 4th chapter of 
their text which is completely on data screening.

I'm convinced it is a model that should be used for graduate level 
statistics courses in general.

Paul

On Monday, November 3, 2003, at 03:15  PM, [EMAIL PROTECTED] wrote:

I am a big fan of data screening, but find it
difficulty to fit into the course.  At least at a
level that seems to matter.  Currently I spend about
a week on outliers and a week on non-normal
distributions.  But that does not seem to create any
changes in practice.
I would be interested in finding out what successful
folks are doing differently. 
 
.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
.                  http://jse.stat.ncsu.edu/                    .
=================================================================