Herman Rubin wrote:
>
> In article <[EMAIL PROTECTED]>,
> Neil J. Salkind <[EMAIL PROTECTED]> wrote (inpart)
> >11/4/03
>
> >Dear teaching colleague,
> >I could sure use your help. I was just asked to speak next January at
> >the National Institute for the Teaching of Psychology and would like
> >to ask you for a few minutes of your time. My presentation at NITOP is
> >titled, "What's difficult to teach in introductory statistics and how
> >to do it".
and Herman Rubin wrote (in part)
>
> The worst of all consists of transformations to normality.
> This destroys all structure, and in the real world, nothing
> is normal.
I think we have to distinguish here between "transformations" (such as
taking logs, square roots, etc) which are bijections and do NOT destroy
any structure (provided that the use of the transformation is recorded
and it is used in interpreting the final results) , and the use of
normal scores (replacing each datum by the inverse cdf of its empirical
quantile) which does.
-Robert Dawson
.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
. http://jse.stat.ncsu.edu/ .
=================================================================
