Agreed.  I always provide a context, typcially a published research article
from which I deduced the model for simulating the data.  For example:
http://core.ecu.edu/psyc/wuenschk/SimData/XD-T.htm .

Karl W.
----- Original Message ----- 
From: <[EMAIL PROTECTED]>
To: "Wuensch, Karl L" <[EMAIL PROTECTED]>; <[EMAIL PROTECTED]>
Sent: Tuesday, November 04, 2003 5:31 PM
Subject: Re: [edstat] Data Screening (was Teaching Statistics)



And it is a bad idea to use concocted out-of-context data.
Without a context there is no way to decide what to do with
"kinky" points.

Kjetil Halvorsen

On 4 Nov 2003 at 10:25, Wuensch, Karl L wrote:
>
>
> I am guilty of spoiling my students by usually providing them with
> clean data, generated such that distributional assumptions are all met,
> there are no out-of-range data, and so on.  Every once and a while,
however,
> I do something that I really should do more often -- I throw into the data
> for their assignments a few kinks and expect them to find them.  If they
do
> not screen the data, they don't find them.  Of course, the more
challenging
> question is what to do with the kinky data -- excluding them from the
> analysis is often a bad idea.
>
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> Karl L. Wuensch, Department of Psychology,
> East Carolina University, Greenville NC  27858-4353
> Voice:  252-328-4102     Fax:  252-328-6283
> [EMAIL PROTECTED]
> http://core.ecu.edu/psyc/wuenschk/klw.htm
> .
> .
> =================================================================
> Instructions for joining and leaving this list, remarks about the
> problem of INAPPROPRIATE MESSAGES, and archives are available at:
> .                  http://jse.stat.ncsu.edu/                    .
> =================================================================


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

Reply via email to