On Tue, 4 Feb 2003 12:27:13 +0000 (UTC)
<[EMAIL PROTECTED]> wrote:

> Simon, Steve, PhD <[EMAIL PROTECTED]> wrote:
> 
> : I would not be as critical as some of the others on the list. Sometimes a
> : categorical variable is easier to interpret. A lot of dietary research, for
> : example, looks at the highest quintile of fat consumption and compares it to
> : the lowest quintile. I can visualize those two groups pretty well.

This is not very informative research.  Categorical variables are only easier to 
interpret because the results are misinterpreted.  Are we in the highest quintile of 
fat consumption or is our fat consumption equal to a specific value?  I think the 
latter.  Hence the widespread use of smooth nonlinear regression fits, to allow 
variables to remain continuous.

> 
> But this is different from dichotomization.  Sam Green published a paper a
> few years ago (I believe it was in Multivariate Behavioral Research)  
> showing that if you use enough categories, say, six or seven, the loss of
> information is not too bad. 

It depends on the true shape of the regression function, but having to use 6 d.f. 
results in high mean variance of estimates as compared with, say, a regression spline 
fit.

> 
> : Furthermore, categorization mitigates some of the problems caused by
> : measurement error.

Not in any way.  When dichotomizing a variable, if a subject is assigned to the wrong 
category, the error is 100%.  The probability of an error times the magnitude of an 
error is not at all helped by categorization.

> 
> Can you say more about this?  My understanding was that by definition, 
> dichotomizing will reduce the reliability of a continuous measure.
> 
> Mike Babyak
> Duke University Medical Center


-- 
Frank E Harrell Jr              Prof. of Biostatistics & Statistics
Div. of Biostatistics & Epidem. Dept. of Health Evaluation Sciences
U. Virginia School of Medicine  http://hesweb1.med.virginia.edu/biostat
.
.
=================================================================
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