"albinali" <[EMAIL PROTECTED]> escribi� en el mensaje
news:[EMAIL PROTECTED]
> Hi,
>   With non-ordinal categorical data, I was told that logistic regression
is
> likely to do a better job, why is that, whats the problem with linear
> regression?

In the case of what the dependient variable be dichotomous, if you do a
multiple regression you are violating the assumptions required for
inference. In particular the regression errors are not normally distribuited
neither their variance is constant.
If you do a multiple regression the final conclusions can be misleading.


Best wishes.
Carlos


------------PRESTIGE OIL SPIL NOT AGAIN---


.
.
=================================================================
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