Re: Regression Analysis

From: Richard Ulrich <Rich.Ulrich@xxxxxxxxxxx>
Date: Sat, 29 Oct 2005 15:24:21 -0400

On 28 Oct 2005 20:15:12 -0700, "George" <dialloalpha@xxxxxxxxx> wrote:

> Hi all,
> I have two questions related to regression analysis:
> 1) I woulk like to know how to run a regression analysis with dummy
> variables when the variance in the error terms is not constant.
>
> 2) How to run a regression analysis when the dependant variable are
> highly correlated with one of the independant variable. (the two are
> perfectly correlated)

If one independent variable is perfectly correlated with
the dependent, then you are finished. There is no residual
to support any further prediction.

Please browse in any basic statistics book and then in
a book on regression.

--
Rich Ulrich, wpilib@xxxxxxxx
http://www.pitt.edu/~wpilib/index.html
.

Follow-Ups:
- Re: Regression Analysis
  - From: Reef Fish
- Re: Regression Analysis
  - From: Art Kendall

References:
- Regression Analysis
  - From: George

Prev by Date: Re: Who is this "Reef Fish" character?
Next by Date: Difference between probability, expected value, and average value???
Previous by thread: Regression Analysis
Next by thread: Re: Regression Analysis
Index(es):
- Date
- Thread

Relevant Pages

Regression Analysis
... I woulk like to know how to run a regression analysis with dummy ... Prev by Date: ...
(sci.stat.edu)
Re: Regression Analysis
... Richard Ulrich wrote: ... >> 1) I woulk like to know how to run a regression analysis with dummy ... exactly on the straight line and there will be no residuals. ...
(sci.stat.edu)