Thom Baguley wrote:
>
> Gary McClelland wrote:
>
> Lots of excellent advice. I would add that plots also have the advantage of
> suggesting a possible solution to any serious violation. For example variance
> about the regression line increasing with X tends would lead me first to look
> at transformations such as the log or square root. Failure on a homogeneity
> test is typically lacking in diagnostic information.
>
> Thom
This is not quite true.
If the variance of
e_i=(Y_i-hatY_i)
is increasing with the hatY (I use LOWESS for to estimate hatY)
then you can sometimes find a transformation Y'=f(Y) and you get
homoskedastic
residual variance.
If the variance of e_i is increasing (decreasin) with X, the
situation is
more complicated. Sometimes you can find variance stabilizing
transformation(s)
for to make the residual variance homoskedastic. Y'=f(Y), X'=g(X).
Please read R. Dennis Cook & Sanford Weisberg, Applied Regression
Including Computing and
Graphics. Wiley 1999.
http://www.stat.umn.edu/arc/
Juha
--
Juha Puranen
Department of Statistics
P.O.Box 54 (Unioninkatu 37), 00014 University of Helsinki, Finland
http://noppa5.pc.helsinki.fi
===========================================================================
This list is open to everyone. Occasionally, less thoughtful
people send inappropriate messages. Please DO NOT COMPLAIN TO
THE POSTMASTER about these messages because the postmaster has no
way of controlling them, and excessive complaints will result in
termination of the list.
For information about this list, including information about the
problem of inappropriate messages and information about how to
unsubscribe, please see the web page at
http://jse.stat.ncsu.edu/
===========================================================================
