[R] Non-normal residuals.

Bert Gunter gunter.berton at gene.com
Wed Oct 28 22:48:34 CET 2009


If generalities -- with the attendant risk of occasional specific caveats
and violations -- can be tolerated, then George Box's (paraphrased) comments
of circa 40-50 years ago seem apropos: why do statisticians obsess over
normality, to which most analyses -- i.e. inference (especially from
balanced designs)-- are robust, when lack of independence of the
observations is the violation of assumptions that can reek the greatest
havoc on the statistical analysis?

Time series analysis and mixed effects models are among modern statistics
ways of dealing with such lack of indepndence, btw.


Bert Gunter
Genentech Nonclinical Biostatistics

-----Original Message-----
From: r-help-bounces at r-project.org [mailto:r-help-bounces at r-project.org] On
Behalf Of David Scott
Sent: Wednesday, October 28, 2009 2:09 PM
To: Kjetil Halvorsen
Cc: Karl Ove Hufthammer; r-help at stat.math.ethz.ch
Subject: Re: [R] Non-normal residuals.

Kjetil Halvorsen wrote:
> On Wed, Oct 28, 2009 at 7:25 AM, David Scott <d.scott at auckland.ac.nz>
>> Karl Ove Hufthammer wrote:
>>> On Tue, 27 Oct 2009 18:06:02 -0400 Ben Bolker <bolker at ufl.edu> wrote:
>>>>  If transforming your data brings you closer to satisfying
>>>> the assumptions of your analytic methods and having a sensible
>>>> analysis, then that's good.  If it makes things worse, that's bad.
>>>> Other choices, depending on the situation, include robust methods
>>>> (for "outlier" problems); generalized linear models etc. (for
>>>> discrete data from standard distributions); models using t- instead
>>>> of normally distributed residuals;
>>> I have sometimes wondered about this: Which functions/packages do you
>>> to fit a (perhaps just a simple linear) model with t-distributed
>>> (or residuals of a different distribution)?
>> Package sn has this facility I believe.
> Yes, for independent data, but for time series???
> Kjetil

No, not for time series---I was responding to
"fit a (perhaps just a simple linear) model with t-distributed residuals"


David Scott	Department of Statistics
		The University of Auckland, PB 92019
		Auckland 1142,    NEW ZEALAND
Phone: +64 9 923 5055, or +64 9 373 7599 ext 85055
Email:	d.scott at auckland.ac.nz,  Fax: +64 9 373 7018

Director of Consulting, Department of Statistics

R-help at r-project.org mailing list
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.

More information about the R-help mailing list