[R] Preconditions for a variance analysis

[Christoph Buser]

Dear David

I'd not recommend the tests (for normality, equal variances) as
they are described on your second link (wikipedia).

I would use graphical tools such as Tukey-Anscombe Plot
(residuals against fitted values), quantile (or normal) plot,
leverage plot.

see also ?plot.lm for some short descriptions of these plots in
R.

Best regards,

Christoph

Daniel Tahin writes:
 > Thanx for your answer. I don't have the book, but found something on 
 > the web:
 > http://www.basic.northwestern.edu/statguidefiles/oneway_anova.html   
 > and
 > http://en.wikipedia.org/wiki/Analysis_of_variance#Assumptions
 > 
 > Seems to be the same on both of the sites :-)
 > Is this, that was meant?
 > 
 > Thanx again,
 > Daniel
 > 
 > 
 > 
 > 
 > > Dear David
 > > 
 > > Yes. There are assumptions that should be verified in an
 > > analysis of variance. Without checking them, the results are not
 > > reliable. 
 > > I'd recommend e.g.
 > > 
 > > Robert O. Kuehl, Design of Experiments: Statistical Principles
 > >   of Research Design and Analysis, Duxbury Press, 2000
 > > 
 > > You will find a chapter about assumptions and how to check them
 > > by residual analysis,
 > > 
 > > And also
 > > 
 > > W. N. Venables and B. D. Ripley, Modern Applied Statistics 
 > >   with S, Springer-Verlag, New York, 2002
 > > 
 > > in which you find residual analysis and how to obtain it in R.
 > > 
 > > Best regards,
 > > 
 > > Christoph
 > > 
 > > --------------------------------------------------------------
 > > 
 > > Credit and Surety PML study: visit our web page www.cs-pml.org
 > > 
 > > --------------------------------------------------------------
 > > Christoph Buser <buser at stat.math.ethz.ch>
 > > Seminar fuer Statistik, LEO C13
 > > ETH Zurich	8092 Zurich	 SWITZERLAND
 > > phone: x-41-44-632-4673		fax: 632-1228
 > > http://stat.ethz.ch/~buser/
 > > --------------------------------------------------------------
 > > 
 > > 
 > > Daniel Tahin writes:
 > >  > Hello everbody,
 > >  > 
 > >  > i'm currently using the anova()-test for a small data.frame of 40 
 > >  > rows and 2 columns. It works well, but is there any preconditions for 
 > >  > a valid variance analysis, that i should consider?
 > >  > 
 > >  > Thank you for your answer,
 > >  > Daniel
 > >  > 
 > >  > ______________________________________________
 > >  > R-help at stat.math.ethz.ch mailing list
 > >  > https://stat.ethz.ch/mailman/listinfo/r-help
 > >  > PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
 > >  > and provide commented, minimal, self-contained, reproducible code.
 >