Hi R-help members, I have read a lot in the Archive about the "Type I" vs "Type III" sum of square. I think I have read confusing post so I want to have a clear idea of the problem. Here is an example. I have 3 groups of subjects of unequal sample size (G1 (n=7), G2 (n=7), G3 (n=4)). for Each subject I have 4 measures corresponding to the crossing of 2 factor (A & B) of two levels each. my dependant variable is X. After reading a lot of tutorials on R I have tried the summary(aov(X~GROUP*A*B+Error(SUJECT/(A*B) ) This results are with "type I SS". What's wrong with these results ? Should I use type III SS and, if so how to enter my design in Anova (car package, I still have not the J. Fox's book) ? I have clearly not understood the difference between type I & III (with the limits of each approach). A link to a good tutorial on this topic will help me a lot. Sylvain CLEMENT "Neuropsychology & Auditory Cognition Team" Lille, FRANCE
More to the point, you are confusing multistratum AOV with single-stratuam AOV. For a good tutorial, see MASS4 (bibliographic information in the R FAQ). For unbalanced data we suggest you use lme() instead. On Tue, 14 Feb 2006, WPhantom wrote:> Hi R-help members, > > I have read a lot in the Archive about the "Type I" vs "Type III" sum > of square. I think I have read confusing post so > I want to have a clear idea of the problem. > > Here is an example. > I have 3 groups of subjects of unequal sample size (G1 (n=7), G2 > (n=7), G3 (n=4)). > for Each subject I have 4 measures corresponding to the crossing of > 2 factor (A & B) of two levels each. > > my dependant variable is X. > > After reading a lot of tutorials on R I have tried the > summary(aov(X~GROUP*A*B+Error(SUJECT/(A*B) ) > > This results are with "type I SS". > > What's wrong with these results ? Should I use type III SS and, if so > how to enter my design in Anova (car package, I still have not the > J. Fox's book) ? > I have clearly not understood the difference between type I & III > (with the limits of each approach). A link to a good tutorial on > this topic will help me a lot. > > > > Sylvain CLEMENT > "Neuropsychology & Auditory Cognition Team" > Lille, FRANCE-- Brian D. Ripley, ripley at stats.ox.ac.uk Professor of Applied Statistics, http://www.stats.ox.ac.uk/~ripley/ University of Oxford, Tel: +44 1865 272861 (self) 1 South Parks Road, +44 1865 272866 (PA) Oxford OX1 3TG, UK Fax: +44 1865 272595
I have forwarded comments on these issues, in a separate message that is entitled: "Strata and Degrees of freedom in anova and multi-level modeling" John Maindonald email: john.maindonald@anu.edu.au phone : +61 2 (6125)3473 fax : +61 2(6125)5549 Mathematical Sciences Institute, Room 1194, John Dedman Mathematical Sciences Building (Building 27) Australian National University, Canberra ACT 0200. On 15 Feb 2006, at 10:00 PM, r-help-request@stat.math.ethz.ch wrote:> From: Peter Dalgaard <p.dalgaard@biostat.ku.dk> > Date: 15 February 2006 3:26:27 AM > To: WPhantom <wp1@tiscali.fr> > Cc: r-help@stat.math.ethz.ch > Subject: Re: [R] A concrete type I/III Sum of square problem > > > WPhantom <wp1@tiscali.fr> writes: > >> Thanks Brian for the reference. >> I just discover that it is available in our >> library so I going to take it & read it soon. >> Actually, I don't even know the difference >> between a multistratum vs a single-stratum AOV. A >> quick search on google returned me the R materials so that I imagine >> that these concepts are quite specific to R. > > You have to be careful not to confuse Google's view of the world with > Reality... > > The concept of error strata is much older than R, and existed for > instance in Genstat, anno 1977 or so. However, Genstat seems to have > left little impression on the Internet. > >> I will read the book first before asking for more informations. > > The executive summary is that the concept of error strata relies > substantially on having a balanced design (at least for the random > effects), so that the analysis can be decomposed into analyses of > means, contrasts, and contrasts of means. For unbalanced designs, you > usually get meaningless analyses. > > >> Thanks >> >> Sylvain Clément >> >> At 12:38 14/02/2006, you wrote: >>> More to the point, you are confusing >>> multistratum AOV with single-stratuam AOV. For >>> a good tutorial, see MASS4 (bibliographic >>> information in the R FAQ). For unbalanced data >>> we suggest you use lme() instead. >>> >>> -- >>> Brian D. Ripley, ripley@stats.ox.ac.uk >> >> ______________________________________________ >> R-help@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 >> > > -- > O__ ---- Peter Dalgaard Øster Farimagsgade 5, Entr.B > c/ /'_ --- Dept. of Biostatistics PO Box 2099, 1014 Cph. K > (*) \(*) -- University of Copenhagen Denmark Ph: (+45) > 35327918 > ~~~~~~~~~~ - (p.dalgaard@biostat.ku.dk) FAX: (+45) > 35327907[[alternative HTML version deleted]]
> WPhantom <wp1 at tiscali.fr> writes: > >>> Thanks Brian for the reference. >>> I just discover that it is available in our >>> library so I going to take it & read it soon. >>> Actually, I don't even know the difference >>> between a multistratum vs a single-stratum AOV. A >>> quick search on google returned me the R materials so that I imagine >>> that these concepts are quite specific to R. > > You have to be careful not to confuse Google's view of the world with > Reality... > > The concept of error strata is much older than R, and existed for > instance in Genstat, anno 1977 or so. However, Genstat seems to have > left little impression on the Internet. > >>> I will read the book first before asking for more informations. > > The executive summary is that the concept of error strata relies > substantially on having a balanced design (at least for the random > effects), so that the analysis can be decomposed into analyses of > means, contrasts, and contrasts of means. For unbalanced designs, you > usually get meaningless analyses. >Can you (prof. Dalgaard) please point us to relevant book with these topics. I am very interested in it since my data are often unbalanced.>>> Thanks >>> >>> Sylvain Cl?ment >>> >>> At 12:38 14/02/2006, you wrote: >> >>>> >More to the point, you are confusing >>>> >multistratum AOV with single-stratuam AOV. For >>>> >a good tutorial, see MASS4 (bibliographic >>>> >information in the R FAQ). For unbalanced data >>>> >we suggest you use lme() instead.I do not have the whole book in my head as prof. Ripley probably does, but I can not recall to read about this in MASS4. I am sure I am wrong and would you (prof. Ripley) be please so kind and point us to relevant chapters/pages. Many thanks. -- Lep pozdrav / With regards, Gregor Gorjanc ---------------------------------------------------------------------- University of Ljubljana PhD student Biotechnical Faculty Zootechnical Department URI: http://www.bfro.uni-lj.si/MR/ggorjan Groblje 3 mail: gregor.gorjanc <at> bfro.uni-lj.si SI-1230 Domzale tel: +386 (0)1 72 17 861 Slovenia, Europe fax: +386 (0)1 72 17 888 ---------------------------------------------------------------------- "One must learn by doing the thing; for though you think you know it, you have no certainty until you try." Sophocles ~ 450 B.C.