[R] R vs SPSS contrasts

Peter Dalgaard p.dalgaard at biostat.ku.dk
Sun Oct 12 22:33:34 CEST 2008


Gabor Grothendieck wrote:
> I found this link:
>
> http://webs.edinboro.edu/EDocs/SPSS/SPSS%20Regression%20Models%2013.0.pdf
>
> which indicates that the contrast in SPSS that is used
> depends not only on the contrast selected but also on the
> reference category selected and the two can be chosen
> independently.  Thus one could have simple/first, simple/last,
> deviation/first, deviation/last, etc.  An R contr.SPSS function
> would have to specify both the deviation type and the
> first/last in order to handle all SPSS variations.
>   

Not sure about "have to"; if so, contr.SAS also needs extension since 
SAS does different things in different procedures (the help file does 
say so).

Notice, BTW, that the cited reference defines contrast matrices 
inversely to R (it maps from group levels to parameters).

-- 
   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 at biostat.ku.dk)              FAX: (+45) 35327907



More information about the R-help mailing list