[R] ANOVA precision

[Peter Dalgaard BSA]

Bill Venables <William.Venables at cmis.csiro.au> writes:

> > Rers,
> > 
> > Hi, I just ran a one-way ANOVA using R and compared the output to a
> > one-way ANOVA I created using SAS on the same data set.
> > 
> > Here are the results:
> > 
> > R
> > 
> >            DF  Sum Sq   Mean Sq  F value  Pr(>F)
> > Base       3   27.7449  9.2483   107.99   < 2.2e-16 *** 
> > Residuals 32    2.7449  0.0856
> > 
> > SAS
> > 
> >             Sum Of
> > Source DF   Squares Mean   Square       F Value   Pr > F
> > Model   3   27.69085439    9.23028480   107.55    <.0001
> > Error  32    2.74643538    0.08582611
> > 
> > In a "philosophical" sense, are the rounding differences worth worrying
> > about (they may not be significant in this case, but sometimes the
> > differences are in the second digit), or is this an example of "use at
> > your own risk"?
> 
> Well, if SAS has an accuracy problem I'm sure they would be glad to
> hear of it and keen to fix it up...  :-)

Actually, this discrepancy looks a bit too large to be expected in
either of the two programs. Is it certain that it really is the exact
same data values that are being used? The difference in total
variation matches quite well what would happen if one program used
values rounded to nearest .1 and the other to .01.

What is the output of PROC MEANS and tapply(response,Base,mean)
respectively?


-- 
   O__  ---- Peter Dalgaard             Blegdamsvej 3  
  c/ /'_ --- Dept. of Biostatistics     2200 Cph. N   
 (*) \(*) -- University of Copenhagen   Denmark      Ph: (+45) 35327918
~~~~~~~~~~ - (p.dalgaard at biostat.ku.dk)             FAX: (+45) 35327907
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._