[R] Drop1 and weights

Prof Brian Ripley ripley at stats.ox.ac.uk
Thu Mar 2 14:25:47 CET 2006


It looks like at some point weighted.residuals() was changed, and broke 
this.  The models are fitted correctly, but AIC is wrong.  However, note 
that it does work correctly for a glm() fit which gives you a workaround.

You example is not reproducible, so I was unable to test the workaround 
nor the correction which will short;y be in R-patched.

On Wed, 1 Mar 2006, Yan Wong wrote:

> Hi,
>
> If I used drop1 in a weighted lm fit, it seems to ignore the weights
> in the AIC calculation of the dropped terms, see the example below.
> Can this be right?
>
> Yan
>
> --------------------
>
> library(car)
> > unweighted.model <- lm(trSex ~ (river+length +depth)^2-
> length:depth, dno2)
> > Anova(unweighted.model)
> Anova Table (Type II tests)
>
> Response: trSex
>              Sum Sq  Df F value  Pr(>F)
> river        0.1544   1  3.3151 0.07100 .
> length       0.1087   1  2.3334 0.12911
> depth        0.1917   1  4.1145 0.04461 *
> river:length 0.0049   1  0.1049 0.74652
> river:depth  0.3008   1  6.4567 0.01226 *
> Residuals    5.9168 127
> ---
> Signif. codes:  0 ?***? 0.001 ?**? 0.01 ?*? 0.05 ?.? 0.1 ? ? 1
> >
> > drop1(unweighted.model)
> Single term deletions
>
> Model:
> trSex ~ (river + length + depth)^2 - length:depth
>              Df Sum of Sq     RSS     AIC
> <none>                       5.92 -401.97
> river:length  1  0.004889    5.92 -403.86
> river:depth   1      0.30    6.22 -397.37
>
> ### Both drop1 & Anova suggest dropping river:length
> ### Compare with the following:
>
> > weighted.model <- lm(trSex ~ (river+length +depth)^2-length:depth,
> dno2, weights=males+females)
> > Anova(weighted.model)
> Anova Table (Type II tests)
>
> Response: trSex
>              Sum Sq  Df F value  Pr(>F)
> river         1.471   1  3.3775 0.06843 .
> length        1.002   1  2.2999 0.13187
> depth         2.974   1  6.8295 0.01005 *
> river:length  0.075   1  0.1733 0.67790
> river:depth   2.020   1  4.6397 0.03313 *
> Residuals    55.303 127
> ---
> Signif. codes:  0 ?***? 0.001 ?**? 0.01 ?*? 0.05 ?.? 0.1 ? ? 1
> >
> > drop1(weighted.model)
> Single term deletions
>
> Model:
> trSex ~ (river + length + depth)^2 - length:depth
>              Df Sum of Sq     RSS     AIC
> <none>                      55.30 -104.71
> river:length  1    -49.33    5.97 -402.79
> river:depth   1    -49.04    6.26 -396.39
>
> ______________________________________________
> 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
>

-- 
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




More information about the R-help mailing list