[R] Problem with anova.lmRob() "robust" package

[Ista Zahn]

I found the question really confusing as well, but see below.

On Thu, Jul 28, 2011 at 11:42 AM, David Winsemius
<dwinsemius at comcast.net> wrote:
>
> On Jul 28, 2011, at 9:13 AM, m.fenati at libero.it wrote:
>
>>
>> Dear R users,
>> I'd like to known your opinion about a problem with anova.lmRob() of
>> "Robust" package that occurs when I run a lmRob() regression on my dataset.
>> I check my univariate model by single object anova as anova(lmRob(y~x)).
>> If I compare my model with the null model (y~1), I must obtain the same
>> results,
>> but not for my data.
>> Is it possible?
>>
>> My example:
>>
>> x<-c(rep(0,8),rep(1,8),rep(2,7))
>>
>> y<-c(1,0.6,-0.8,0.7,1.6,-0.2,-1.2,-3.8,-1.8,-2.6,-1.7,-2.1,-0.3,-1.4,1.4,-0.3,-0.3,0.5,0.4,-0.9,-1.6,0.4,0.4)
>> library(robust)
>> lmR<-lmRob(y~factor(x))
>> anova(lmR)
>> lmR0<-lmRob(y~1)
>> anova(lmR,lmR0)
>>
>> If I run the code omitting the factor() (then treating "x" as continuous),
>> the results are the same..
>>
>
> I do not get the same results with that code. And the code does not appear
> to track your description, since the second model does not have an "x" term
> in it. Even when I create the model that it sounded as though you would have
> written, namely lmR0 <- lmRob(y ~ x),  it is clearly _not_ the same result.
>
>> coef(lmR)
> (Intercept)  factor(x)1  factor(x)2
>  0.2428571  -1.3432007  -0.4000000
>> coef(lmR0)
>  (Intercept)            x
> -0.509524217  0.005820355
>>
>> What is the explanation of these different results?
>
> Since you didn't post your results and since your complaint was that they
> are "the same", it's hard to know what you are talking about.

I think the question is why

lm1 <- lm(y ~ factor(x))
lm0 <- lm(y ~ 1)
anova(lm1)
anova(lm0, lm1)

gives the same result, but

lmR<-lmRob(y~factor(x))
lmR0<-lmRob(y~1)
anova(lmR)
anova(lmR,lmR0)

does not.

I don't know the answer, but I think it is an interesting question.

Best,
Ista

>
> --
> David Winsemius, MD
> West Hartford, CT
>
> ______________________________________________
> R-help at r-project.org 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.
>



-- 
Ista Zahn
Graduate student
University of Rochester
Department of Clinical and Social Psychology
http://yourpsyche.org