[R-sig-ME] Negative estimates of variance component

[Ben Bolker]

  I claim that this is in the category of "consistency with older
method-of-moments estimates" ... I know that population geneticists
still like to think in terms of variance components, but wouldn't one
ideally want to deal with the negative correlation built into the
system by estimating it more or less directly (i.e. via a structured
variance-covariance model that has nonnegative variances but could
have negative covariances) rather than by estimating a negative
variance component?

  Ben Bolker (not a population geneticist so possibly missing the point)

On Fri, Jul 30, 2010 at 12:39 PM, Jerome Goudet <jerome.goudet at unil.ch> wrote:
> Hi all,
>
> Here is an example: genes are nested in individuals, themselves nested in
> populations.  If individuals avoid mating with relatives, then the variance
> component of allele frequencies among individuals within population is
> expected to take negative values, as 2 genes taken from 2 different
> individuals are more similar than two genes taken from the same individual.
> See Weir & Cockerham (1984) Evolution for instance.
>
>
>
>
> Ben Bolker wrote:
>
>   Pardon my asking, but why? For consistency with older (arguably less
> correct) method-of-moments estimates that could give negative
> estimates?
>
> On Fri, Jul 30, 2010 at 1:32 AM, Chris Brien <Chris.Brien at unisa.edu.au>
> wrote:
>
>
> Dear mixed modellers,
>
> I have a data set that gives me an estimate of 0 for one of the variance
> components. I wanted to allow for the estimate of the component to be
> negative. My search of the documentation led me to believe that this is not
> possible. Am I right, or did I miss something?
>
> Cheers,
>
> Chris Brien
> -----
> University of South Australia
> ADELAIDE  5001  South Australia
> WEB page:  <http://people.unisa.edu.au/Chris.Brien>
>
> _______________________________________________
> R-sig-mixed-models at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
>
>
>
> _______________________________________________
> R-sig-mixed-models at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
>
>
>
> --
> Jérôme Goudet
> Dept. Ecology & Evolution
> UNIL-Sorge, CH-1015 Lausanne
>
> mail:jerome.goudet at unil.ch
> Tel:+41 (0)21 692 4242
> Fax:+41 (0)21 692 4265