Hello,
I find your recommended procedure not very sensible but:
a) sum(fit$residuals^2)/fit|$d||f.residual
|
b) This is a statistical question but, for instance, in the case of
repeated mesures. But you'd use the precision (1/variance) of the
original data, computed by groups of regressor.
Hope this helps,
Rui Barradas
Em 13-10-2012 20:07, Eiko Fried escreveu:> Hello.
>
> I'm am trying to follow a recommendation to deal with a dependent
variable
> in a linear regression.
>
> I read that, due to the positive trend in my dependent variable residual vs
> mean function, I should
> 1) run a linear regression to estimate the standard deviations from this
> trend, and
> 2) run a second linear regression and use 1 / variance as weight.
>
> These might be terribly stupid questions, but:
> a) Where do I find the standard deviations from this trend in the lm
> output? Except for coefficients I only see residual std error, DF, R^2, F
> and p values.
> b) Why exactly does one adjust a linear regression by a 1/variance weight,
> and in what cases?
>
> Thanks
> T
>
> [[alternative HTML version deleted]]
>
> ______________________________________________
> R-help@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.
[[alternative HTML version deleted]]