[R] Correlation of coefficients?

[Peter Ehlers]

If you're talking about regression models, then I'm puzzled that
this occurs "frequently" in R, since the summary.lm() function
defaults to 'correlation = FALSE'. S-PLUS defaults to TRUE.
Can you give an example where R gives you the correlations of
coefficients?

Anyway, parameter estimators are correlated random variables and,
in the case of regression, summary(model, correlation = TRUE) gives
the estimated correlation matrix. This must be somewhere in MASS,
the book; any decent linear models text should will have a
discussion, e.g. Draper and Smith, Applied Regression Analysis.

Peter Ehlers

asako Ishii wrote:

> Hi R users!
> 
> One thing I cannot understand with R is the frequently appearing "
> Correlation of coefficients". I do not find a through going explanation
> of this concept either on the help pages of R or on the internet. A
> Google search has shown many examples of R/S-Plus outputs but no
> explanations of meaning.
> 
> I am afraid this turns out to be a silly and unworthy question. Would
> anyone be kind enough to show relevant literature or a text book to a
> layperson studying on his/her own? 
> 
> Thanks in advance.
>