[R] Some matrix and sandwich questions

Michael Ash mash at econs.umass.edu
Tue Oct 30 17:36:24 CET 2007


On 10/30/07, Christian Ritz <ritz at life.ku.dk> wrote:

> Use 'model.matrix' to construct a design matrix:
>
> x <- runif(10,0,10)
> model.matrix(~x)

Thanks!  That's super helpful.

> The following article by A. Zeileis provides more details on the
> package 'sandwich':
> http://www.jstatsoft.org/v16/i09/
> (see also the references).

This is also very helpful, although I still have questions about
terminology (because I am unfamiliar with the phrase "empirical
estimating function").  Is the term "empirical estimating function"
equivalent to "the expectation of the score set equal to zero" where
score is the vector of first derivatives of the (log) likelihood
function?

I think that the paper is written at a higher level of abstraction
about estimators than I am used to.  I'd be grateful if someone could
briefly help me understand the Zeileis paper using either MLE, or even
OLS, *examples*.

Thanks very much.

Best,
Michael



-- 
Michael Ash, Associate Professor
  of Economics and Public Policy
Department of Economics and CPPA
University of Massachusetts
Amherst, MA 01003
Tel +1-413-545-6329 Fax +1-413-545-2921
Email mash at econs.umass.edu
http://people.umass.edu/maash



More information about the R-help mailing list