Skip to content

R-sig-mixed-models Digest, Vol 47, Issue 39

1 message · Ryan King

#
It's just a caveat; neither the variance inflation of Kackar and
Harville (84) nor the t-statistics degree of freedom juggling that SAS
uses are being applied, so the variance is understated relative to
those.  My understanding is that it makes little difference in many
cases in practice, and I have no idea if the coverage properties
perform as advertised.

Ryan King
Dept Health Studies
University of Chicago

On Mon, Nov 22, 2010 at 5:00 AM,
<r-sig-mixed-models-request at r-project.org> wrote: