[R-sig-ME] Fixed effects cannot be interpreted if Y is transformed in (g)lme's [was: conditional vs. marginal coefficients in GLMM]

[David Duffy]

On Mon, 14 Mar 2011, lorenz.gygax at art.admin.ch wrote:

> Now, of course, in such a generalized model the response is 
> (logit-)transformed.

If you use a log link, then this is not a problem (eg Ritz &
Spiegelman Stat Methods Med Res 2004 13:309-323).  This also
will give you risk ratios directly, which are more natural for your
setup.

-- 
| David Duffy (MBBS PhD)                                         ,-_|\
| email: davidD at qimr.edu.au  ph: INT+61+7+3362-0217 fax: -0101  /     *
| Epidemiology Unit, Queensland Institute of Medical Research   \_,-._/
| 300 Herston Rd, Brisbane, Queensland 4029, Australia  GPG 4D0B994A v