Skip to content
Prev 8645 / 20628 Next

mean and variance of random effects in glmer

Don't assume that there is not quasi-separation, but rather check it to be sure that it is not there. I'm pretty sure that it is a case of complete separation. Deal with that first.
________________________________________
Van: Ken Kelley [kkelley at nd.edu]
Verzonden: dinsdag 24 juli 2012 20:38
Aan: ONKELINX, Thierry
CC: r-sig-mixed-models at r-project.org
Onderwerp: Re: mean and variance of random effects in glmer

Hi Thierry,

Thanks for your thoughts on this. I hadn't considered quasi-separation, but I don't think that is it. Actually, the issue is that I slipped into thinking that the random effects were the conditional means (like in a linear mixed effects model). Rather, they are the conditional modes. Thus, the mean of the random effects need not be zero as I initially expected (and as would be the case in a linear mixed effects model).

But, I still expected the variance of the random effects to match the output (it is 18.9 in the output yet 7.8 when I calculate it on the random effects directly).

Best wishes,
Ken
On Jul 24, 2012, at 5:23 AM, "ONKELINX, Thierry" <Thierry.ONKELINX at inbo.be> wrote:

            
* * * * * * * * * * * * * D I S C L A I M E R * * * * * * * * * * * * *
Dit bericht en eventuele bijlagen geven enkel de visie van de schrijver weer en binden het INBO onder geen enkel beding, zolang dit bericht niet bevestigd is door een geldig ondertekend document.
The views expressed in this message and any annex are purely those of the writer and may not be regarded as stating an official position of INBO, as long as the message is not confirmed by a duly signed document.