var(ranef(Random Effect)) not the same as the variance component
On 9/7/20 9:22 PM, Simon Harmel wrote:
Ben, This might seem irrelevant to my previous question, but it is from the post you linked in your previous answer. So, is it correct language to say: By including Random-Effects (e.g., random intercepts) of some subjects,?we are **controlling/adjusting/holding constant** subjects's?random?variations in that random-effect (e.g., variation in subjects' initial?status)?
I would probably say something like "incorporating among-subject variation in that term (e.g., the initial status) in the model". Or "accounting for".
On Mon, Sep 7, 2020 at 7:53 PM Simon Harmel <sim.harmel at gmail.com
<mailto:sim.harmel at gmail.com>> wrote:
Much?appreciated, Ben. I will study those resources to better
understand the estimation process.
Thanks again,
Simon
On Mon, Sep 7, 2020 at 6:25 PM Ben Bolker <bbolker at gmail.com
<mailto:bbolker at gmail.com>> wrote:
? ?Yes, that's correct.
?From
https://stats.stackexchange.com/questions/392283/interpreting-blups-or-varcorr-estimates-in-mixed-models/392307#392307
?> the covariance matrix of the empirical Bayes estimates
obtained from
ranef() is related to the covariance of this posterior
distribution [of
conditional modes/BLUPs] whereas VarCorr() is giving the D
matrix, which
is the covariance matrix of the prior distribution of the random
effects. These are not the same.
On 9/7/20 7:22 PM, Simon Harmel wrote:
> Hello All,
>
> A very basic question. Generally, `var(ranef(Random Effect))`
may not
> necessarily be the same as the variance component reported
for that Random
> Effect in the model output, correct?
>
>
> Thank you all,
> Simon
>
>? ? ? ?[[alternative HTML version deleted]]
>
> _______________________________________________
> R-sig-mixed-models at r-project.org
<mailto:R-sig-mixed-models at r-project.org> mailing list
_______________________________________________
R-sig-mixed-models at r-project.org
<mailto:R-sig-mixed-models at r-project.org> mailing list
https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models