Skip to content
Prev 15559 / 20628 Next

Question on random effect

Going back to the independence assumption, see your comment and below my
comment (thanks):

Joaqu?n asked: do the levels of the random effect have to be independent?
For example, if my random effect is the identity of sites and it is
associated to the intercept, do sites have to be independent?

Ben said:
Yes, the assumption is that the random effects are (conditionally)
independent.  It can help to specify covariates  (such as
latitude/longitude or eastings/northings, or environmental conditions
[temperature, elevation, etc.]) for sites to mop up some of the
independence. It is theoretically possible, although I don't know of
an easy off-the-shelf way to do it, to impose (e.g.) spatial
correlation structures at the level of the random effects ...  or you
could examine the spatial dependence of the conditional modes/random
effects and try to convince yourself it was weak ...

Joaqu?n current comment: However, I?ve seen that response observations
inside a random effect variable may be correlated, and this can be
quantified by Intraclass Correlation Coefficients (ICCs). Is this lack of
independence among random effect levels?

On Tue, Jul 11, 2017 at 7:14 AM, Joaqu?n Aldabe
<joaquin.aldabe at gmail.com> wrote:
2017-07-11 8:22 GMT-03:00 Ben Bolker <bbolker at gmail.com>: