Use the text field to specify
random effects
by which to model the
covariance
structure of the response
variable
.
Random effects are typically comprised of
class variables
and their interactions. Be sure to specify any class variables in the
Additional Class Variables
field above.
Do
not
repeat any effects here that are confounded with
Time
,
Treatment
, or any effect specified in the
Additional Fixed Effects
field.
Commonly used random effects are
SITEID
(for multi-center trials) and
STUDYID
(for data assembled from multiple studies).
|
If you specify more than one random effect, then you must separate the effects with
spaces
. For example, assuming
A
and
B
are listed as Class Variables, entering
A B
in the text field specifies variables
A
and
B
as random effects.
|