![]() ![]()
The addition of Store ID as
a random-effects factor reduces the overall unexplained variability,
marked as Error,
from 4151946.415 to 3411647.185.
It has also reduced the variation
explained by the main effects, USECOUP
and SHOPFOR. This is quite natural and simply
shows that some of the variation originally explained by these model terms
can be better explained by Store ID.
The important fact here is that the proportional reduction in the variation
explained by each of the main effects was not as great as the reduction
in the unexplained variation. Thus, the partial eta squared has increased
for each of the model effects.
|
|
Test of Between-Subjects Effects |