¥Random
effects factor = differences between levels in this factor are
modeled as random fluctuations
HUseful for categories not under experimenterÕs control or
observation
HIn FMRI, is especially useful for subjects; a good rule is
åtreat
subjects as a separate random effects factor rather than
åas
multiple independent measurements inside fixed-effect factors
HIn such a case, usually have 1 measurement per cell (each cell is
the combination of a level from the other factor with 1 subject)
åThis is sometimes called a Òrepeated measures ANOVAÓ, when we have
multiple measurements
on each subject (in this case, across different stimulus classes)
HTreating subjects as a random factor in a fully crossed analysis
is a generalization of the paired t-test
åintra-subject
and inter-subject data variations are modeled separately
åwhich can let you detect small intra-subject changes due to the
fixed-effect factors that might otherwise be overwhelmed by larger
inter-subject fluctuations
HMain effect for a random effects factor tests if fluctuations
among levels in this factor have additional variance above that from the other
random fluctuations in the data
åe.g., Are inter-subject fluctuations bigger than
intra-subject fluctuations?
åNot
usually very interesting when random factor = subject
HIt is hard to think of a good FMRI example where both factors
would be random
H3dANOVA2: Usually have 1 fixed factor
and 1 random factor = mixed effects analysis