I've had a question about a random effects analysis, where someone has
reported t-values, but not the degrees of freedom. Is the t in SPM at the
second level somehow not comparable to t statistics done in the behavioural
domain, and therefore df aren't required?
Thanks in advance,
Jamie
|