I read a paper today where someone used the t-statistic as the input to a group-level t-test. My understanding is that it is the beta weight that should be propagated up to 2nd level analyses. Am I wrong? If not, how wrong is it to use the t-statistic?
cheers
Iain