Re: [Marsbar-users] Nutty F-values

[Matthew B. <mat...@gm...>]

Hi,

> I’ve been using marsbar now for quite a few years and over this time I have
> occasionally run into some rather striking anomalies between the T and F
> values that I get for my ROIs (for example, a T of 2.5 and an F of 125 for
> an equivalent contrast [i.e. 1 1 -1 -1]). In this case the F-test should
> have a p value twice that of the one-tailed T-test.

That does sound strange.  Can you find some data and a design that
replicates this and send it to me by email or some other way?

Best,

Matthew