 |
 |
Hi - in FLAME both ME (either FLAME 1 or 1+2) and FE use the same variance weighting that is doing exactly the kind of thing you are describing. So that's already sorted, and the question of whether to use FE or ME is a separate issue, i.e., do you want to be able to make a statement about the general population from which the subejcts were drawn, in which case use ME, which is the normal thing to do. Cheers. On 2 Feb 2008, at 23:39, Jessica Kirkland wrote: > Hi all, > > I have a question about variance weighting. I have a task in which I > am > modeling correct and incorrect trials for a number of conditions > separately. > I have 80 subjects, all of whom have different numbers correct and > incorrect, and some who have zero or very few trials for a given EV > (e.g. > Angry face incorrect). At the group level, I would like to take > individual > subject variance into account, such that those subjects with little to > contribute to a group contrast are not unduly influencing the group > contrast. In reading through the FEAT online help, I am still > confused about > whether I should be using FE or ME analysis. I would have thought > ME, using > FLAME, but will this allow for "weighted" group estimates, as I > would like? > Thank you for the help! > > Jessica Kirkland > --------------------------------------------------------------------------- Stephen M. Smith, Professor of Biomedical Engineering Associate Director, Oxford University FMRIB Centre FMRIB, JR Hospital, Headington, Oxford OX3 9DU, UK +44 (0) 1865 222726 (fax 222717) [log in to unmask] http://www.fmrib.ox.ac.uk/~steve ---------------------------------------------------------------------------