Hello,
I am currently preparing an analysis regarding with a 4x2 factorial design, where some cells have a very small number of scans (e.g.: 10-15). I don't have any a priori reason to believe that some cell could have a *widely* different variance than another one, neither when moving along factor 1 nor factor 2. However I don't have either any a priori reason to believe that they're going to be identical. Indeed, my best a priori assumption would be that some of those cells *might* have a *slightly* higher variance. Yet again, due to the small numerosity in that cell, I'm pretty sure that the estimate of the variance would suffer from lots of noise and be very imprecise (even if I'm not sure how SPM estimates it. Does it use some "tricks" to improve it or just plain definition?).
Therefore I'm asking you... On which side should I err?
Cheers,
Luca
Rispetta l’ambiente: non stampare questa mail se non è necessario.
Respect the environment: if it's not necessary, don't print this mail.
|