Yes, the difference between the two statistical tests is very subtle.
-fa generates two subbricks: the first one is the
intensity of the effect for factor A, and the other the F statistic for the
equality of factor A effect. Intensity of effect A, defined as square root of mean sum of squares of A (MSA), measures the variation accounted for by factor A, and the F statistic of factor A, MSA/MSE (or MSA/MSAB in the case of mixed design), reveals the relative magnitude of the variation accounted for by factor A to that by random errors. In other words, the F value tests whether all the factor levels are equal or there is at least one level different from others.
-acontr 1 1 1 1 generates two subbricks too: the first one is the estimate of the
magnitude (or contrast versus zero) of all factor A levels, and the other is the corresponding t statistic, demonstrating the
significance of the magnitude. Here magnitude is different from the intensity in -fa because magnitude measures the sum of all factor A level effects, not the variation. And -acontr 1 1 1 1 is focused on whether the total magnitude of factor levels is significantly different from zero, compared to the equality test of factor levels in -fa.
Gang
