The best way to test this would be to assemble a group of kamihime, all with attack increase assist skills, and see if this increases the damage the soul does (vs. a group with 4 kamihime with no assist skills that influence damage, or a group that only includes the soul). Theoretically a stacking buff from 4 kamihime, even if the individual buffs were small, would add up to a very significant effect that would be readily observable. To maximize the relative effect, the soul should have no buffs from assault skill (so change the soul's element accordingly). Obviously the tests would have to be run on the same enemy (like a Disaster boss), since different enemies could have different base defense values.
Assuming a cumulative effect of at least 20% or so, 10-20 normal attacks each in the experiment and control group should be sufficient to spot it statistically.