Correlated. Just like squares are correlated to rectangles.

Proof that T causes high aggression in isolation, in humans, thus far, has not been decisive. There is no consensus on this. Looking at the absence of the inverse (low T decreases aggression) would tip a few things.