If a unimodal dataset x with mean 0
E(x) = 0
is positively skewed
E(x^3) > 0,
does it imply that mean squared value of positive data (y: x/ x>0) is more than negative data (z: x/ x<0) within x
E(y^2) > E(z^2) ??
or that variance of y is more than variance of z??
