The
dispersion matrix for the parameter estimates

is the matrix

, where

is the covariance of

and

. The dispersion matrix is calculated according to the formula

where S
2 is the estimated variance, as defined above, and
X and

are the regression matrix and its transpose, respectively.