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.