ABSTRACT
The asymptotic variance-covariance matrix of the maximum likelihood estimates (P.,, &) is obtained by inverting the Fisher information matrix with elements that are negatives of expected values of the second-order partial derivatives of the loglikelihood function with respect to the parameters. Accordingly, as given by Cohen (1961), we obtain
0"2 V(P.,) = N f.Lll•
0"2 Cov(P.,,&) = N f.L12·
(2.6.1)
where f.L 1 ~> f.L 12 , and J..L22 are functions of <!> 11 , <!> 12 , and <!>22 , which are elements of the information matrix.