A Monte Carlo study of local identifiability and degrees of freedom in the asymptotic likelihood ratio test
Roderick P. McDonald, William R. Krane · British Journal of Mathematical and Statistical Psychology · 1979
In applications to restricted factor analysis, it is shown that the determination of local identifiability and degrees of freedom in the asymptotic likelihood ratio test, from the eigenvalues of certain matrices, is not unambiguous in all cases. Some results are given that serve to indicate appropriate reparameterizations of unidentified problems.