This paper is concerned with the null distribution of the likelihood ratio test statistic −2log Λ for testing the adequacy of a random-effects covariance structure in a parallel profile model. It is known that the null distribution of −2log Λ converges to χ2f or 0.5χ2f + 0.5χ2f+1 when the sample size tends to infinity. In order to extend this result, we derive asymptotic expansions of the null distribution of −2log Λ. The accuracy of the approximations based on the limiting distribution and an asymptotic expansion are compared through numerical experiments.
No takes yet. Share an insight, caveat, or question.
Inatsu et al. (2016) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: