Growth curve model with covariance structures (Q1327736): Difference between revisions

The paper studies the statistical inference in growth curve models with arbitrary covariance structure, uniform covariance structure and serial covariance structure. The paper consists of two parts which discuss the cases of complete data and missing data, respectively. In the first part, the properties of the MLE's in growth curve models with various covariance structures and their density functions are given. Asymptotic expansions of the distribution of maximum likelihood estimators (MLE's) and the likelihood ratio statistics for testing the uniform covariance structures are derived. Testing hypotheses, and confidence intervals and regions for mean parameters are also considered. The second part deals with the case of missing data. The MLE's of mean parameters and their asymptotic properties are studied. The author compares asymptotic properties of the MLE's in the cases of complete data and missing data. Finally, the author considers testing hypotheses for mean parameters in the case of missing data.