论文信息 - Variance Components Estimators for the Balanced Two-Way Mixed Model

Variance Components Estimators for the Balanced Two-Way Mixed Model

For a mixed two-way balanced design model, the maximum likelihood (ML) and restricted ML (REML) estimators of the variance components are obtained and compared under the nonnegativity requirements of the variance components. The true ML and REML estimators are always nonnegative and have uniformly smaller mean squared error (MSE) than the ML and REML variance estimators of Corbeil and Searle (1976, Biometrics 32, 779-791), denoted by CML and CREML, respectively.

C. Kapadia | Kwan R. Lee | C. H. Kapadia

[1] John A. Nelder,et al. The interpretation of negative components of variance , 1954 .

[2] H. Sahai. A Comparison of Estimators of Variance Components in the Balanced Three-Stage Nested Random Effects Model Using Mean Squared Error Criterion , 1976 .

[3] W. A. Thompson. The Problem of Negative Estimates of Variance Components , 1962 .

[4] S. R. Searle,et al. A Comparison of Variance Component Estimators , 1976 .

[5] David A. Harville,et al. Alternative Formulations and Procedures for the Two-Way Mixed Model , 1978 .

[6] L. R. Verdooren. On estimation of variance components , 1980 .

[7] H. D. Patterson,et al. Recovery of inter-block information when block sizes are unequal , 1971 .

[8] S. Zacks,et al. Mean Square Efficiency of Estimators of Variance Components , 1969 .