Open Access
June, 1993 Matrix Weighting of Several Regression Coefficient Vectors
Alan T. James, William N. Venables
Ann. Statist. 21(2): 1093-1114 (June, 1993). DOI: 10.1214/aos/1176349166

Abstract

For small sample random effects models, results are derived which show in certain cases, and indicate in general, that an estimated random effects variance matrix may be used in the weight matrices without causing undue error in the empirically weighted mean. Exact error variances are derived mathematically for the empirically weighted mean for the two sample case in one and two dimensions. Simulation is used to determine errors for a practical example of six five-variate samples. For estimation of their mean, the differences between the samples are ancillary. The biases of the average and weighted mean estimators conditional on these ancillaries is illustrated in a diagram plotting values obtained by simulation. A curious range anomaly is illustrated which arises if random effects are ignored when present.

Citation

Download Citation

Alan T. James. William N. Venables. "Matrix Weighting of Several Regression Coefficient Vectors." Ann. Statist. 21 (2) 1093 - 1114, June, 1993. https://doi.org/10.1214/aos/1176349166

Information

Published: June, 1993
First available in Project Euclid: 12 April 2007

zbMATH: 0778.62048
MathSciNet: MR1232534
Digital Object Identifier: 10.1214/aos/1176349166

Subjects:
Primary: 62H12
Secondary: 62J10

Keywords: conditional bias , cutoff function , efficiency , estimated generalized least squares , Matrix weighting , moment estimator , random effects model , range anomaly , residual maximum likelihood , simulation , small sample random effects model , unbalanced data

Rights: Copyright © 1993 Institute of Mathematical Statistics

Vol.21 • No. 2 • June, 1993
Back to Top