Abstract
In the following we consider the correlation betweenS 2 and the least squares estimator in the linear regression model. We are interested in situations where these two statistics are uncorrelated though the errors are correlated. Conditions are developed without normality assumption, only assuming finite fourth moments of the error distributions.
Similar content being viewed by others
References
Christensen, R. (1987): Plane Answers to Complex Questions: The Theory of Linear Models.Springer, New York.
Kleffe, J. (1978): Simultaneous Estimation of Expectation and Covariance Matrix in Linear Models,Mathematische Operationsforschung und Statistik, Series Statistics 9, p. 443–478.
Rao, C.R. andKleffe, J. (1988): Estimation of Variance Components and Applications,North Holland, Amsterdam.
Rasch, D. (1989): Einführung in die mathematische Statistik I, 3. Auflage,Deutscher Verlag der Wissenschaften, Berlin.
Author information
Authors and Affiliations
Additional information
Support by Deutsche Forschungsgemeinschaft Grant No. Tr 253/1-2 is gratefully acknowledged.
An erratum to this article is available at http://dx.doi.org/10.1007/BF02926398.
Rights and permissions
About this article
Cite this article
Knautz, H., Trenkler, G. A note on the correlation betweenS 2 and the least squares estimator in the linear regression model. Statistical Papers 34, 237–246 (1993). https://doi.org/10.1007/BF02925544
Received:
Revised:
Issue Date:
DOI: https://doi.org/10.1007/BF02925544