Repeated Measures with Missing Data: Parameter Estimation Via Covariance Analysis

Luz Mery González & Luis Alberto López

 

 

Abstract

In this paper the estimation of parameters and the different adjusted sums of squares for balanced designs in repetead measures with incomplete information is done through three procedures: the Bartlett's method of covariance analysis; a multivariate method with complete data, and finally, an alternative multivariate method using the available information in the experimental arrangements. Numerical applications using the above procedures are done.

 

Keywords: Covariance analysis, missing information, longitudinal data, repeated measures, generalized least squares, multivariate analysis.

 

PDF (Spanish)