Research Paper: A COMPARATIVE STUDY OF ACTUAL VALUE AND ESTIMATED MISSING VALUE IN SOME ORTHOGONAL BLOCK DESIGNS

A COMPARATIVE STUDY OF ACTUAL VALUE AND ESTIMATED      MISSING VALUE IN SOME ORTHOGONAL BLOCK DESIGNS

A. V. Oladugba and C. E. Ogbonnaya

Department of Statistics

University of Nigeria, Nsukka

ABSTRACT

The purpose of this work is to find out if the same conclusion will be reached when “actual value” and “estimated missing value” of some orthogonal block designs are compared using three imputation of estimating missing value methods (the mean method, least squares method and regression imputation method), and also to see if there is any significant difference in the three imputation methods chosen. Estimation of missing values using these imputation methods do not void the result obtained from statistical analysis since the conclusion reached using any imputation method agrees with the conclusion reached using the complete data. From the result obtained in the significance test, we conclude that the difference in estimates for missing observations obtained using the mean method, least squares method and regression imputation is not significant. This implies that the difference in estimates obtained using the three methods are ignorable.

For further details, contact the author on chexyworld@yahoo.com.

Further works on Estimation of Missing Data in Latin Square & Graceo-Latin Squares coming up.