مقارنة اربع طرائق لتقدير القيم المفقودة في تصميم المربع اللاتيني

Abstract

Four methods for tackling missing values in Latin square design have been
presented: Yates, Harry, Rubin, and the method of Haseman and Gaylar. To
make preference among these methods some statistical measurements have
been used, which are: lowest value of the mean square error (MSe), highest
value of the mean square treatments (MSt) and the value of significant
differences between treatments (F cal.). The easiest path of statistical
analysis has been taken into account. It has been found that the most
preferable method is Yates' method which has the most complicated
application whenever the number of missing values are increased, followed
by Harry method which has a moderate application difficulty, then Haseman
and Gaylar method, and finally Rubin method. The last two methods have
the easiest applications.