Page: 01-06. Test on groups of repeated measures experiments with unequal error variance structure.

 A. S. M. R. ROUF1

1=Abu Sayeed Md. Ripon Rouf, Associate Professor, Department of Statistics, Jagannath University, Dhaka, Bangladesh. Email: ripon1969@gmail.com

ABSTRACT

Any group of repeated measures experiments can be analyzed with equal variance co-variance structure under consideration. The same repeated measures experiments may be conducted in different places. When the variance covariance are not equal then the analysis of groups of repeated measures experiment are questionable. This study derives some tests and techniques to solve the problem with an example from a group of experiment on Mahogany trees conducted in different station of Jahangirnagar University, Savar, Dhaka, Bangladesh. Although the variances of different places are not homogenous, it would be possible to conduct the analysis at a time and overcome the situational bias by the proposed test procedure. It is observed that significant variation of growth of the plants over the time and the places are present in the experiment. The experiments were carried out by the Green-Belt project of Jahangirnagar University, Savar, Dhaka, Bangladesh and the study period was November 1993 to July 1994.

Keywords: Spectral De-composition, Treatment, Contrast and Tolerance limit.