Vol. 19 No. 2
Exploration and consequences of outlier on potato field crop through factorial experiment
Author(s): MD M. H. MIDDYA, M. ROY, A. SAHA, J. S. BASAK AND A. MAJUMDER
Abstract: Unwanted observations in a field experiment can lead to inaccurate outcomes and violate normality assumptions of ANOVA. Even one outlier can disrupt multi-factor analyses and yield with faulty conclusions. The present study has explored the outlier observation (observations) in factorial experiments using cook statistics. Mean shift model has been taken into consideration for constructing the test statistics. Each erroneous observation’s mean differs from the rest of the observations’ means in the mean shift model. Himadri (2013) has also used another test statistics namely Qt statistics for identifying the outlying observation (or observations) in field experiment. The above test statistics are mainly used for identifying a single or multiple outliers among the set of data. These test statistics have been materialised on real experimental observation. For examining the validity of a data through cook statistics or Qt statistics, a single observation has been taken into account. After identifying the outlying observation, this observation should be deleted. Then analysis was accomplished by replacing the outlier with their estimated missing value through method of least square technique. Remarkable differences have been noticed by executing the analysis with outlying observation and without outlying observation.g observation.
Download Full Article: Click Here