An alternate statistical method for dealing outliers in perennial crop experiment

Abstract

A statistical method based on Robust ANOVA to handle outliers induced high coefficient of variation (CV) in pooled (2011-2018) analysis of long-term Mango cv. Totapuri rootstock trail was suggested. Based on the results, it was concluded that the rootstock treatment T3: Olour (average yield over the period 2011 to 2018 as 57.21 kg/tree) as the best. Precision gained as estimated by reduction in CV (%) was in the range of 11.01 % to 78.9 %. SAS IML codes were built-in for analysis. Hence, this study calls for employing robust ANOVA approach in testing the significance of evaluated treatments in a designed perennial crop experiment with high CV that would have reduced the sensitivity of testing the significance of treatment differences otherwise.