Test for Outlier and Normality of the Residuals for the Pseudo-1st Order Kinetic Modelling of Glyphosate Adsorption onto Palm Oil Fronds

Abstract

Generally, in nonlinear regression, the residuals of the curve must be distributed normally, and residuals must be tested for the presence of outliers. This is normally done using normality tests such as the Kolmogorov-Smirnov, Wilks-Shapiro and D'Agostino-Pearson and the Grubb's test, the latter test for the presence of an outlier, which is the focus of this study. Normality tests for residues used in general nonlinear regression revealed that the usage of the pseudo-1st order model in the fitting of glyphosate biosorption to palm oil fronds was adequate due to lack of an outlier. The Grubb’s test was applied to the residual results. The critical value of Z from statistical table for Grubbs’ test for a single outlier using mean and SD was 2.289 (n=10). The Grubbs (Alpha = 0.05) g value was 1.844. As the test statistic is lesser than the critical assigned value, no outlier was deemed present.