Application of range-test in multiple linear regression analysis in presence of outliers
Application of range-test in multiple linear regression analysis in the presence of outliers is studied in this paper. First, the plot of the explanatory variables (i.e. Administration, Social/Commercial, Economic services and Transfer) on the dependent variable (i.e. GDP) was done to identify the statistical trend over the years. The identified trend is linear and positive upward. Secondly, a multiple linear regression model is constructed to describe the relationship between the dependent variable and independent variables. Thirdly, it is shown how outliers could be handled using the Range Test, because outliers can have deleterious effects on statistical analyses. When researchers ignore such abnormal observations, especially with respect to dependent variables, the empirical results can be misleading. From our findings, we conclude that treating outliers from regression models give better fit of the model in terms of R-square.