A comparative study of rank tests for regression

Five different nonparametric approaches, based on ranks, for testing hypothesis in a general linear regression model will be considered, and comparisons will be made between the performances of each of these five statistics, under different hypotheses and conditions. Comparisons will also be made between these nonparametric approaches and the classical parametric F test in order to determine any possible significant differences and similarities between these theoretically different methods. Some versions of these comparisons for some of these nonparametric approaches have been made in Hettmansperger and McKean[l], but this new study considers cases not studied in their study and considers some new ideas that have not yet been presented.