On the Role of Diversity Measures for Multi-Objective Test Case Selection