Opposition-based differential evolution (ODE) is a recently proposed DE variant, which has shown faster convergence speed and more robust search abilities than classical DE. The concept of opposition was utilized for the first time in optimization area to propose ODE. It is based on two important steps, generation jumping and elite selection. Some studies have pointed out that the first step improves diversity and provides more potential points to be searched (diversification), while the second step decreases diversity and accelerates convergence speed (intensification). However, there is not any experimental study to support this explanation. In this paper, we present an experimental study to analyze how the diversity changes in ODE. The experimental results confirm the explanation, and show that ODE makes a good balance between generation jumping and elite selection.