The two-sample Mood statistic for clustered data

Abstract The classical Mood statistic is frequently used to test the equality of scale parameters in two-sample problems. However, its application requires the assumption of independent observations. In many practical situations, correlated observations often arise due to clustered data. To address this issue, we develop a test for clustered data under the assumption of homoscedasticity. This study derives the asymptotic distribution of the proposed statistic under the null hypothesis. Simulations are performed to investigate the robustness and power performance of the proposed statistics in various situations. The simulation results demonstrate that the proposed statistic is a strong competitor to existing statistics and offers advantages in certain cases. Two illustrations using oribatid mite data and BMI data are provided. Finally, the concluding remarks are presented.