A High-Breakdown MCD-Based Robust Concordance Correlation Coefficient

The concordance correlation coefficient (CCC) is a popular measure of agreement between two continuous variables but is highly sensitive to outliers and data contamination. In this study, we propose a robust reformulation of the CCC by replacing classical moment estimators with Minimum Covariance Determinant (MCD) estimators. The proposed robust CCC preserves the interpretability of the classical coefficient while providing substantially improved robustness. Comprehensive Monte Carlo simulations under normal and non-normal distributions, varying sample sizes, correlation levels, and contamination schemes compare the proposed coefficient with the classical CCC and existing robust alternatives. The results show that the proposed robust CCC achieves superior stability and accuracy in contaminated settings while remaining competitive under clean data. Theoretical properties of the estimator are discussed, and its practical usefulness is demonstrated using real glucose measurement and blood pressure data sets. The proposed method is implemented in the MVTests R package, enabling straightforward application to real-world data.