¹  While there exist conditions for \(T_{ml}\) to follow a chi-square distribution when data are not normally distributed, there is not an effective way to verify the so-called asymptotic robustness conditions in practice.