Interval estimation for the difference of two correlated gamma means: a generalized inference method and hybrid methods

Gamma distribution plays an important role in applied fields due to its flexibility of accommodating right-skewed data. Although inference methods for single gamma mean and for difference of two independent gamma means have been well studied, inference methods for difference of two correlated gamma means are sparse. This paper considers the problem of interval estimation for the difference of two correlated gamma means. We propose several inference methods including a method based on the concepts of generalized inference and two hybrid methods based on the method of variance estimates recovery. An extensive Monte Carlo simulation study was conducted to assess the performance of the proposed methods in terms of coverage probabilities and average lengths of the estimated intervals. Two real data examples from medical and engineering studies are analyzed using the proposed methods.