Detecting Difference Between Coefficients in Linear Model Using Jackknife Empirical Likelihood

Empirical likelihood has been found very useful in many different occasions. It usually runs into serious computational difficulties while jackknife empirical likelihood (JEL) is shown to be effective when applied to some complicated statistics. In this paper, to test the difference between coefficients of two linear regression models, the authors apply JEL to construct the confidence regions. Based on the JEL ratio test, a version of Wilks’ theorem is developed. Furthermore, to improve the coverage accuracy of confidence regions, a Bartlett correction is applied. Simulation studies are carried out to show the effectiveness of the proposed method in aspects of coverage accuracy. A real data set is analyzed with the proposed method as an example.