Empirical likelihood for the response mean of generalized linear models with missing at random responses

In this article, we present a new empirical likelihood ratio for constructing the confidence interval of the response mean of generalized linear models with missing at random responses. Compared with the existing methods, the proposal can avoid the so-called curse of dimensionality problem when the dimension of covariates is high, and is still chi-squared distributed asymptotically, nevertheless. Simulation studies are also provided to illustrate the performance of the developed method.