A Goodness-of-Fit Test for the Additive Risk Model with a Binary Covariate

In this article, we propose a class of weighted estimators for the excess risk in additive risk model with a binary covariate. The proposed estimator is consistent and asymptotically normal. When the assumed model is inappropriate, however, the estimators with different weights converge to nonidentical constants. This fact enables us to develop a goodness-of-fit test for the excess assumption by comparing estimators with diffrent weights. It is shown that the proposed test converges in distribution to normal with mean zero and is consistent under the model misspecifications. Furthermore, the finite-sample properties of the proposed test procedure are investigated and two examples using real data are presented.