Robust inference in generalized linear models

BIANCO, ANA; BOENTE, GRACIELA; RODRIGUES, ISABEL

Paris

Conferencia; 19th International Conference on Computational Statistics (COMPSTAT 2010); 2010

In this work, we introduce robust procedure to estimate the regression parameter under a glm model in order to build test statistics for this parameter when missing data occur in the responses. We derive the asymptotic behaviour of the robust estimators for the regression parameter under the null hypothesis and under contiguous alternatives in order to study the robust Wald test. The influence function of the test functional is also studied. The finite sample properties of the proposed procedure are investigated through a Monte Carlo study where the robust test is also compared with nonrobust and robust alternatives.under a glm model in order to build test statistics for this parameter when missing data occur in the responses. We derive the asymptotic behaviour of the robust estimators for the regression parameter under the null hypothesis and under contiguous alternatives in order to study the robust Wald test. The influence function of the test functional is also studied. The finite sample properties of the proposed procedure are investigated through a Monte Carlo study where the robust test is also compared with nonrobust and robust alternatives.