Marginal estimation under a general regression model with missing responses and covariates

BIANCO, ANA; BOENTE, GRACIELA; GONZALEZ-MANTEIGA, WENCESLAO; PEREZ-GONZALEZ, ANA

Wollongong

Conferencia; International Conference on Robust Statistics (ICORS 2017); 2017

University of Wollongong, NIASRA, AMSI

We consider a general regression model where missing data occur, in the responses and in the covariates. Our aim is to discuss estimation procedures for any marginal functional of the responses such as the median or any alpha-quantile of the response variable. A missing at random condition is assumed in order to prevent from bias in the estimation of the marginal measures undera non-ignorable missing mechanism. Different approaches for the estimation of the marginal functional of interest are presented, including inverse probability weighting, the convolution of the estimators of the distributions of the errors and of the regression function and also double robust estimators, which protect against misspecication of the regression or the missing probability models. The proposals may be combined with robust procedures for the estimation of theregression function. The small sample behaviour of the proposals is illustrated through a Monte Carlo study on a partially linear model.