Empirical-likelihood-inference-based Estimating Equations for Complex Surveys with Data Missing at Random 

Title: Empirical-likelihood-inference-based Estimating Equations for Complex Surveys with Data Missing at Random 

Speaker: Song Cai (Carleton U.)

Time: 1:30-2:30 pm, Mar 24, 2017

Room: KED B015

Abstract: We propose an empirical likelihood (EL) method for constructing confidence intervals on population parameters defined by an estimation equation (EE) with data from complex surveys that are missing at random. Instead of imputing the original data, we impute the EE using a fractional random imputation with fixed number of draws. We then construct an EL ratio based on the imputed EE and show that this ratio has a chi-bar-square limiting distribution in general. We also show that, in the special case when no missing data are present, the proposed EL ratio has a simple chi-square limiting distribution under probability proportional to size sampling with replacement. Moreover, we propose a proper bootstrap procedure to approximate the limiting distribution of the EL ratio for constructing confidence intervals on the parameters of interest.