Abstract:Missing data occur frequently in empirical studies in health and social sciences, often compromising our ability to make accurate inferences. An outcome is said to be missing not at random (MNAR) if, conditional on the observed variables, the missing data mechanism still depends on the unobserved outcome. In such settings, identification is generally not possible without imposing additional assumptions. Identification is sometimes possible, however, if an instrumental variable (IV) is observed for all subjects which satisfies the exclusion restriction that the IV affects the missingness process without directly influencing the outcome. In this paper, we provide necessary and sufficient conditions for nonparametric identification of the full data distribution under MNAR with the aid of an IV. In addition, we give sufficient identification conditions that are more straightforward to verify in practice. For inference, we focus on estimation of a population outcome mean, for which we develop a suite of semiparametric estimators that extend methods previously developed for data missing at random. Specifically, we propose inverse probability weighted estimation, outcome regression-based estimation and doubly robust estimation of the mean of an outcome subject to MNAR. For illustration, the methods are used to account for selection bias induced by HIV testing refusal in the evaluation of HIV seroprevalence in Mochudi, Botswana, using interviewer characteristics such as gender, age and years of experience as IVs.

Estimation in Semiparametric Models with Missing Data

Nonparametric Imputation of Missing Values for Estimating Equation Based Inference

Semiparametric Inference for Estimating Equations with Nonignorably Missing Covariates

Improving Semiparametric Estimation by Using Surrogate Data

Empirical Likelihood for Estimating Equations with Missing Values

Identification And Inference With Nonignorable Missing Covariate Data

Full-semiparametric-likelihood-based inference for non-ignorable missing data

State Estimation with Missing Measurements Using IMM

Semiparametric fractional imputation using Gaussian mixture models for handling multivariate missing data

Semiparametric Estimation with Data Missing Not at Random Using an Instrumental Variable.

Nonparametric regression with discrete covariate and missing values

Parameter Estimation in Spatial Autoregressive Models with Missing Data and Measurement Errors

On a simultaneous parameter inference and missing data imputation for nonstationary autoregressive models

Robust estimation of the generalised partial linear model with missing covariates

An Investigation of Methods for Handling Missing Data with Penalized Regression

Inverse Probability-Weighted Estimation for Dynamic Structural Equation Model with Missing Data

Maximum likelihood estimation of missing data probability for nonmonotone missing at random data

Double robust estimation of functional outcomes with data missing at random

Imputations for High Missing Rate Data in Covariates Via Semi-supervised Learning Approach

Double robust semiparametric weighted M-estimators of a zero-inflated Poisson regression with missing data in covariates

Imputation-based empirical likelihood inferences for partially nonlinear quantile regression models with missing responses