Semiparametric Estimation of Missingness Mechanism for Missing Not at Random Data

• Activity Number: 361 - Contributed Poster Presentations: Section on Nonparametric Statistics
• Type: Contributed
• Date/Time: Wednesday, August 5, 2020 : 10:00 AM to 2:00 PM
• Sponsor: Section on Nonparametric Statistics
• Author(s): Samidha Sudhakar Shetty* and Yanyuan Ma and Jiwei Zhao
• Companies: Pennsylvania State University and The Pennsylvania State University and State University of New York at Buffalo
• Keywords: Missing data; Missing not at random; Semiparametric estimation; Efficient score

Abstract

Missing data is rampant in data sets in every field of science. In the past few decades, there has been interest in understanding the underlying pattern of missingness, formally known as the missingness mechanism. There are three main types of missingness mechanisms: Missing Completely at Random (MCAR), Missing at Random (MAR) and Missing Not at Random (MNAR). Most likelihood or imputation-based methods developed assume the MAR or the MCAR condition, which are the more well studied conditions. We discuss the third condition which is less well studied. MNAR is the hardest to deal with but also the most likely to occur. Under the MNAR assumption, the missing response depends on a set of covariates and the value of the response itself. We model the missingness mechanism using a semiparametric approach. The resulting estimator is consistent and further under certain conditions is proved to be efficient.