On the identifiability and estimation of generalized linear models with parametric nonignorable missing data mechanism

Xia Cui, Jianhua Guo, Guangren Yang

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

We address the problem of identifying and estimating generalized linear models when the response variable is nonignorably missing. Three types of monotone missing data mechanism are assumed, including Logit model, Probit model and complementary Log–log model. In this situation, likelihood based on observed data may not be identifiable. In this article, we prove the model parameters are identifiable under very mild conditions and then construct estimators based on a likelihood-based approach. The proposed estimators are shown to be consistent and asymptotically normal. Simulation studies demonstrate that the proposed inference procedure performs well in many settings. We apply the proposed method to a data set from research in Chinese Household Income Project study.

Original languageEnglish (US)
Pages (from-to)64-80
Number of pages17
JournalComputational Statistics and Data Analysis
Volume107
DOIs
StatePublished - Mar 1 2017

Fingerprint

Nonignorable Missing Data
Missing Data Mechanism
Identifiability
Generalized Linear Model
Likelihood
Estimator
Probit Model
Logit Model
Monotone
Simulation Study
Model
Demonstrate

All Science Journal Classification (ASJC) codes

  • Statistics and Probability
  • Computational Mathematics
  • Computational Theory and Mathematics
  • Applied Mathematics

Cite this

@article{4b554e832a3148f38273c77f7d0dbdb1,
title = "On the identifiability and estimation of generalized linear models with parametric nonignorable missing data mechanism",
abstract = "We address the problem of identifying and estimating generalized linear models when the response variable is nonignorably missing. Three types of monotone missing data mechanism are assumed, including Logit model, Probit model and complementary Log–log model. In this situation, likelihood based on observed data may not be identifiable. In this article, we prove the model parameters are identifiable under very mild conditions and then construct estimators based on a likelihood-based approach. The proposed estimators are shown to be consistent and asymptotically normal. Simulation studies demonstrate that the proposed inference procedure performs well in many settings. We apply the proposed method to a data set from research in Chinese Household Income Project study.",
author = "Xia Cui and Jianhua Guo and Guangren Yang",
year = "2017",
month = "3",
day = "1",
doi = "10.1016/j.csda.2016.10.017",
language = "English (US)",
volume = "107",
pages = "64--80",
journal = "Computational Statistics and Data Analysis",
issn = "0167-9473",
publisher = "Elsevier",

}

On the identifiability and estimation of generalized linear models with parametric nonignorable missing data mechanism. / Cui, Xia; Guo, Jianhua; Yang, Guangren.

In: Computational Statistics and Data Analysis, Vol. 107, 01.03.2017, p. 64-80.

Research output: Contribution to journalArticle

TY - JOUR

T1 - On the identifiability and estimation of generalized linear models with parametric nonignorable missing data mechanism

AU - Cui, Xia

AU - Guo, Jianhua

AU - Yang, Guangren

PY - 2017/3/1

Y1 - 2017/3/1

N2 - We address the problem of identifying and estimating generalized linear models when the response variable is nonignorably missing. Three types of monotone missing data mechanism are assumed, including Logit model, Probit model and complementary Log–log model. In this situation, likelihood based on observed data may not be identifiable. In this article, we prove the model parameters are identifiable under very mild conditions and then construct estimators based on a likelihood-based approach. The proposed estimators are shown to be consistent and asymptotically normal. Simulation studies demonstrate that the proposed inference procedure performs well in many settings. We apply the proposed method to a data set from research in Chinese Household Income Project study.

AB - We address the problem of identifying and estimating generalized linear models when the response variable is nonignorably missing. Three types of monotone missing data mechanism are assumed, including Logit model, Probit model and complementary Log–log model. In this situation, likelihood based on observed data may not be identifiable. In this article, we prove the model parameters are identifiable under very mild conditions and then construct estimators based on a likelihood-based approach. The proposed estimators are shown to be consistent and asymptotically normal. Simulation studies demonstrate that the proposed inference procedure performs well in many settings. We apply the proposed method to a data set from research in Chinese Household Income Project study.

UR - http://www.scopus.com/inward/record.url?scp=84995394456&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84995394456&partnerID=8YFLogxK

U2 - 10.1016/j.csda.2016.10.017

DO - 10.1016/j.csda.2016.10.017

M3 - Article

AN - SCOPUS:84995394456

VL - 107

SP - 64

EP - 80

JO - Computational Statistics and Data Analysis

JF - Computational Statistics and Data Analysis

SN - 0167-9473

ER -