Ignorability in general incomplete-data models

Research output: Contribution to journalArticle

55 Citations (Scopus)

Abstract

SUMMARY: Rubin (1976) defined ignorability conditions for frequentist and Bayes/likelihood analyses of data subject to missing observations. More recently, Heitjan & Rubin (1991) and Heitjan (1993) generalised the Rubin model to encompass other forms of incompleteness, establishing ignorability conditions for Bayes/likelihood inferences only. This paper extends the Heitjan-Rubin model by explicitly defining the observed degree of coarseness as a data element. This permits the development of a frequentist theory, including a generalisation of 'missing completely at random', the frequentist ignorability condition for missing data. The model is applied in a number of incomplete-data problems of general interest.

Original languageEnglish (US)
Pages (from-to)701-708
Number of pages8
JournalBiometrika
Volume81
Issue number4
DOIs
StatePublished - Dec 1 1994

Fingerprint

Ignorability
Incomplete Data
Data Model
Data structures
Bayes
Missing Completely at Random
development theory
Missing Observations
Likelihood Inference
Incompleteness
Missing Data
Likelihood
Model
Incomplete data

Keywords

  • Coarsened at random
  • Coarsened completely at random
  • Missing at random
  • Missing completely at random
  • Observed at random

ASJC Scopus subject areas

  • Statistics, Probability and Uncertainty
  • Applied Mathematics
  • Mathematics(all)
  • Statistics and Probability
  • Agricultural and Biological Sciences (miscellaneous)
  • Agricultural and Biological Sciences(all)

Cite this

Ignorability in general incomplete-data models. / Heitjan, Daniel F.

In: Biometrika, Vol. 81, No. 4, 01.12.1994, p. 701-708.

Research output: Contribution to journalArticle

@article{c96bbf331fe34821a1e94d53b541c58b,
title = "Ignorability in general incomplete-data models",
abstract = "SUMMARY: Rubin (1976) defined ignorability conditions for frequentist and Bayes/likelihood analyses of data subject to missing observations. More recently, Heitjan & Rubin (1991) and Heitjan (1993) generalised the Rubin model to encompass other forms of incompleteness, establishing ignorability conditions for Bayes/likelihood inferences only. This paper extends the Heitjan-Rubin model by explicitly defining the observed degree of coarseness as a data element. This permits the development of a frequentist theory, including a generalisation of 'missing completely at random', the frequentist ignorability condition for missing data. The model is applied in a number of incomplete-data problems of general interest.",
keywords = "Coarsened at random, Coarsened completely at random, Missing at random, Missing completely at random, Observed at random",
author = "Heitjan, {Daniel F.}",
year = "1994",
month = "12",
day = "1",
doi = "10.1093/biomet/81.4.701",
language = "English (US)",
volume = "81",
pages = "701--708",
journal = "Biometrika",
issn = "0006-3444",
publisher = "Oxford University Press",
number = "4",

}

TY - JOUR

T1 - Ignorability in general incomplete-data models

AU - Heitjan, Daniel F.

PY - 1994/12/1

Y1 - 1994/12/1

N2 - SUMMARY: Rubin (1976) defined ignorability conditions for frequentist and Bayes/likelihood analyses of data subject to missing observations. More recently, Heitjan & Rubin (1991) and Heitjan (1993) generalised the Rubin model to encompass other forms of incompleteness, establishing ignorability conditions for Bayes/likelihood inferences only. This paper extends the Heitjan-Rubin model by explicitly defining the observed degree of coarseness as a data element. This permits the development of a frequentist theory, including a generalisation of 'missing completely at random', the frequentist ignorability condition for missing data. The model is applied in a number of incomplete-data problems of general interest.

AB - SUMMARY: Rubin (1976) defined ignorability conditions for frequentist and Bayes/likelihood analyses of data subject to missing observations. More recently, Heitjan & Rubin (1991) and Heitjan (1993) generalised the Rubin model to encompass other forms of incompleteness, establishing ignorability conditions for Bayes/likelihood inferences only. This paper extends the Heitjan-Rubin model by explicitly defining the observed degree of coarseness as a data element. This permits the development of a frequentist theory, including a generalisation of 'missing completely at random', the frequentist ignorability condition for missing data. The model is applied in a number of incomplete-data problems of general interest.

KW - Coarsened at random

KW - Coarsened completely at random

KW - Missing at random

KW - Missing completely at random

KW - Observed at random

UR - http://www.scopus.com/inward/record.url?scp=0000579766&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0000579766&partnerID=8YFLogxK

U2 - 10.1093/biomet/81.4.701

DO - 10.1093/biomet/81.4.701

M3 - Article

AN - SCOPUS:0000579766

VL - 81

SP - 701

EP - 708

JO - Biometrika

JF - Biometrika

SN - 0006-3444

IS - 4

ER -