Nonparametric analysis of factorial designs with random missingness

Bivariate data

Michael G. Akritas, Efi S. Antoniou, Jouni Kuha

Research output: Contribution to journalArticle

4 Citations (Scopus)

Abstract

We propose a nonparametric approach to the analysis of factorial designs where each subject is observed at two time points and both observations are subject to missingness. The procedures are fully nonparametric in that they do not require continuity, and do not impose models to describe the relation of the response distribution in different factor-level combinations. The approach for estimating and testing treatment and time effects is based on a method, which we introduce, for estimating a distribution function. The method requires a pattern-mixture-type assumption on the missingness mechanism, which is weaker than the missing-completely-at-random assumption but neither weaker nor stronger than the missing-at-random assumption. This missingness assumption is the minimal requirement for nonparametric analysis. Comparisons with normal-based likelihood ratio tests indicate that the proposed tests fare well when the data are normal and homoscedastic, and outperform them in many other cases. Simulations also confirm that the proposed method has higher power than common nonparametric complete-pairs tests for observations missing completely at random. Finally, a dataset on the delinquent values of boys released from correctional institutions is analyzed and discussed.

Original languageEnglish (US)
Pages (from-to)1513-1526
Number of pages14
JournalJournal of the American Statistical Association
Volume101
Issue number476
DOIs
StatePublished - Dec 1 2006

Fingerprint

Factorial Design
Missing Completely at Random
Missing at Random
Likelihood Ratio Test
High Power
Distribution Function
Testing
Nonparametric analysis
Factorial design
Requirements
Simulation
Observation
Model

All Science Journal Classification (ASJC) codes

  • Mathematics(all)
  • Statistics and Probability

Cite this

@article{96c5f8f9251742d28c3f03ecdb861651,
title = "Nonparametric analysis of factorial designs with random missingness: Bivariate data",
abstract = "We propose a nonparametric approach to the analysis of factorial designs where each subject is observed at two time points and both observations are subject to missingness. The procedures are fully nonparametric in that they do not require continuity, and do not impose models to describe the relation of the response distribution in different factor-level combinations. The approach for estimating and testing treatment and time effects is based on a method, which we introduce, for estimating a distribution function. The method requires a pattern-mixture-type assumption on the missingness mechanism, which is weaker than the missing-completely-at-random assumption but neither weaker nor stronger than the missing-at-random assumption. This missingness assumption is the minimal requirement for nonparametric analysis. Comparisons with normal-based likelihood ratio tests indicate that the proposed tests fare well when the data are normal and homoscedastic, and outperform them in many other cases. Simulations also confirm that the proposed method has higher power than common nonparametric complete-pairs tests for observations missing completely at random. Finally, a dataset on the delinquent values of boys released from correctional institutions is analyzed and discussed.",
author = "Akritas, {Michael G.} and Antoniou, {Efi S.} and Jouni Kuha",
year = "2006",
month = "12",
day = "1",
doi = "10.1198/016214506000000537",
language = "English (US)",
volume = "101",
pages = "1513--1526",
journal = "Journal of the American Statistical Association",
issn = "0162-1459",
publisher = "Taylor and Francis Ltd.",
number = "476",

}

Nonparametric analysis of factorial designs with random missingness : Bivariate data. / Akritas, Michael G.; Antoniou, Efi S.; Kuha, Jouni.

In: Journal of the American Statistical Association, Vol. 101, No. 476, 01.12.2006, p. 1513-1526.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Nonparametric analysis of factorial designs with random missingness

T2 - Bivariate data

AU - Akritas, Michael G.

AU - Antoniou, Efi S.

AU - Kuha, Jouni

PY - 2006/12/1

Y1 - 2006/12/1

N2 - We propose a nonparametric approach to the analysis of factorial designs where each subject is observed at two time points and both observations are subject to missingness. The procedures are fully nonparametric in that they do not require continuity, and do not impose models to describe the relation of the response distribution in different factor-level combinations. The approach for estimating and testing treatment and time effects is based on a method, which we introduce, for estimating a distribution function. The method requires a pattern-mixture-type assumption on the missingness mechanism, which is weaker than the missing-completely-at-random assumption but neither weaker nor stronger than the missing-at-random assumption. This missingness assumption is the minimal requirement for nonparametric analysis. Comparisons with normal-based likelihood ratio tests indicate that the proposed tests fare well when the data are normal and homoscedastic, and outperform them in many other cases. Simulations also confirm that the proposed method has higher power than common nonparametric complete-pairs tests for observations missing completely at random. Finally, a dataset on the delinquent values of boys released from correctional institutions is analyzed and discussed.

AB - We propose a nonparametric approach to the analysis of factorial designs where each subject is observed at two time points and both observations are subject to missingness. The procedures are fully nonparametric in that they do not require continuity, and do not impose models to describe the relation of the response distribution in different factor-level combinations. The approach for estimating and testing treatment and time effects is based on a method, which we introduce, for estimating a distribution function. The method requires a pattern-mixture-type assumption on the missingness mechanism, which is weaker than the missing-completely-at-random assumption but neither weaker nor stronger than the missing-at-random assumption. This missingness assumption is the minimal requirement for nonparametric analysis. Comparisons with normal-based likelihood ratio tests indicate that the proposed tests fare well when the data are normal and homoscedastic, and outperform them in many other cases. Simulations also confirm that the proposed method has higher power than common nonparametric complete-pairs tests for observations missing completely at random. Finally, a dataset on the delinquent values of boys released from correctional institutions is analyzed and discussed.

UR - http://www.scopus.com/inward/record.url?scp=33846036997&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33846036997&partnerID=8YFLogxK

U2 - 10.1198/016214506000000537

DO - 10.1198/016214506000000537

M3 - Article

VL - 101

SP - 1513

EP - 1526

JO - Journal of the American Statistical Association

JF - Journal of the American Statistical Association

SN - 0162-1459

IS - 476

ER -