Modeling nonignorable missing data processes in item calibration

Cornelis A.W. Glas, Jonald Pimentel

Research output: Book/ReportReportOther research output

Abstract

In this report, it is shown that the problem of nonignorable missing data in the calibration phase for computerized adaptive testing can be handled by introducing an item response theory (IRT) model for the missing data indicator. In the first simulation study, it is shown that treating data with nonignorable missing data as if ignorability held, does indeed lead to a significant increase in the estimation errors, and introducing an IRT model for the missing data process solves the problem to a large extent. In the second simulation study, it is shown that the correction obtained using the IRT model is almost as good as the correction that can be obtained if the covariates determining the missing data were actually observed.
Original languageUndefined
Place of PublicationNewton, PA, USA
PublisherLaw School Admission Council
Number of pages7
Publication statusPublished - Jan 2006

Publication series

NameLSAC research report series
PublisherLaw School Admission Council
No.04-07

Keywords

  • IR-104265

Cite this

Glas, C. A. W., & Pimentel, J. (2006). Modeling nonignorable missing data processes in item calibration. (LSAC research report series; No. 04-07). Newton, PA, USA: Law School Admission Council.
Glas, Cornelis A.W. ; Pimentel, Jonald. / Modeling nonignorable missing data processes in item calibration. Newton, PA, USA : Law School Admission Council, 2006. 7 p. (LSAC research report series; 04-07).
@book{8e1955d53ad74a0f8978e4974e5b3baf,
title = "Modeling nonignorable missing data processes in item calibration",
abstract = "In this report, it is shown that the problem of nonignorable missing data in the calibration phase for computerized adaptive testing can be handled by introducing an item response theory (IRT) model for the missing data indicator. In the first simulation study, it is shown that treating data with nonignorable missing data as if ignorability held, does indeed lead to a significant increase in the estimation errors, and introducing an IRT model for the missing data process solves the problem to a large extent. In the second simulation study, it is shown that the correction obtained using the IRT model is almost as good as the correction that can be obtained if the covariates determining the missing data were actually observed.",
keywords = "IR-104265",
author = "Glas, {Cornelis A.W.} and Jonald Pimentel",
year = "2006",
month = "1",
language = "Undefined",
series = "LSAC research report series",
publisher = "Law School Admission Council",
number = "04-07",

}

Glas, CAW & Pimentel, J 2006, Modeling nonignorable missing data processes in item calibration. LSAC research report series, no. 04-07, Law School Admission Council, Newton, PA, USA.

Modeling nonignorable missing data processes in item calibration. / Glas, Cornelis A.W.; Pimentel, Jonald.

Newton, PA, USA : Law School Admission Council, 2006. 7 p. (LSAC research report series; No. 04-07).

Research output: Book/ReportReportOther research output

TY - BOOK

T1 - Modeling nonignorable missing data processes in item calibration

AU - Glas, Cornelis A.W.

AU - Pimentel, Jonald

PY - 2006/1

Y1 - 2006/1

N2 - In this report, it is shown that the problem of nonignorable missing data in the calibration phase for computerized adaptive testing can be handled by introducing an item response theory (IRT) model for the missing data indicator. In the first simulation study, it is shown that treating data with nonignorable missing data as if ignorability held, does indeed lead to a significant increase in the estimation errors, and introducing an IRT model for the missing data process solves the problem to a large extent. In the second simulation study, it is shown that the correction obtained using the IRT model is almost as good as the correction that can be obtained if the covariates determining the missing data were actually observed.

AB - In this report, it is shown that the problem of nonignorable missing data in the calibration phase for computerized adaptive testing can be handled by introducing an item response theory (IRT) model for the missing data indicator. In the first simulation study, it is shown that treating data with nonignorable missing data as if ignorability held, does indeed lead to a significant increase in the estimation errors, and introducing an IRT model for the missing data process solves the problem to a large extent. In the second simulation study, it is shown that the correction obtained using the IRT model is almost as good as the correction that can be obtained if the covariates determining the missing data were actually observed.

KW - IR-104265

M3 - Report

T3 - LSAC research report series

BT - Modeling nonignorable missing data processes in item calibration

PB - Law School Admission Council

CY - Newton, PA, USA

ER -

Glas CAW, Pimentel J. Modeling nonignorable missing data processes in item calibration. Newton, PA, USA: Law School Admission Council, 2006. 7 p. (LSAC research report series; 04-07).