Performance of missing data approaches under nonignorable missing data conditions
Author(s) / Creator(s)
Pohl, Steffi
Becker, Benjamin
Abstract / Description
Approaches for dealing with item omission include incorrect scoring, ignoring missing values, and approaches for nonignorable missing values and have only been evaluated for certain forms of nonignorability. In this paper we investigate the performance of these approaches for various conditions of nonignorability, that is, when the missing response depends on i) the item response, ii) a latent missing propensity, or iii) both. No approach results in unbiased parameter estimates of the Rasch model under all missing data mechanisms. Incorrect scoring only results in unbiased estimates under very specific data constellations of missing mechanisms i) and iii). The approach for nonignorable missing values only results in unbiased estimates under condition ii). Ignoring results in slightly more biased estimates than the approach for nonignorable missing values, while the latter also indicates the presence of nonignorablity under all simulated conditions. We illustrate the results in an empirical example on PISA data.
Keyword(s)
missing data nonignorability item response theory item nonresponse large-scale assessmentPersistent Identifier
Date of first publication
2020-06-18
Journal title
Methodology
Volume
16
Issue
2
Page numbers
147–165
Publisher
PsychOpen GOLD
Publication status
publishedVersion
Review status
peerReviewed
Is version of
Citation
Pohl, S., & Becker, B. (2020). Performance of missing data approaches under nonignorable missing data conditions. Methodology, 16(2), 147-165. https://doi.org/10.5964/meth.2805
-
meth.v16i2.2805.pdfAdobe PDF - 331.97KBMD5: 32a75701d4b6a8967f9c98d83dfa9d82
-
There are no other versions of this object.
-
Author(s) / Creator(s)Pohl, Steffi
-
Author(s) / Creator(s)Becker, Benjamin
-
PsychArchives acquisition timestamp2022-04-14T11:24:36Z
-
Made available on2022-04-14T11:24:36Z
-
Date of first publication2020-06-18
-
Abstract / DescriptionApproaches for dealing with item omission include incorrect scoring, ignoring missing values, and approaches for nonignorable missing values and have only been evaluated for certain forms of nonignorability. In this paper we investigate the performance of these approaches for various conditions of nonignorability, that is, when the missing response depends on i) the item response, ii) a latent missing propensity, or iii) both. No approach results in unbiased parameter estimates of the Rasch model under all missing data mechanisms. Incorrect scoring only results in unbiased estimates under very specific data constellations of missing mechanisms i) and iii). The approach for nonignorable missing values only results in unbiased estimates under condition ii). Ignoring results in slightly more biased estimates than the approach for nonignorable missing values, while the latter also indicates the presence of nonignorablity under all simulated conditions. We illustrate the results in an empirical example on PISA data.en_US
-
Publication statuspublishedVersion
-
Review statuspeerReviewed
-
CitationPohl, S., & Becker, B. (2020). Performance of missing data approaches under nonignorable missing data conditions. Methodology, 16(2), 147-165. https://doi.org/10.5964/meth.2805en_US
-
ISSN1614-2241
-
Persistent Identifierhttps://hdl.handle.net/20.500.12034/5688
-
Persistent Identifierhttps://doi.org/10.23668/psycharchives.6292
-
Language of contenteng
-
PublisherPsychOpen GOLD
-
Is version ofhttps://doi.org/10.5964/meth.2805
-
Is related tohttps://doi.org/10.23668/psycharchives.2905
-
Is related tohttps://www.oecd.org/pisa/data/pisa2012database-downloadabledata.htm
-
Keyword(s)missing dataen_US
-
Keyword(s)nonignorabilityen_US
-
Keyword(s)item response theoryen_US
-
Keyword(s)item nonresponseen_US
-
Keyword(s)large-scale assessmenten_US
-
Dewey Decimal Classification number(s)150
-
TitlePerformance of missing data approaches under nonignorable missing data conditionsen_US
-
DRO typearticle
-
Issue2
-
Journal titleMethodology
-
Page numbers147–165
-
Volume16
-
Visible tag(s)Version of Recorden_US