Multiple imputation to balance unbalanced designs for two-way analysis of variance
Author(s) / Creator(s)
van Ginkel, Joost R.
Kroonenberg, Pieter M.
Abstract / Description
A balanced ANOVA design provides an unambiguous interpretation of the F-tests, and has more power than an unbalanced design. In earlier literature, multiple imputation was proposed to create balance in unbalanced designs, as an alternative to Type-III sum of squares. In the current simulation study we studied four pooled statistics for multiple imputation, namely D₀, D₁, D₂, and D₃ in unbalanced data, and compared them with Type-III sum of squares. Statistics D₁ and D₂ generally performed best regarding Type-I error rates, and had power rates closest to that of Type-III sum of squares. Additionally, for the interaction, D₁ produced power rates higher than Type-III sum of squares. For multiply imputed datasets D₁ and D₂ may be the best methods for pooling the results in multiply imputed datasets, and for unbalanced data, D₁ might be a good alternative to Type-III sum of squares regarding the interaction.
Keyword(s)
unbalanced designs multiple imputation two-way analysis of variance missing data type-III sum of squaresPersistent Identifier
Date of first publication
2021-03-31
Journal title
Methodology
Volume
17
Issue
1
Page numbers
39–57
Publisher
PsychOpen GOLD
Publication status
publishedVersion
Review status
peerReviewed
Is version of
Citation
van Ginkel, J. R., & Kroonenberg, P. M. (2021). Multiple imputation to balance unbalanced designs for two-way analysis of variance. Methodology, 17(1), 39-57. https://doi.org/10.5964/meth.6085
-
Meth.v17i1.6085.pdfAdobe PDF - 318.03KBMD5: 67384ce7b45fe0cf4756b11c97025a08
-
There are no other versions of this object.
-
Author(s) / Creator(s)van Ginkel, Joost R.
-
Author(s) / Creator(s)Kroonenberg, Pieter M.
-
PsychArchives acquisition timestamp2022-04-14T11:19:23Z
-
Made available on2022-04-14T11:19:23Z
-
Date of first publication2021-03-31
-
Abstract / DescriptionA balanced ANOVA design provides an unambiguous interpretation of the F-tests, and has more power than an unbalanced design. In earlier literature, multiple imputation was proposed to create balance in unbalanced designs, as an alternative to Type-III sum of squares. In the current simulation study we studied four pooled statistics for multiple imputation, namely D₀, D₁, D₂, and D₃ in unbalanced data, and compared them with Type-III sum of squares. Statistics D₁ and D₂ generally performed best regarding Type-I error rates, and had power rates closest to that of Type-III sum of squares. Additionally, for the interaction, D₁ produced power rates higher than Type-III sum of squares. For multiply imputed datasets D₁ and D₂ may be the best methods for pooling the results in multiply imputed datasets, and for unbalanced data, D₁ might be a good alternative to Type-III sum of squares regarding the interaction.en_US
-
Publication statuspublishedVersion
-
Review statuspeerReviewed
-
Citationvan Ginkel, J. R., & Kroonenberg, P. M. (2021). Multiple imputation to balance unbalanced designs for two-way analysis of variance. Methodology, 17(1), 39-57. https://doi.org/10.5964/meth.6085en_US
-
ISSN1614-2241
-
Persistent Identifierhttps://hdl.handle.net/20.500.12034/5100
-
Persistent Identifierhttps://doi.org/10.23668/psycharchives.5704
-
Language of contenteng
-
PublisherPsychOpen GOLD
-
Is version ofhttps://doi.org/10.5964/meth.6085
-
Is related tohttps://doi.org/10.23668/psycharchives.4714
-
Keyword(s)unbalanced designsen_US
-
Keyword(s)multiple imputationen_US
-
Keyword(s)two-way analysis of varianceen_US
-
Keyword(s)missing dataen_US
-
Keyword(s)type-III sum of squaresen_US
-
Dewey Decimal Classification number(s)150
-
TitleMultiple imputation to balance unbalanced designs for two-way analysis of varianceen_US
-
DRO typearticle
-
Issue1
-
Journal titleMethodology
-
Page numbers39–57
-
Volume17
-
Visible tag(s)Version of Recorden_US