Options
Multiple imputation of binary multilevel missing not at random data
Hammon, Angelina; Zinn, Sabine (2020): Multiple imputation of binary multilevel missing not at random data, in: Journal of the Royal Statistical Society: Series C (Applied Statistics), Oxford: Wiley-Blackwell, Jg. 69, Nr. 3, S. 547ā564, doi: 10.1111/rssc.12401.
Faculty/Chair:
Author:
Title of the Journal:
Journal of the Royal Statistical Society: Series C (Applied Statistics)
ISSN:
1467-9876
0035-9254
Publisher Information:
Year of publication:
2020
Volume:
69
Issue:
3
Pages:
Language:
English
DOI:
Abstract:
We introduce a selection model-based multilevel imputation approach to be used within the fully conditional speciļ¬cation framework for multiple imputation. Concretely, we apply a censored bivariate probit model to describe binary variables assumed to be missing not at random. The ļ¬rst equation of the model deļ¬nes the regression model for the missing data mechanism. The second equation speciļ¬es the regression model of the variable to be imputed. The non-random selection of the binary data is mapped by correlations between the error terms of the two regression models. Hierarchical data structures are modelled by random intercepts in both equations. To ļ¬t the novel imputation model we use maximum likelihood and adaptive GaussāHermite quadrature. A comprehensive simulation study shows the overall performance of the approach.We test its usefulness for empirical research by applying it to a common problem in social scientiļ¬c research: the emergence of educational aspirations. Our software is designed to be used in the R package mice.
GND Keywords:
Fehlende Daten ; Imputationstechnik
Keywords:
Fully conditional speciļ¬cation; Missingness not at random; Multilevel data; Multiple imputation; Selection model
DDC Classification:
RVK Classification:
Type:
Article
Activation date:
June 17, 2022
Versioning
Question on publication
Permalink
https://fis.uni-bamberg.de/handle/uniba/54235