What does this research mean for the field?

The proposed expectation model selection algorithm for omitted items (EMS-OI) effectively performs latent variable selection and parameter estimation in multidimensional item response theory models with nonignorable missingness. Novelty: ClaimNovelty.METHODOLOGICAL. Consensus alignment: ConsensusAlignment.NEUTRAL.

What question did this study set out to answer?

Develop and evaluate a latent variable selection method for multidimensional item response theory models when item omissions represent nonignorable missingness.

September 2, 2026Journal of Educational and Behavioral Statistics

Latent Variable Selection in Multidimensional Item Response Theory Models With Nonignorable Missingness

View Full Paper

Ask AI

Bookmark

Authors

QZQian‐Zhen Zheng PXPing‐Feng Xu LSLaixu Shang

Discussion

Member takes

Overview

Methodological study demonstrates accurate latent variable selection in multidimensional test responses with missing data, highlighting improved parameter estimation in educational testing.

Key Points

Develop and evaluate a latent variable selection method for multidimensional item response theory models when item omissions represent nonignorable missingness.
Formulated a joint latent variable model combining a Rasch model for missingness propensities with a multidimensional two-parameter logistic model for item responses.
Engineered an Expectation Model Selection algorithm for omitted items (EMS-OI) to simultaneously perform variable selection and item parameter estimation.
Evaluated performance using synthetic simulation experiments and an empirical validation using the PISA 2012 assessment dataset.
EMS-OI demonstrated competitive accuracy in identifying item-ability structures and estimating model parameters compared with traditional EMS and expectation-maximization baselines.
Application to the PISA 2012 dataset verified the algorithm's empirical feasibility and robust estimation behavior in large-scale assessment settings.

Cite This Study

Zheng et al. (2026) studied this question.

synapsesocial.com/papers/6a97e25ec562ede874ec66cf https://doi.org/10.3102/10769986261480063

View Full Paper

Ask AI

Bookmark

September 2, 2026Journal of Educational and Behavioral Statistics

Latent Variable Selection in Multidimensional Item Response Theory Models With Nonignorable Missingness

View Full Paper

Ask AI

Bookmark

Authors

QZQian‐Zhen Zheng PXPing‐Feng Xu LSLaixu Shang

Discussion

Member takes

Overview

Methodological study demonstrates accurate latent variable selection in multidimensional test responses with missing data, highlighting improved parameter estimation in educational testing.

Key Points

Develop and evaluate a latent variable selection method for multidimensional item response theory models when item omissions represent nonignorable missingness.
Formulated a joint latent variable model combining a Rasch model for missingness propensities with a multidimensional two-parameter logistic model for item responses.
Engineered an Expectation Model Selection algorithm for omitted items (EMS-OI) to simultaneously perform variable selection and item parameter estimation.
Evaluated performance using synthetic simulation experiments and an empirical validation using the PISA 2012 assessment dataset.
EMS-OI demonstrated competitive accuracy in identifying item-ability structures and estimating model parameters compared with traditional EMS and expectation-maximization baselines.
Application to the PISA 2012 dataset verified the algorithm's empirical feasibility and robust estimation behavior in large-scale assessment settings.

Cite This Study

Zheng et al. (2026) studied this question.

synapsesocial.com/papers/6a97e25ec562ede874ec66cf https://doi.org/10.3102/10769986261480063

View Full Paper

Ask AI

Bookmark