Reference
Kang, Hyeon-Ah. (2023). Sequential Generalized Likelihood Ratio Tests for Online Item Monitoring. Psychometrika, 88(2), 672-696. https://doi.org/10.1007/s11336-022-09871-9
Review
In her 2023 article published in Psychometrika, Kang delves into a critical dimension of psychometric testing: the continuous and intermittent monitoring of item functioning. At the heart of the article lies the introduction of sequential generalized likelihood ratio tests designed to surveil multiple item parameters across various sampling techniques. Kang's focus on the stability of item parameters across time sets significant precedence in an age where psychometric tests, especially online ones, see broad usage and necessitate consistent quality checks.
Through a combination of simulated and real assessment data, Kang validates the efficacy of the proposed monitoring procedures. One of the standout features of these methods, as highlighted in the study, is their ability to identify significant parameter shifts in a timely fashion while keeping error rates within acceptable margins. The research commendably compares these newly introduced methods against existing ones, showcasing their superior performance. Such empirical results strengthen the credibility of these procedures and their potential applicability in real-world settings.
The article suggests that multivariate parametric monitoring, anchored on robust likelihood-ratio tests, holds the promise of being a formidable tool in upholding the quality of psychometric items. Kang's emphasis on joint monitoring of multiple-item parameters provides a holistic approach to maintaining consistency and reliability. Grounded on the empirical findings, the study also offers tangible strategies for online item monitoring, an invaluable asset for practitioners in the field.
