Identifying Heywood Cases in Item Response Theory
Author Information
Author(s): Jay Verkuilen, Peter J. Johnson
Primary Institution: CUNY Graduate Center, New York, NY, USA
Hypothesis
The study proposes a new method for identifying Heywood cases in item response theory models using Fisher information.
Conclusion
The proposed method provides a more objective and automated way to detect Heywood cases, which can significantly impact scoring and model identification.
Supporting Evidence
- Heywood cases indicate model misspecification and can lead to incorrect scoring.
- The proposed method uses Fisher information to provide a more objective detection of Heywood cases.
- Empirical examples demonstrate the effectiveness of the new method in identifying problematic items.
Takeaway
This study helps researchers find problems in test questions that can lead to incorrect scoring, making it easier to fix those issues.
Methodology
The study uses empirical examples and proposes a new metric called Item Fraction of Total Information (IFTI) to identify Heywood cases.
Limitations
The study acknowledges that the proposed methods may not eliminate all issues related to Heywood cases and that careful design is still necessary.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website