Related Experiment Videos
Diagnosing item score patterns on a test using item response theory-based person-fit statistics
1Faculty of Behavioural Sciences, Department of Measurement and Data Analysis, University of Twente, Enschede, the Netherlands. meijer@edte.utwente.nl
Psychological Methods
|May 14, 2003
Summary
Person-fit statistics help identify deviations from item response theory (IRT) models. Parametric statistics are more effective than nonparametric ones for detecting different types of person misfit in intelligence tests.
Area of Science:
- Psychometrics
- Educational Measurement
- Item Response Theory
Background:
- Person-fit statistics are crucial for evaluating how well individual response patterns align with Item Response Theory (IRT) models.
- Investigating misfit in response patterns is essential for data quality and model validation in educational and psychological assessments.
Purpose of the Study:
- To explore the utility of person-fit statistics in detecting various types of misfit.
- To compare the effectiveness of different person-fit statistics across distinct IRT models.
Main Methods:
- Analysis of intelligence test data using person-fit statistics.
- Application of the G. Rasch (1960) and R. J. Mokken (1971, 1997) IRT models.
- Comparison of parametric and nonparametric person-fit statistics.
Main Results:
- Different person-fit statistics effectively identify distinct types of person misfit.
- Parametric person-fit statistics demonstrated higher power in detecting misfit compared to nonparametric methods.
- The choice of IRT model influences the detection of misfitting item score patterns.
Conclusions:
- Person-fit statistics are valuable tools for diagnosing misfit in IRT analyses.
- Parametric person-fit statistics offer a more powerful approach for identifying response pattern deviations.
- Model selection is an important consideration when using person-fit statistics for misfit detection.