Many-facet Rasch analysis with crossed, nested, and mixed designs

Area of Science:

• Educational measurement
• Psychometrics
• Social sciences

Background:

• Many-facet Rasch analysis is crucial for evaluating rater performance and task difficulty.
• Typical designs involve crossed facets, allowing direct comparison between judges and conditions.
• Nested designs restrict comparisons, limiting the ability to link measures across different conditions.

Purpose of the Study:

• To explore how many-facet Rasch analysis can be adapted for mixed measurement designs.
• To demonstrate the utility of mixed designs in establishing a common vertical ruler for linking measures.
• To address the connectivity requirement for comparing facet measures across different frames of reference.

Main Methods:

• Investigated the application of many-facet Rasch analysis to mixed measurement designs.
• Presented examples of crossed, nested, and mixed designs.
• Illustrated modifications to the analysis to ensure measure commensurability.

Main Results:

• Mixed designs in many-facet Rasch analysis can achieve a common vertical ruler.
• This approach allows for the comparison of measures across facets that are not directly crossed.
• Connectivity requirements for linking measures are met through appropriate design and analysis.

Conclusions:

• Many-facet Rasch analysis can be effectively implemented with mixed designs.
• Mixed designs offer a flexible framework for establishing a common vertical ruler in complex measurement situations.
• This methodology enhances the fairness and meaningfulness of decisions based on individual ratings.