The Multicomponent Latent Trait Model for Diagnosis: Applications to Heterogeneous Test Domains.

Area of Science:

• Educational Measurement
• Psychometrics
• Psychological Testing

Background:

• Heterogeneous item content is common in tests measuring global traits like intelligence and achievement.
• These tests often cover a broad domain with items varying in required attributes and skills.
• Hierarchical structures are frequently needed to manage the complexity of such test domains.

Purpose of the Study:

• To present and apply the multicomponent latent trait model for diagnosis (MLTM-D) to heterogeneous tests.
• To enable diagnostic assessment at multiple levels of specificity within a test domain.
• To diagnose mastery of broad and specific skills in middle school mathematics.

Main Methods:

• Application of the multicomponent latent trait model for diagnosis (MLTM-D).
• Utilizing a confirmatory modeling approach.
• Describing the methods for applying MLTM-D to diagnose skill mastery.

Main Results:

• The multicomponent latent trait model for diagnosis (MLTM-D) was successfully applied to middle school mathematics achievement data.
• The model demonstrated capability in diagnosing student mastery at both broad and specific skill levels.
• The study validated the utility of MLTM-D for heterogeneous testing contexts.

Conclusions:

• The multicomponent latent trait model for diagnosis (MLTM-D) is a viable tool for analyzing heterogeneous tests.
• MLTM-D provides detailed diagnostic information, supporting targeted educational interventions.
• The model's hierarchical diagnostic capability is particularly useful for complex subject areas like mathematics.