Deterministic Input, Noisy Mixed Modeling for Identifying Coexisting Condensation Rules in Cognitive Diagnostic...

Area of Science:

• Cognitive psychology
• Educational measurement
• Psychometrics

Background:

• Condensation rules in cognitive diagnosis models link required attributes to item responses.
• Multiple or coexisting condensation rules reflect complex cognitive processes and potential inconsistencies with expert assumptions.

Purpose of the Study:

• To evaluate the deterministic input with a noisy mixed (DINMix) model for identifying coexisting condensation rules.
• To provide feedback for item revision to enhance the validity of cognitive process measurement.

Main Methods:

• The study proposed and evaluated the DINMix model.
• Two simulation studies were conducted to assess the model's psychometric properties.
• An empirical example was analyzed to demonstrate practical application.

Main Results:

• The DINMix model demonstrated adaptive and accurate identification of coexisting condensation rules.
• The model successfully identified rules existing simultaneously within an item or separately across multiple items.

Conclusions:

• The DINMix model is effective for identifying complex cognitive processes and coexisting condensation rules.
• The model offers advantages for item revision and improving the validity of cognitive diagnostic assessments.