Cognitive diagnostic assessment with different weight for attribute: based on the Dina model

Area of Science:

• Educational Measurement
• Psychometrics
• Cognitive Science

Background:

• Cognitive diagnostic models (CDMs) are crucial for understanding student knowledge states.
• Accurately weighting attributes is essential for effective cognitive diagnosis.
• Existing methods may be limited to specific CDM types.

Purpose of the Study:

• To propose a novel, model-agnostic attribute weight calculation method for cognitive diagnosis.
• To evaluate the performance of the proposed method using simulation studies.
• To compare the effectiveness of weighted versus unweighted attribute status in classification.

Main Methods:

• A Bayesian network and the least squares distance method were integrated for attribute weight calculation.
• The proposed method was designed to be independent of specific cognitive diagnostic models.
• Simulation studies were conducted to assess data fit and classification accuracy.

Main Results:

• The least squares distance method demonstrated excellent data fit.
• Weighted attribute status significantly improved correct classification rates compared to unweighted status.
• The weighted approach showed promising performance in recognizing examinee knowledge states across various conditions.

Conclusions:

• The proposed attribute weight calculation method is a practical and effective tool for cognitive diagnosis.
• This model-agnostic approach enhances the accuracy of knowledge state assessment in diverse diagnostic settings.
• Attribute weighting is a valuable consideration for improving examinee classification in educational assessments.