Hierarchical diagnostic classification models morphing into unidimensional 'diagnostic' classification models-a...

Area of Science:

• Psychometrics
• Educational Measurement
• Cognitive Psychology

Background:

• Cognitive diagnostic models (CDMs) are increasingly used for assessing skills and knowledge.
• Templin and Bradshaw's 2013 paper proposed hierarchical diagnostic classification models.
• These models assume specific structures for attribute relationships.

Purpose of the Study:

• To critically evaluate the modeling approaches and terminology in Templin and Bradshaw's work.
• To discuss the implications of assuming attribute hierarchies and interactions in CDMs.
• To highlight potential issues and offer avenues for further research in diagnostic classification.

Main Methods:

• A critical commentary and analysis of existing literature.
• Examination of the assumptions underlying hierarchical diagnostic classification models.
• Illustration of conceptual issues with concrete examples.

Main Results:

• Identified several concerns regarding the application of CDMs with pre-defined attribute hierarchies.
• Highlighted the potential limitations when assuming specific attribute interaction forms.
• Provided a framework for discussing the validity and interpretation of such models.

Conclusions:

• The commentary underscores the importance of carefully considering model assumptions in cognitive diagnostic assessment.
• Researchers should be mindful of the implications of hierarchical structures and attribute interactions.
• Further research is needed to refine CDMs and their application in educational and psychological measurement.