Cognitive Diagnostic Computerized Adaptive Testing
Also known as: CD-CAT
Cognitive Diagnostic Computerized Adaptive Testing (CD-CAT) combines computerized adaptive testing (CAT) with cognitive diagnostic models (CDMs) to efficiently assess students' specific skill profiles. Rather than producing a single overall ability score, CD-CAT adaptively selects items to quickly identify which skills a student has mastered and which need development.
Read the full method
Sign in with a free account to read this section.
Method map
The neighbourhood of related methods — select a node to explore.
When to use it
Apply CD-CAT when you need to diagnose specific skill profiles efficiently, want to minimize testing time while maximizing diagnostic accuracy, or need adaptive difficulty progression that targets skill gaps. Ideal for educational systems, competency certification, and formative assessment where actionable diagnosis is more important than rank-ordering.
Strengths & limitations
- Efficient: typically requires fewer items than fixed tests to achieve accurate diagnosis, reducing testing burden
- Skill-focused: targets specific skill deficits, enabling targeted remediation rather than generic tutoring
- Adaptive difficulty: automatically adjusts item difficulty to challenge without overwhelming the examinee
- Real-time feedback: can provide immediate skill-by-skill performance feedback for instructional guidance
- Reduced testing time: decreases overall assessment duration compared to fixed-length tests, freeing instructional time
- Requires strong CDM: depends on valid cognitive diagnostic model specifying which items assess which skills
- Item bank demands: needs a well-calibrated bank of items aligned to skill attributes
- Complex implementation: requires specialized software and training to administer and interpret
- Skill independence assumption: many CDMs assume skills are independent or have specified dependencies; violations bias results
Frequently asked
What cognitive diagnostic model should I use in CD-CAT?
Choice depends on your context. DINA (Deterministic Inputs, Noisy Outputs) is simple and widely used. DINO and more complex models allow probabilistic skill mastery. Select based on whether skills are truly discrete (mastered/not) or probabilistic (partial mastery).
How many items do I need in my item bank?
Typically 3-5 items per skill attribute minimum; more is better for efficient item selection. With few items per skill, the algorithm may exhaust options. Aim for sufficient variety that students with different skill profiles see different items.
When should testing stop in CD-CAT?
Common stopping rules include: reaching a confidence threshold (e.g., 95% probability) for the examinee's skill profile, exhausting a maximum item count, or when additional items provide minimal information. Rules balance accuracy against testing time.
How do I select items adaptively?
Common criteria include: Fisher information (expected reduction in ability variance), KL divergence (distance between skill profiles), or posterior predictive probabilities. Different criteria optimize for different goals; experiment and validate.
How do I communicate CD-CAT results to students and educators?
Avoid presenting skill profiles as deterministic. Instead, use confidence intervals or color-coding to show which skills are clearly mastered, unclear, or clearly deficient. Include recommendations for next instructional steps based on skill gaps.
Sources
- Choi, K. M., Lee, Y. S., & Park, Y. S. (2015). What CDM can tell about examinees' strengths and weaknesses: Cognitive diagnostic information in TIMSS. Journal of Educational Evaluation for Policy Analysis, 24(1), 79-100. link ↗
- Kaplan, M., de la Torre, J., & Barrada, J. R. (2015). New item selection methods for cognitive diagnosis computerized adaptive testing. Journal of Educational Measurement, 52(4), 393-411. DOI: 10.1177/0146621614554650 ↗
- Hsu, C. L., Wang, W. C., & Chen, S. Y. (2013). Variable-length computerized adaptive testing based on cognitive diagnosis models: A simulation study. Applied Psychological Measurement, 37(1), 3-23. DOI: 10.1177/0146621613488642 ↗
How to cite this page
ScholarGate. (2026, June 3). Cognitive Diagnostic Computerized Adaptive Testing. ScholarGate. https://scholargate.app/en/psychometrics/cognitive-diagnostic-cat
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
- DINA ModelPsychometrics↔ compare
- DINO ModelPsychometrics↔ compare
- Latent Transition AnalysisPsychometrics↔ compare
- Necessary Condition AnalysisPsychometrics↔ compare
- Rule Space MethodologyPsychometrics↔ compare