How does computer-adaptive testing work?
Computer-adaptive testing selects later item difficulty from earlier responses to estimate ability with fewer questions.
Core approach and implementation process
A test may start near medium difficulty; correct answers favor harder items and incorrect answers easier ones. The algorithm combines item difficulty, discrimination and content constraints. Testing stops when uncertainty is low enough or a limit is reached. Learners may see different items while scores remain on one scale.
Classroom practice, monitoring and cautions
A large secure item bank, device access and algorithm monitoring are necessary. Repeated exposure threatens security, and accessibility features must preserve comparability. Adaptive tests may not always provide detailed instructional diagnosis. Algorithms need ongoing review for fairness, privacy and group differences.