Re-Designing Assessments—Recommended Test-Length Based on Estimates of Either Reliability or Classification Accuracy

The project focuses on the design of exams in medical studies. These exams typically consist of 100 to 200 or more questions. One reason for this is to ensure reliability of the exam. This study explores how different statistical criteria influence the selection of the number of tasks and questions. This could potentially open opportunities for smaller exam sets and new ways of testing students' knowledge and skills.