Evaluation of Logic Proof Problem Difficulty through Student Performance Data

The interactions of concepts and problem-solving techniques needed to solve open-ended proof problems are varied, mak-ing it difficult to select problems that improve individual student performance. We have developed a system of data-driven ordered problem selection for Deep Thought, a logic proof tutor. The problem selection system presents prob-lem sets of expert-determined higher or lower difficulty to students based on their measured proof solving proficiency in the tutor. Initial results indicate the system improves student-tutor scores; however, we wish to evaluate problem set difficulty through analysis of student performance to val-idate the expert-authored problem sets