Understanding Question Difficulty and Student Ability in Mathematics Assessments

Education software is abundant on the web, but few support the analysis of students’ performance and test difficulty over user-defined questions and learning objectives. The idea of this project is to create a computer program to provide educators with insights on student performance on assessment items and perceived and actual assessment item difficulty within fine-grained learning objectives. This tool also provides departments with key metrics to identify potential course delivery issues and compare students’ understanding of course material year over year. The software proposed here estimates item difficulty and student ability by building a mathematical model from student performance data on assessment question items. In addition, an interface is envisioned to help educators and students better understand model results based on fine-grained learning objectives.

Faculty Supervisor:

Marcus Santos;Dejan Delic

Student:

Partner:

Bitbolide Inc.

Discipline:

Engineering

Sector:

Education

University:

Toronto Metropolitan University

Program: