Engineering Freshman assessment of mathematical skills for autonomous learning.doc

Engineering Freshman assessment of mathematical skills for autonomous learning
. Desmarais, 1C. Burney-Vincent, 2F. Caron, 1M. Gagnon, 1O. Gagnon, 1F. Lemieux, 1I. Pelczer, 1J.-M. Robert,
École Polytechnique de Montréal, Canada, {,-vincent,,,}***@;
Université de Montréal, Canada, {france.******@}2
Abstract
Formative assessments provide critical feedback to the students. They allow them to appreciate where they stand with respect to a course or a program requirements and with respect to their peers. This information is particularly important in an autonomous learning context and for distance learning. For such contexts, the ability to provide an accurate assessment through a small number of questions or exercises represents a substantial advantage. The length of the assessment is particularly important to reduce frustration from long quizzes and allow frequent feedback. We assess the gain that can be obtained by relying on an adaptive testing approach using Item Response Theory. We investigate this potential gain through a simulation study and using data from a mathematical test administered to freshman engineering students. The results indicate that an adaptive testing approach can reduce the number of questions by half to obtain an assessment equivalent to a traditional test