Automatic assessment of mathematical programming exercises with Numbas

Abstract

As programming has become a common feature of undergraduate mathematics degrees, there has been an increasing focus on how to teach and assess the subject to mathematicians. The potential benefits of e-assessment of basic programming exercises have many parallels with assessment in mathematics where e-assessment tools are widely used: the chance to give instant feedback to students offers an opportunity to allow students to work at their own pace, accommodating the disparate background in programming that often exists in undergraduate mathematics cohorts. And the randomisation of question content not only offers a powerful tool for practice, with students able to repeat similar problems over and over, it also can offer some protection against plagiarism in a subject where, just like a solution to some mathematical problems, student answers to identical problems are likely to be very similar. This paper considers an extension to Numbas to automatically assess programming exercises and the successful implementation of the resource in undergraduate modules using the programming languages R and Python.