Evaluating digital student work through model backtracking

Abstract

Designing a digital assessment system for mathematics brings along a number of challenges, such as creating an intuitive interface or providing the right feedback. We investigate how these challenges are met in different methods for evaluating digital student work. After done so, we formulate an idea for a new approach: model backtracking. That is, retracing students’ calculations through their final or intermediate answers. We explain the mathematical principle behind model backtracking and conclude that the method is promising, despite its limitations. Finally, we hypothesize about possible ways of meeting design challenges using model backtracking and provide recommendations for further research.