An Online Environment to Compare Students’ and Expert Solutions to Ill-Structured Problems

Abstract

Practitioners often face ill-structured problems. However, it is difficult for instructors to assess their students’ work on such problems, as a broad set of solutions exist and may depend on the context. One way to assess student learning is through the evaluation of their mental models, which can be presented in the form of a causal network or ‘map’. While comparing a student’s map to an expert’s map can assist with the evaluation, this is a challenging process, in part, due to variations in language, resulting in the use of different terms for the same construct. The first step of the comparison is to address these variations by aligning as many of the students’ terms with their equivalent in the expert’s map. We present the design and implementation of a software to assist with the alignment task. The software improves on previous work by optimizing usability (e.g., minimizing the number of clicks to create an alignment) and by leveraging previous alignments to recommend new ones. In addition, alignments can be done collaboratively, as our system is available online: one instructor can invite others to edit or see the alignments. Further improvements to this system may be achieved using content-based recommender systems or natural language processing.

Publication Title

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

Share

COinS