Language Technologies to Support Formative Feedback

Abstract

Formative feedback enables comparison to be made between a learner’s current understanding and a desired learning goal. Obtaining this information is a time consuming task that most tutors cannot afford. We therefore wished to develop a support software tool, which provides tutors and learners with information that identifies a learner’s progress, and requires only limited human intervention. The central idea is to use language technologies to create concepts maps automatically from texts, such as students’ essays or Blogs. By comparing maps from students over time, or with maps created from tutor’s materials, or by other students, it should be possible to ascertain learners’ progress and identify remedial actions. We review existing tools for automatic construction of concepts maps and describe our initial explorations of one of these tools. This paper then introduces the theoretical background of the proposed tool, design considerations and requirements. An initial validation, which explored tutors’ perceptions of the tool showed that tutors found the approach relevant, but its implementation in practice requires to consider teachers’ practices, the tools already in use, as well as institutional policies.