NATURAL LANGUAGE PROCESSING AND SCORING STUDENT WRITING IN WEB-BASED INTELLIGENT TUTORING SYSTEMS

The most complex part of any web based learning environment’s development is the assessment of the student work. The scoring is integral to assessing the students’ answers and proving the most appropriate and useful feedback and guidance. Natural language processing (NLP) has a long history of complexities that range from simple parsing of text to actually trying to make meaning of what the students’ writing (Graesser, ; Landauer et al., ). This paper presents the NLP challenges faced in the development of the Intelligent Tutoring System for the Structure Strategy (ITSS), the solution implemented and the enhancements reviewed. This line of research should aid any other team considering the development of interactive learning environments that require students to write responses and for the system to assess the responses and provide feedback.

The paper begins with a short description of ITSS followed by challenges in scoring the full recalls that students write in response to questions, scoring system, and approaches to addressing these challenges.