How Can We Use AI-Based Assessment on Reflection Narrative for Program Improvement?
Non-cognitive skills such as GRIT are easy to measure, but it is difficult to specify why and how each non-cognitive skill can be developed. At Dalton Tokyo Middle and High School, a New York-based Dalton Plan school in Japan, we have developed an evidence-based system for improving curriculum and programmes to develop non-cognitive skills. It is a comprehensive analytical process of measuring and specifying the pedagogical impact on non-cognitive skills through intensive assessment analysis of depth of learning credentials. In this process we are testing how we can assess them through AI supported tools with student reflective narrative rather than numerical self-assessment. This resulted in a linked analysis between the non-cognitive skills score as a result and the AI narrative analysis as an input trigger for student growth. For example, students with high growth scores had higher than average empathy scores, which stemmed from the growth in mindset and the behaviour of learning from failure. This can be analyzed using AI-based narrative assessment. In this presentation, we will present the method and improvement process of the evidence-based development scheme, including the analytical framework and the AI’s attempt to measure the depth of learning.
Facilitated By
Terry Furuya
Dalton Tokyo Middle and High School
R&D and Business Development in Dalton Tokyo Middle and High School and Executive Officer in Kawaijuku Holdings. Prior to this position, Terukane was Director of advertising business in X (Former Twitter) covering South Korea and Japan, and Head of self-serve advertising business in Amazon to look after Japan and China businesses. Terukane received a master degree of Public Policy from University of Michigan and bachelor degree of Economics from The University of Tokyo.