Data-informed Decision-making in TEFA Processes: An Empirical Study of a Process Derived from Peer-Instruction

Abstract

When formative assessment involves a large number of learners, Technology-Enhanced Formative Assessments are one of the most popular solutions. However, current TEFA processes lack data-informed decision-making. By analyzing a dataset gathered from a formative assessment tool, we provide evidence about how to improve decision-making in processes that ask learners to answer the same question before and after a confrontation with peers. Our results suggest that learners' understanding increases when the proportion of correct answers before the confrontation is close to 50%, or when learners consistently rate peers' rationales. Furthermore, peer ratings are more consistent when learners' confidence degrees are consistent. These results led us to design a decision-making model whose benefits will be studied in future works.