Individualized Learning Patterns Require Individualized Conversations – Data-Driven Insights from the Field on How Chatbots Instruct Students in Solving Exercises

Abstract

AbstractChatbots can foster the learning success of students in educational settings. This has been shown in prior research studies, e.g., using laboratory studies in online learning settings. Actual evaluations of using educational chatbots in the field are nevertheless rarely reported. Thus, insights into the students’ interactions with chatbots in long-term field settings are missing. In this research study, we aim at gaining insights into the students’ interactions with an educational chatbot in a programming course. To this aim, we follow an explorative data-driven discourse analysis approach and show how students interacted with the chatbot during a field study lasting several months. We ground our analysis on a dataset from 54 students interacting with a chatbot while working on programming exercise tasks. We reveal how students interact with the chatbot and identify different types of usage patterns. The results imply that adaptive learning paths are one of the most important aspects of educational chatbot design for dealing with heterogeneous usage patterns.KeywordsEducational chatbotsDigital tutorsLearning processField studyEvaluation