Investigating the Impact of Unsolicited Next-Step and Subgoal Hints on Dropout in a Logic Proof Tutor (Abstract Only)
| Autor: | Christa Cody, Behrooz Mostafavi |
|---|---|
| Rok vydání: | 2017 |
| Předmět: |
Computer science
business.industry Process (engineering) 05 social sciences 050301 education 02 engineering and technology Intelligent tutoring system Test (assessment) 020204 information systems ComputingMilieux_COMPUTERSANDEDUCATION 0202 electrical engineering electronic engineering information engineering Mathematics education State (computer science) Artificial intelligence TUTOR business 0503 education computer Dropout (neural networks) computer.programming_language |
| Zdroj: | SIGCSE |
| DOI: | 10.1145/3017680.3022426 |
| Popis: | We have been incrementally adding data-driven methods into the Deep Thought logic tutor for the purpose of creating a fully data-driven intelligent tutoring system. Our previous research has shown that the addition of data-driven hints, worked examples, and problem assignment can improve student performance and retention in the tutor. In this study, we investigate the influences two unsolicited hint types have on students' ability to complete the tutor. We have used data collected from two test conditions: one with unsolicited next step hints (NSH) presenting the immediate next step of a logic proof to a student's current proof-solving state, and the other with unsolicited subgoal hints (SGH) presenting a step of a logic proof two or three steps of the student's current state. Our results show that students who received unsolicited SGH had more interactions within the tutor and skipped more problems. Furthermore, the SGH group had a significantly higher dropout percentage. These results suggest that hint types can affect student behavior and the ability to learn the material. Therefore, determining what type of hint to give during problem solving is important to the learning process and should be taken into consideration when designing an intelligent tutoring system (ITS). Future work will include using historical student data to determine the best hint type to give a student by analyzing student behavior and identifying the most effective hint type for the behavior being exhibited. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|