Adaptive games for education: Studying engagement and performance of active learners

• In Event: Poster Session 01
  In Poster Session: PS 01 Topic - Technology, Media & Child Development

Abstract

This study investigates whether preschoolers learn more effectively when they choose what material to study. We focus on one dimension of active learning: the consequence of allowing learners to make decisions about the information they want to experience (Gureckis and Markant, 2012). Previous work has supported an active learning advantage in recognition memory and category learning in children (Ruggeri et al., 2016; 2019; Sim et al., 2016). We ask whether children effectively monitor and guide their learning when they choose what material they want to study (letters vs. numbers, instead of individual stimuli). In order to test whether an active learning advantage exists in letter and number recognition, we will compare engagement and performance in two conditions: active learners will be able to control time allocation between letter and number games that adapt to their level; control learners will follow a standard preschool curriculum with no adaptive game dynamics. With the current prevalence and increasing importance of remote learning and its disruption to teachers' ability to gauge their students’ knowledge and guide them accordingly, it is critical to assess approaches that can help children directly through nudges in their digital environment.
In our planned week-long intervention, 60 4-year-olds will play independently with tablet games that teach letter and number recognition with minimal parental influence for at least four 15-minute sessions. In each game, participants will hear the name of a letter or number and select the corresponding character from several choices, presented on flowers. If the child answers incorrectly, they can try again with the help of a bee that flies slowly towards the correct answer. Participants in the active condition will be able to choose between letter and number games in the main menu and switch between them at any time. Within each game, an adaptive algorithm chooses the next target stimulus, re-prioritizes the content, and adjusts the difficulty (i.e., the number of foils and the speed of the bee) based on the participant’s performance history. Participants in the control condition cannot actively switch between the letter and number games and are instead presented with the stimuli in a predetermined order, with letter and number stimuli interleaved at a set interval. We will measure engagement and performance through app usage data and parent surveys gauging each child’s prior knowledge and overall enjoyment. We expect that children who are given active control over which content they study will engage more with the apps than children in the control condition (playing more frequently and/or playing for longer periods of time in each session), and show greater gains in knowledge (speed and accuracy). Moreover, we expect that children who have greater parent-rated interest and/or prior knowledge in one of the content areas (numbers or letters) will spend more time playing the related game.
We have piloted the apps and will recruit participants online via mailing lists from a local preschool to run the experiment in October 2020.

Authors

• Gloria Yi, Stanford University
  Presenting Author

• Alexandra Carstensen, University of California - San Diego
  Non-Presenting Author

• George Kachergis, Stanford University
  Non-Presenting Author

• Michael C. Frank, Stanford University
  Non-Presenting Author