Leveraging Minecraft to measure and promote middle school STEM interest

• In Event: Contributing factors to digital game efficacy for learning among children and early adolescents

Abstract

In addition to transfer effects from educational video games, it is also important to consider the ways in which video game play may affect how learners feel about the learning content and how it may influence their future learning. One such game that is currently influencing several generations of learners, Minecraft, is now the most popular video game in history with over 3 million daily users. Our second presenter will report on their work to identify patterns of play and use of STEM learning resources that reflect the development of interest in STEM. This study draws on extant work that has been done to investigate how Minecraft can be used as a STEM learning environment.
The project, What-if Hypothetical Implementations in Minecraft (WHIMC), provides numerous Minecraft worlds for learners to explore that each present alternative versions of Earth. These worlds are often radically different from the Earth we have and invite comparisons. For example, Earth with no moon would spin about three times as fast as the actual Earth leading to unique surface features, such as shorter and stronger trees, more smoothly rounded mountains, and less vigorous tidal behaviors. Learners use a custom set of science tools to measure these environments (e.g., check radiation levels, temperature, atmospheric composition) with the goal of investigating the habitability of these worlds. On the server, learners can also explore a NASA launch site, space station, moon base, four more “What if” worlds, and three known exoplanets (planets that orbit stars other than our Sun).
Data on 48 middle school learners was collected in the summers of 2021 and 2022 in a summer camp format (3 hours per day, 5 days of the week). To measure changes in learning and interest (Renninger & Hidi, 2015), five categories of data were included: 1) interviews, 2) surveys, 3) observations (textual descriptions of world features), 4) exploration behaviors (behavioral log data on movement and interaction), and 5) self-explanations (1-2 sentence long responses to in-game prompts). Interviews and surveys were done outside of the Minecraft environment, while the remaining three were collected in the game. Importantly, we detected significant correlations between self-explanations and observations in both groups (r(20) = 0.50, r(26) = 0.49, p < .05) suggesting a link between scientifically accurate and relevant observations and being able to explain them accurately. Additional significant correlations in our 2021 data were found between interest development (surveys) with exploration behaviors (r(20) = 0.55) and self-explanations (r(20) = 0.54, p < .05) further suggesting that interest development in STEM is associated with how children explore our different worlds and the quality of self-explanations given in response to in-game prompts.

Authors

• H. Chad Lane, University of Illinois at Urbana-Champaign
  Presenting Author

• Matthew Gadbury, University of Illinois-Urbana-Champaign

• Jeff Ginger, University of Illinois-Urbana-Champaign

• Sam Hum, University of Illinois-Urbana-Champaign