In Search of Strategic Exploration: Using Games to Assess Students' Inquiry Strategies Within Physics Simulations

• In Event: Interactive Simulations as a Window Into Cognition: Theory-Driven Characterizations of Process Data

Abstract

Physics simulations provide rich environments for learners to explore causal relations and discover underlying principles. They also provide the possibility of assessing students’ inquiry strategies. It can be difficult to interpret student choices within an open-ended simulation, because they may have goals that are unknown to the analyst. To address this interpretive problem, we have been exploring gamified simulations. In the current instance, we modified the PhET Balance Act simulation to include a “game room,” that required answering 8 increasingly difficult questions in a row. The gaming goal helped us to make sense of student choices in the open-ended portion of the simulation where they could explore to learn how to win in the game room. We describe the results of two classroom studies.

In study 1, we taught 8th graders (n~100) one of two inquiry strategies. Half of the classes learned the Control of Variables Strategy (CVS), which involves making unconfounded comparisons by changing only one variable at a time. The other classes learned to make comparisons by looking at common outcomes and inducing the invariant structure, a move we call the General Principle Strategy (GPS). At the end of 10 lessons, students completed a written posttest on the use of inquiry strategies in a new physics domain. Then they played the gamified PhET Balance Act simulation. We found that students who used either strategy on the written posttest were more likely to successfully complete the Balance Act challenge mode, χ2(2, N = 100) = 6.14, p < .05, and on average answered more consecutive challenges correctly, F(2,98) = 6.143, p = .003. These results suggest that the gamification portion of the simulation can capture whether students are using effective inquiry strategies.

In the second study, we wanted to examine the exploration choices directly. In particular, we wanted to determine if we could differentiate inquiry strategies based on choices in the exploration room. Therefore, we told the students the strategies we wanted them to use, to see if we could detect those strategies in their choice patterns. The study again involved 8th graders (n~120). Half of the children were instructed to use CVS and half GPS. Preliminary analysis of gameplay suggests that we could identify the CVS users, because they only changed one variable a time, whereas the GPS users focused on making a “balance situation” in different ways. In both cases, the use of strategies happened in explore mode immediately after a failed challenge in the game mode, which further indicates the value of gamifying (or providing a clear criterion of performance) within a simulation intended as an assessment. Further analysis will explore how gameplay patterns representing inquiry strategies can be codified into a machine-learning algorithm, to facilitate data mining of strategic exploration within simulations.

Authors

• Luke David Conlin, Stanford University

• Daniel L. Schwartz, Stanford University