Impact of Digital Game-Based Learning Experiences on Reasoning Skills and Science Engagement

A study of 143 seventh-grade students at a public middle school in New Jersey examines whether a choose-your-own-adventure simulation game can strengthen scientific reasoning, specifically the Control of Variables Strategy (CVS), and whether it can increase engagement with science. Drawing on research in science education, embodied cognition, and game-based learning, the study uses a Preparation for Future Learning design in which students were randomly assigned to unlimited play, structured play, or a control condition reading an explanatory CVS document. Reasoning was measured at baseline, immediately after training, and one week later, with science motivation and game experience captured through validated questionnaires.

Results were mixed across the three time points. Baseline scores were lowest in the unlimited play group and highest in the control group, a surprising imbalance the authors attribute to possible unobserved factors. After training, the unlimited play group achieved the strongest post-test performance, roughly 150% higher odds of answering a question correctly than controls, while the structured group scored in the middle. On the one-week retention test, the structured group performed significantly worse than both other groups, while unlimited play and control were nearly identical. Survey analysis found the unlimited play group reported higher self-determination than the structured group and higher intrinsic motivation and self-efficacy than controls, though only the self-determination difference survived multiple-comparison correction.

The authors note ceiling effects on the CVS instrument, which was originally validated with younger children, and acknowledge that the reversal from baseline to post-test is suggestive but not statistically verified. They also report that a fire alarm interrupted one group's post-test, prompting a reanalysis that partially altered the pairwise comparisons. Future work is recommended to separate the game's direct effects from those of formal instruction, record actual gameplay time, and test adaptations for students at risk of disengagement or with learning disabilities.