How to structure an unstructured activity: Generating physics rules from simulation or contrasting cases

Studies show that having students attempt to invent a scientific rule ("invention activity") before receiving direct instruction benefits their learning. However, the design affordances of these invention activities have not been fully addressed. In this paper, we compare the effect of two different brief invention activities on student learning about buoyancy. In one treatment condition, students are provided with contrasting cases of a phenomenon to invent the rule. In the other treatment condition, students explore the phenomenon using an interactive simulation. Students in the contrasting cases condition invented more complete rules and performed significantly better in solving buoyancy problems. We hypothesize that this difference between conditions results from different levels of attention to important features of the phenomenon, which we illustrate with differences in what students explore in the simulation. This study suggests that proper scaffolding to ensure sufficient exposure to the underlying structure is essential in designing an invention activity.