Abstract
In line with theories of embodied cognition, hands-on experience is typically assumed to support learning. In the current paper, we explored this issue within the science domain of sinking objects. Adults had to make a guess about which of two objects in a pair would sink faster. The crucial manipulation was whether participants were handed real-life objects (real-objects condition) or were shown static images of objects (static-images condition). Results of Experiment 1 revealed more systematic mistakes in the real-objects than the static-images condition. Experiment 2 investigated this result further, namely by having adults make predictions about sinking objects after an initial training. Again, we found that adults made more mistakes in the real-objects than the static-images condition. Experiment 3 showed that the negative effect of hands-on experiences did not influence later performance. Thus, the negative effects of hands-on experiences were short-lived. Even so, our results call into question an undifferentiated use of manipulatives to convey science concepts. Based on our findings, we suggest that a nuanced theory of embodied cognition is needed, especially as it applies to science learning.