Computational tools hold promise for science education, but each models information in a different way. This paper explore ways to connect what students already know with the different ways it might be represented.

Agent-based modeling can get students thinking about mechanism. But to get students refining their models toward accurate explanations, comparing models with data is just as important.