Nowadays, there is a strong emphasis on learning environments in which students are seen as active and self-directed constructors of knowledge.

Discovery learning is a form oflearning that fits well in this approach. In discovery learning, learners use inductive processes to generate hypotheses and design experiments to validate these hypotheses. A type oflearning environment that is specifically suited for discovery learning is (computerbased) simulation. In simulation environments, learners can change values of (input) variables and observe the consequences on values of (output) variables. These basic activities can be used for the inductive and for the validation aspects of discovery learning. The actual discovery processes of learners are determined by a number of factors that are partly outside the learner (e.g., the complexity of the domain) and partly internal to the learner. As internal determinants of discovery learning, we distinguish prior domain knowledge, generic (model) knowledge, discovery skills, and intelligence and general metacognitive skills. In a number of studies, we have examined the effects of these determinants on discovery behavior. This was done in the context of a simulation learning environment on the physics topic of geometrical optics. This chapter summarizes the theoretical background of the research program, the overall setup of the studies, the design of the learning environment, the tests that were developed to measure the determinants, and the overall findings of the program.