Supporting the translation between multiple representations

Modern, computer based, learning environments often embrace a multitude of representations. Research with learning environments that contain multiple representations has revealed that learning with multiple representations can lead to deeper understanding but also that processing different representations can be a demanding process. In a multi-representational learning environment the learner has to understand each representation and, when the representations are (partially) redundant, has to translate between the representations. The current study is the first of a series of studies in which the translation between representations is studied within a simulation based learning environment. Students can be supported in this translation process by linking representations dynamically (which means that concurrent changes occur in separate represen-tation) or spatially (which means that related representations are linked in place). In the present study three versions of the same simulation were compared: a simulation with dynamically linked integrated representations, a simulation with dynamically linked separated representations, and a simulation with non-linked separated representations. Eighty-nine Dutch students from four middle vocational training schools participated in the experiment. Subjects worked with a SimQuest simulation in which they studied the behaviour of moments in mechanical engineering. The experiments were held at the four participating schools and consisted of three sessions: pre-test, working with the simulation, and post-test. Three types of items were used in the tests: items on subject matter content, items on translation between representations, and items on transfer problems. The paper will present the results of these experiments to examine how different forms of translation influence learning.