Abstract
In research on the impact of new media, it can often be observed that medial and (instructional) methodological influences are not kept apart. Therefore, little systematic research has been done so far on the instructional conditions under which the use of a certain software supports learning effectively at all. We propose an evaluation method that starts from the construct of the learning environment and advocate evaluating software under the most comprehensive control of the parameters of the respective learning environment. In this way, insights can be gained into the conditions under which the potential of a particular software comes into its own and how software should be used in the classroom in a didactically justified way.