Does it make a difference if students exercise on the web or in the classroom?

Several Web-based learning environments which can automatically give immediate feedback to the students have been reported within the past few years. The quality of feedback can be relatively high in these systems, but it does not achieve the level a trained teacher can provide. However, the lack of the best possible feedback can be compensated for, to some extent, by providing immediate and individualised feedback at any place or time. The question is whether the systems providing automatic feedback are good enough to compete with humans.This paper reports on a randomised large scale intervention study. We found that there was no significant difference in the final examination results between students doing instructed simulation exercises in a classroom session and students using a web-based learning environment, if the exercises were the same. However, with more challenging exercises, there was a significant difference in the examination results, while the drop out rate was higher. Thus, the chosen teaching method and medium did not effect the level of learning, but the quality of the exercises did.