It was investigated to what extent example variability and the elicitation of sophisticated self-explanations foster the acquisition of applicable and transferable knowledge by learning from worked-out examples. To this end, we had 56 apprentices from a bank learn calculation of compound interest and real interest. The subjects were randomly assigned to the four conditions of a 2´2-factorial design (factor 1: uniform vs. multiple examples; factor 2: spontaneous vs. Elicited self-explanations). The learning results were measured by a post-test comprising application problems (...) and transfer problems. It was found that the acquisition of applicable and transferable knowledge can be supported by eliciting self-explanations. This underlines the "causal" relevance of the quality of self-explanations for knowledge acquisition by learning from worked-out examples. The assumption that multiple examples foster transfer performance, at least when sophisticated self-explanations are elicited, is not unequivocally supported. Without learner support in form of an elicitation procedure, multiple examples even seem to be detrimental. (shrink)