Does the specific structure of advance organizers influence learning outcomes? In the first experiment, 48 psychology students were randomly assigned to three differently structured advance organizers: a well-structured, a well-structured and key-concept emphasizing, and a less structured advance organizer. These were followed by a sorting task, a text study phase, and a posttest. The results indicated that differently structured advance organizers lead to different proto-schemata before and different learning outcomes after the text study phase. The second experiment replicated and extended these findings with 53 mathematics students. As in experiment 1, three differently structured advance organizers were used; but to rule out alternative explanations, the sorting task between the advance organizer and the text study phase was omitted. The results showed strong beneficial effects of well-structured advance organizers on near and far transfer tasks. Taken together, both experiments support the claim that the structure of advance organizers has an effect on preliminary schemata and learning outcomes. On a general level, the results indicate that advance organizers can support the generation of proto-schemata and thus can be more than the activation of "existing" concepts in long-term memory. With regard to education, this implies that educators should not only think about whether prior domain-specific knowledge is present, but also about how to scaffold the generation of proto-schemata at the beginning of instruction.