Concept maps make connections between ideas apparent, and thus would seem ideally suited to demonstrate learning. Yet, they are not widely used by instructors, particularly in large university classes. In this paper we review the strengths and rationale behind concept maps and provide examples we have used to structure content for one of our introductory courses. A rubric that focuses on propositions was used to grade student-created concept maps on a final exam administered to 293 students. These grades are compared to those of multiple-choice questions, short answers, and an essay on the same exam. We find poor correlation between concept maps and the other instruments, and stronger correlations between short answers and multiple choice, as well as between short-answer content grades and essay writing grades. We speculate that concept maps may be less influenced by student language skills and short-term memory than written answers, and that concept maps should be used as an alternative assessment of higher order thinking skills.