Concept maps are graphical tools for organizing and representing knowledge. They are recommended for biology learning to support conceptual thinking. In this study, we compare concept map construction (CM-c, i.e., creating concept maps) and concept map study (CM-s, i.e., observing concept maps). Existing theories and indirect empirical evidence suggest distinct effects of both formats on cognitive, metacognitive and emotional aspects of learning. We developed a CM-c training, a CM-s training, and a brief introduction to concept maps (control training) for junior high school students. We investigated effects on learning performance, concept map quality, cognitive load (cognitive effects), accuracy of self-evaluation (metacognitive effects) and enjoyment (emotional effects) of these trainings in a subsequent learning phase (CM-c learning vs. CM-s learning) in a quasi-experimental two-factorial study with 3 × 2 groups (N = 167), involving the factors training type and learning type. Results reveal that CM-c training increased learning performance and concept map quality. Effects of CM-c training on learning performance transferred onto learning with CM-s. Self-evaluation was slightly more accurate after CM-c training than CM-s training. Students reported moderate, and highly varying enjoyment during CM-c and CM-s learning. The superiority of CM-c over CM-s in learning performance and concept map quality probably lies in its characteristic of being an active learning strategy. We recommend practitioners to favor CM-c training over CM-s training, and foster students’ active engagement and enjoyment.