Learning paths have the potential to play an important role in the way educators serve their learners. Empirical research about learning paths is scarce, particularly in a secondary education setting. The present quasi-experimental study took place in the context of a biology course involving 360 secondary school students. A 2 × 2 factorial research design was adopted. Learners were engaged in learning activities in a learning path. These learning activities (1) differed in design and (2) were either undertaken individually or collaboratively. Gender was considered as a critical co-variable, given the focus on science learning. All learning paths were developed on the basis of visual representations, but in the experimental design conditions, learners worked with learning paths designed according to Mayer's multimedia guidelines (2003). Multilevel analyses were applied to study the impact on learning outcomes according to the design of learning paths, the individual/collaborative setting, and the co-variable gender. The study provides empirical evidence that both the design and the group setting (collaborative versus individual) have an impact on learning outcomes. Although there was no main effect, several significant interaction effects with gender were found. The results are helpful to direct research about the design and implementation of learning paths in a secondary school setting and underpin the relevance of representation modes in science learning.