The use of investigations in science teaching is both common and commonly advocated for in science education literature. We suggest that the use of investigations should differ depending on the complexity of the subject matter. That complexity can vary depending on both the nature of the information and students' expertise levels. The present study used Cognitive Load Theory and tested a hypothesis that the sequential order of having investigations first followed by explicit instruction only is more effective than that of having explicit instruction first when students have acquired sufficiently high levels of knowledge. This hypothesis was tested with 85 middle school students. Participants were randomly assigned to one of the two sequences. The results indicated that for sufficiently knowledgeable students, a dis-ordinal knowledge level by sequence interaction was obtained. There was an advantage to the explicit instruction first approach for the less knowledgeable students for whom element interactivity was high but an advantage to the investigation first approach for the more knowledgeable students for whom element interactivity was lower. The results were discussed in light of a recent debate between the authors and De Jong et al. ("Educational Research Review," 39, 1-14, 2023) on the role of investigations and explicit instruction in science education.