Students experience many difficulties learning the fundamental relationships in Newtonian mechanics, partly due to preexisting mental models that originate from their everyday lives. These preconceptions often persist even after instruction in mechanics and lead to a supposed incompatibility between physics lessons in school and personal experiences. This article presents research that focuses on students' concept knowledge and conceptual understanding in the field of Newtonian dynamics. Interventions are presented that discuss authentic experiments with friction using different software as a tool. Overall, the interventions lead to an increase in conceptual understanding with a large effect size. A hierarchical linear model (HLM) is used to identify factors that lead to a larger or smaller increase in conceptual understanding of the topic. With the HLM, we infer that the pretest score, the students' interest in how physical properties interrelate (interest in "theoretical physics"), prior physics performance (in kinematics in this case, measured in the physics grade), and the (intrinsic and extraneous) cognitive load influence the post-test score. We discuss possible explanations as to why these factors influence students' conceptual change.