Newtons laws of motion are known to be difficult to understand for students. Many persistent preconceptions exist, that are incompatible to the physical theory. Those preconceptions hinder the progress in conceptual understanding of Newtons laws. One reason for the persistence of those conceptions is the focus on idealized situation is physics class. Students have difficulties with those situations and a gap between physics class and the real world can arise. Complex motions from everyday life are too mathematically difficult for physics lessons in school. The computer as a tool can help to bridge that gap. This pre-post-design study with = 274 students from 11 th grade of German (Hessian) schools investigates two different ways of using the computer to discuss complex motions with friction regarding their efficacy in improving the conceptual understanding of Newtons first two laws after the school lessons. For that, two equivalent interventions were designed that only differ in the way the computer is used. In each intervention, four different experiments are discussed. The students always work in groups of two after the respective experiment is shown in the plenum. The test, which was used to measure conceptual understanding and other variables, is partly based on known tests. Other parts of the test were created and piloted for this study. Additionally, screen recordings in combination with the conversations of = 45 students were used to investigate, among other things, the way students use the software.A significant difference with large effect size between conceptual understanding in the pre-test and post-test was found in both cohorts. A comparison showed no difference between computational modelling and video motion analysis in the overall conceptual understanding. However, a significant difference in items regarding Newtons first law was found. The computational modelling cohort improved significantly more. It was also found that the common preconception, that a force in the direction of motion is necessary, was reduced further in the computational modelling cohort. An analysis of the screen recordings reveals that this preconception is expressed more often in that group as well. This leads to the conclusion that modelling forces students to activate the preconception which is then disproved by the comparison of the model and the data in the software. This cognitive conflict leads to a bigger reduction of the preconception which in turn leads to a larger improvement in the items regarding Newtons first law. The computational modelling cohort also improves in model understanding. More differences (e.g. cognitive load and different affective variables) were found and are being discussed.