Inquiry learning is a widely recognized method for fostering inquiry competence in science education. Nevertheless, there is discussion about how to best support students while working on inquiry tasks (in this case: experiments on causal relationships). To identify the kind of support students need in order to design experiments in upper grades, an empirical study consisting of two parts was conducted. The first part, a quantitative test (N?=?96), tested students' competences in designing an experiment. It was found that students need scaffolding in basic aspects of designing (like dependent and independent variables), to reach higher levels, since only about thirty percent reach levels II or III out of three levels. On more reflective aspects (like confounding variables, test times and repetitions), only a maximum of twenty-five percent reached level I or higher; therefore, even more support is needed on these aspects. The second part of the study was a qualitative video-analysis of students' discussions, designs and implementations of an experiment. It was found that support is needed on procedural knowledge and understanding. Possible implications for inquiry support are discussed.