Search

Writing science effectively requires a big amount of practice, even among experts. Yet, reading scientific texts or those on scientific writing is not enough; for scientific argumentation, evidence-based writing, and high-quality content creation are essential writing processes that students must both learn and practice. In previous work on supporting the process of scientific writing, we introduced a modelling language that visualizes the components and their relations that are found in scientific texts. By doing so, for example, not only the production but also the communication of (scientific) content can be improved. We also proposed several text patterns that are represented in the modelling language and that describe proven structures for scientific texts. The classification of the patterns in a pattern language allows for a sound structured approach during the entire design and writing processes, and was based on diverse practical experiences we have collected over the years.This paper examines how the language for text modelling, as well as exemplary text patterns, are used and evaluated by students in different supervision contexts. First, overviews of both the notational elements and the selected patterns are provided. This allows for a straightforward learning of the methodology. Second, the learners create an individual text model as a first draft of, e.g., their final thesis. Finally, a systematic survey is administered in order to assess the methodology, reflect on its pros and cons, detect areas with lack of support, and suggest the next steps of the writing process. Early results show that the students find the formal representation of texts for detailed drafting useful. Although the additional effort is perceived as critical, the respondents also acknowledge the increasing quality of the texts they produce. The pattern language provides useful initial indications for the various phases of scientific writing, from a first draft to the detailed planning of individual text sections.