Digital technologies are changing forms of learning as we know them. Education policy makers are therefore calling for the increasing digitalisation of school education and the development of corresponding digital provision. In the German-speaking world, however, digital education is still in its infancy, with the most common applications more or less limited to multiple-choice tests and comparably simple tasks. A key component of school education – both across classes and subjects – is argumentative writing. “Well thought-out, modern forms of learning are particularly important here,” says Professor Henning Wachsmuth of Paderborn University. The computer scientist is heading up a new research project on e-learning, together with Professor Sara Rezat, also of Paderborn University. The researchers are investigating how algorithmic methods can help pupils acquire written argumentative skills. The project will be funded by the German Research Foundation (DFG) for a period of three years, starting in December, with approximately €540,000.
Identifying developmental levels
Wachsmuth explains: “The applications are designed to automatically analyse argumentative texts in order to provide feedback on which aspects are good and which can be expanded on and to make suggestions for improvement.” The methods focus on the structure of argumentative texts, assessing pupils’ developmental levels and providing feedback tailored to the relevant developmental level. The researchers see the studying of content and the relationship between content and sources as further work. The algorithms are intended to support both pupils and teachers.
Support for forming opinions and structuring arguments
“Addressing counter-positions is a key milestone in the development of argumentative writing. However, studies on acquisition consistently show that this proves a hurdle for many pupils. It’s in particular for this purpose that algorithmic methods are to be developed, which identify pupils’ own positions, justifications and counter-positions in texts and evaluate the relevant developmental level on the basis of this,” explains Rezat from the Department of German Studies and Comparative Literary Studies. “The results of these analyses will then be used as input for methods that generate developmental level-specific, pupil-sensitive feedback. For example, pupils could be made aware of the fact that their essay is lacking counter-positions and of passages in their text where these could be added,” says Wachsmuth. The computer scientist is also involved in another DFG research project, in which machine learning is leveraged to automatically and objectively summarise justifications in texts. The basis for this is the search engine ‘args.me’, developed by Wachsmuth, that has been online since 2017 and that compares pro and contra arguments on search topics to aid independent opinion formation.
Possibilities and limitations
As part of their research, the researchers are creating a German-language corpus, i.e. a digital text collection, of some 1,500 manually annotated argumentative learner texts from three age groups. This will also serve as a basis for further studies in the future and will be available to the public. In the qualitative evaluation, the methods will be combined with didactic knowledge to determine the possibilities and limitations of developmental level-based writing support, both technically and socially. Wachsmuth and Rezat expect to obtain the first meaningful results in the second half of 2022.
