Educational trajectories and educational participation of children from immigrant and refugee families

In Germany, there is a wide range of high-quality research on the educational pathways of children with migration or refugee backgrounds. However, this research is often scattered and difficult for political or civil society actors to access. The project therefore aimed to develop a concept for a research database that systematically makes scientific results, data sets, and key findings available.

In contrast to conventional databases, it should be possible to search not only by keywords but also by content links. Supported by machine learning methods, users should be able to dynamically explore connections between concepts – such as interethnic friendships, gender, and prejudice.

The project thus served as methodological, technical, and conceptual groundwork for a possible prototype that could be transformed into an open research infrastructure in the long term.

The project responds to the lack of a way to automatically and dynamically capture thematic and conceptual connections in educational research. It thus closes a methodological gap between quantitative data collection and the practical application of research results for politics, science, and civil society.

• Development of a viable concept for a research database on educational pathways and educational inequalities.
• Improvement of the visibility of existing research results.
• Identification of research gaps and future needs.
• Examination of technical possibilities for the application of machine learning methods for the conceptual networking of research topics.

A detailed concept for the development of a content-based research database was created.

The conceptual basis for an innovative database architecture is in place. It provides a systematic and technically sound foundation for recording and making research findings more efficiently accessible. In addition, thematic analyses on discrimination in schools and the inclusion of refugees in higher education are available. However, due to personnel restructuring at the DeZIM Institute, the technical prototype was not implemented. In addition, two exemplary research topics were analyzed in depth:

1. Repeating a grade in the context of discrimination in schools.
2. Structural opening of universities to refugees.

• Research funding should be focused more on digital infrastructure projects that link knowledge bases.
• Politicians and administrators need low-threshold access to scientific findings on educational inequalities.
• Cooperation between educational research, computer science, and practical partners should be expanded in order to make machine learning specifically usable for social science analyses.