From Narrative Interviews to Quantitative Data

This research project aims to thoroughly investigate the multifaceted concept of integration among refugees who have come to Germany since 2015 and to shed light on how integration experiences affect their social and personal lives. To this end, it employs innovative approaches that combine computer-assisted analytical methods such as machine learning with qualitative approaches (e.g., grounded theory).

The goal is to understand integration from the perspective of refugees themselves by exploring their perceptions, social dynamics, and personal transformations, while also contributing to the integration of different methodological approaches that allow research subjects to speak for themselves.

To achieve these objectives, the project focuses on the use of a large dataset, ideally collected interactively through semi-structured interviews conducted via a chatbot.

Care is taken to maximize both sample size and the efficiency of data collection by employing suitable technologies and methods such as machine learning and natural language processing for text analysis. Through these innovative approaches, the project seeks to provide comprehensive insights into the complex concept of refugee integration in Germany while also expanding the methodological repertoire of integration research.

The project addresses several key research gaps: there are still few empirical studies that capture integration from the perspective of refugees themselves and systematically examine their subjective meanings, emotions, and experiences. Existing research often treats integration as a political or structural objective rather than as a life-changing and temporally dynamic process.

Furthermore, the project addresses a methodological gap by combining qualitative approaches with technology-based methods such as machine learning and chatbot-supported interviews. This creates a new way to link large-scale data with the depth of individual narratives, thereby expanding the methodological toolkit of integration research.

The project aims to understand integration from the perspective of refugees themselves and to conceptualize it not as a political program but as an individual, social, and emotional process. It seeks to make visible how refugees experience, negotiate, and shape integration, and what personal transformations, conflicts, and new forms of belonging emerge from it. At the same time, the project aims to develop innovative methodological approaches that combine qualitative research with computational methods such as machine learning and chatbot interviews. In this way, it contributes to a deeper, empirically grounded understanding of integration processes and to the advancement of social science methods.

The project’s approach is based on a multi-stage, mixed-method research design that combines qualitative depth with technological efficiency.

First, narrative data are collected through open-ended and complementary structured questions that invite refugees to share their experiences as personal stories. These narratives shed light on the moral, religious, ideological, and emotional dimensions of integration. To ensure diversity and inclusion, the questions are offered in multiple languages and are formulated in a culturally sensitive manner.

For data collection, a chatbot is used to conduct semi-structured interviews, thereby efficiently reaching a large and heterogeneous sample. If a fully functional chatbot is not available, alternative data collection methods such as simplified versions or web-based questionnaires are employed.

Data analysis is carried out using machine learning and natural language processing techniques to identify thematic patterns, emotional tendencies, and integration trajectories.

First, the findings show that refugees define integration in highly diverse ways—as an opportunity, an obligation, a pressure to adapt, or a process of mutual learning. This diversity highlights that integration is not a uniform goal, but a multifaceted, subjective, and evolving experiential process.

Second, the study identifies social and personal patterns of integration that strongly depend on how refugees themselves understand integration. Typical dynamics emerge—for example, the desire for belonging alongside the fear of alienation from one’s culture of origin.

Third, the project demonstrates that technology-based methods such as chatbot interviews and machine learning can generate high-quality, rich data. They enable the efficient analysis of large datasets while still making individual voices and narratives visible.

Finally, the study contributes to expanding the methodological repertoire of integration research by showing how qualitative and quantitative approaches can be productively combined to empirically capture the complexity of integration from the perspective of refugees.