Survey instruments beyond the German race taboo? A new open-ended approach to identifications with ethnic-racial labels

Quantitative research on ethnic-racial identities in Germany faces a fundamental challenge: while ethnic-racial terminology is increasingly established and debated in academic, activist, and political contexts, no shared terminology for ethnic-racial labels can be assumed among the general population. Most existing survey instruments rely on a limited set of predefined answer options, which risk narrowing the understanding of ethnic-racial labels and imposing categories that may not resonate with participants. This article addresses this gap in three steps: it (1) defines an analytical framework to analyze identifications with ethnic-racial labels in surveys, (2) applies this framework to seven established survey instruments and (3) introduces and empirically examines a new open-ended survey instrument that allows participants to describe their identification in their own terminology and syntax. Drawing on data from a large, randomly sampled general population survey in Germany (DeZIM.panel, N = 3,700), the open-ended responses were systematically categorized within this framework. Building on the concept of ethnic-racial labels, responses were analyzed in six dimensions: nation states, regions, migration status, explicit ethnic-racial labels, other aspects, and identifications with multiple labels. One of the main findings shows that only a small fraction of responses referred to explicit ethnic-racial identities (4%, excluding “European”) or migration status (3%), even though these dimensions were explicitly primed by the instrument. Instead, references to nation states (64%) and regions (14%) dominated the responses, with “German” emerging as the single most frequent label (61%). In addition, 7% of the responses contained combinations of the label “German” with other labels. Explicit ethnic-racial labels and German multiple labels were significantly more common among younger participants and among first- and second-generation individuals, while only weak effects were found for education and political engagement. This study demonstrates both the analytical potential and the challenges of deductively categorizing responses generated by an open-ended survey instrument.