How survey design influences people's understanding of democracy

While it is well established that closed-ended survey questions can prime respondents, less is known about how such priming affects the content of subsequent open-ended responses. We investigate this effect in the context of different understandings/perceptions of the term democracy using a survey experiment conducted in Germany in December 2023. Respondents (N = 3,123) were randomly assigned to a control or treatment group: the control group answered an open-ended question about what democracy meant to them first, while the treatment group first responded to closed-ended questions about democratic decision-making before also answering the open-ended question. We then compare the open-ended responses across groups. Using computational text analysis, we test three hypotheses regarding priming effects on (1) response length, (2) lexical overlap with vocabulary from the closedended question, and (3) thematic emphasis. Our findings indicate that while response length merely differs slightly between groups, primed participants are more likely to incorporate vocabulary from the closed-ended questions and to emphasize topics related to democratic decision-making. These results highlight how subtle differences in question order can meaningfully influence the way individuals express their understanding of democracy.