Understanding Survey Participation and Measurement Bias in Dynamic Political Environments: Evidence From a Natural Experiment

Unbiased mirrors of public opinion are crucial for understanding political attitudes and behaviors, especially in a context where political events are expected to sway public opinion. This paper delves into how a changing political context can not only affect political attitudes, but also people's willingness to participate in surveys and, thus, the accuracy of public opinion measurements. Specifically, I zoom in on potential biases in the 'rally around the flag' effect following terrorist attacks. Using two terrorist attacks in Paris in 2015 that fall into the fieldtime of European Social Survey and Eurobarometer survey waves as natural experiments, I investigate how these events affect who responds to surveys. I explore how they might skew measures of national leadership approval, isolating whether increases in support are influenced by changes in survey participation rather than genuine shifts in public sentiment. The results indicate that while these events render people with high educational attainment, political sophistication and more right-leaning ideological stances systematically more responsive, these patterns do not translate into a substantive bias in measures of the 'rally around the flag' effect. This study provides valuable insights into the nexus of survey research and its political context, offering methodological lessons for interpreting survey and public opinion data in dynamic political environments.