The uncertain geographic context problem and the unknown effect of extreme weather on climate change attitudes

Are people exposed to extreme weather events more likely to support climate change policies? Previous research demonstrated that humans can perceive weather anomalies and extremes, attribute them to climate change, and call for more political action. However, systematic reviews also indicated that this association may only be slight to moderate. Moreover, some studies found only spurious correlations, and others only effects among people with specific political convictions. Hence, research yields mixed results, attributing the inconsistencies to theoretical or methodological reasons. This paper contends that theoretical and methodological issues are deeply intertwined in the spatio-temporal analysis of individual-level data. Researchers face challenges in specifying the temporal scale and the spatial reach, e.g., of survey respondents’ experience with extreme weather, known in geography as the uncertain geographic context proble. Our study employs georeferenced survey data from Germany on attitudes toward climate change linked with highly customizable weather data. Within theoretical guardrails, we explore an extensive range of spatio-temporal parameters for exposure to extreme weather. Our results show small effect sizes across certain variable specifications, with inconsistent and sometimes conflicting results. The estimates are much more sensitive to temporal rather than spatial variations, highlighting the impact of spatio-temporal modeling decisions on applied social science research.