Examine the Survey Bias Caused by Local Warming Effect

The survey method is a common tool in environmental social science research, used to widely collect public or participant perspectives on environmental issues, and analyze variables such as attitudes, backgrounds, and actions. However, the reliability and validity of the survey method have long been challenged. For instance, self-reported questionnaires often lead to self-enhancement bias, and recalling historical experiences may rely on the availability heuristic, enhancing the influence of recent events.

Previous studies have proposed the local warming (or weather) effect, which suggests that there will be response bias caused by current weather conditions, such as the recent temperature when surveying, which can influence the beliefs and risk perceptions about climate change. However, these kinds of responses are unstable, especially when emotion takes place; they fail to predict long-term actions or habits. This will cause overlooked research, potentially leading to exaggerated claims of effect sizes.

To test the bias caused by the local warming effect, this study utilizes data from the 2020 Taiwan Social Change Survey: Environment, which surveyed over 2,000 residents across Taiwan about environmental issues. The survey recorded participants' administrative districts and interview times. We selected variables that might be influenced by temperature, including environmental concern (from Protection Motivation Theory), environmental justice (from the Norm Activation Model), temporal distance and spatial distances (from Construal Level Theory), and high and low-cost environmental action willingness (from Low-Cost Hypothesis). We examined the regression relationships between these variables and absolute temperatures and temperature anomalies over 3-day/1-week/1-month/1-year before surveying.

The results indicate that Environmental concern is influenced by absolute temperatures across all time scales and temperature variations over one week to one month. Environmental justice is affected by absolute temperatures within a month and 1-week~1-month temperature anomaly. Temporal distance is positively impacted under all temperature scenarios, while spatial distances are influenced by absolute temperatures within a month.

Regarding environmental actions, both high-cost and low-cost actions are influenced by absolute temperatures within a month, and 1-month ~ 1-year temperature anomaly. Mediation analysis reveals that 3-day absolute temperatures influence environmental action willingness through the mediating of environmental justice and environmental concern. On the other hand, the mediation effect of environmental concern does not appear under 1-week ~ 1-year absolute temperatures.

To confirm that temperature only induces changes temporally, we also examined questions focusing on past habitual behaviors. Results show that environmental information browsing is influenced only by monthly to yearly temperature scales, while environmental actions are only affected by yearly temperatures.

In conclusion, our findings suggest that responses obtained through questionnaires are significantly influenced by recent weather and may not fully reflect participants' long-term stable conditions. However, if temperature anomalies are long enough, it still has the potential to affect environmental habits.