Changing minds about climate change: a pervasive role for domain-general metacognition

Climate change poses a significant existential threat. While scientific consensus exists on its causes and impacts, significant public skepticism remains, hindering effective action. This skepticism is often politically and ideologically charged, impacting attitudes and behaviors. Past research indicates that individuals tend to update beliefs more readily with information confirming pre-existing views (confirmation bias), and that metacognitive abilities (the ability to accurately assess the correctness of one's own beliefs) play a crucial role in belief formation and polarization. Impaired metacognitive sensitivity, or poor insight into the accuracy of one's beliefs, has been linked to the holding of extreme beliefs. While previous work has explored the relationship between metacognition and socio-political beliefs, including climate change beliefs, the mechanisms by which metacognition influences belief updating in relation to climate change remain less understood. This study aimed to investigate the role of domain-general metacognitive abilities (measured using a perceptual task unrelated to climate change) and domain-specific metacognitive abilities (measured using a climate change knowledge test) in influencing the updating of climate-related beliefs, particularly in response to good and bad news about climate change. The study also explored whether the way climate science is communicated (through text only versus text and visuals) affects the updating process.

Existing research highlights the influence of confirmation bias on belief updating about climate change; individuals are more likely to accept information that supports their preconceived notions. Studies emphasize the role of metacognitive abilities, particularly domain-specific metacognition (knowledge about one's own knowledge) in shaping climate change beliefs. Research suggests that dogmatic individuals are less efficient in utilizing internal confidence signals for information seeking, limiting their ability to update beliefs. However, the question of how to effectively change these beliefs and the role of metacognition in this process is underexplored. Previous work demonstrates that both domain-general and domain-specific metacognition explain variations in socio-political beliefs, including beliefs about climate change. However, the crucial question of how to modify these beliefs and the role of metacognition in this process needs further investigation. The current study builds upon this foundation by specifically examining the interplay between domain-general metacognition and belief updating about climate change.

Two online studies were conducted using the Gorilla Experiment Builder platform and Amazon Mechanical Turk, recruiting US participants (Study 1: n=364; Study 2: n=335). Ethical approval was obtained, and informed consent was secured from all participants. The studies followed a consistent design, differing only in the belief-updating task's modality (Study 1: textual only; Study 2: visuo-textual). **Part 1: Perceptual Metacognition Task:** Participants completed a perceptual decision-making task adapted from Rollwage et al. (2018) to assess domain-general metacognitive abilities. This involved a calibration phase (determining individual task difficulty) and a confidence task (80 trials), requiring participants to identify which of two squares contained more flickering dots and rate their confidence in their decision. Metacognitive sensitivity, confidence bias, and perceptual accuracy were calculated. **Part 2: Climate Change Belief Updating Task:** Participants received initial information about either a temperature increase or sea-level rise scenario, then made initial estimates and confidence ratings. They were then presented with either 'good' or 'bad' news revising the initial information. They then updated their estimates and confidence levels. Study 2 included maps to visually represent the information. **Part 3: Questionnaires:** Participants completed questionnaires assessing climate change knowledge (Fischer et al., 2019; Sundblad et al., 2009), climate change skepticism (Capstick and Pidgeon, 2014), dogmatism (Altemeyer, 2002), political orientation (Rollwage et al., 2018), and social dominance orientation (Ho et al., 2015). Domain-specific metacognitive abilities (accuracy, confidence, and metacognitive sensitivity) were computed from their climate change knowledge responses. **Data Analysis:** Analyses used linear mixed-effects models and structural equation modeling (SEM) to investigate the relationships between metacognitive abilities, climate change skepticism, and belief updating. Correlations were calculated to examine the relationships between variables.

Study 1 revealed that greater climate change skepticism was associated with less belief updating following bad news, but more updating following good news (confirmation bias). Dogmatism negatively predicted belief updating following bad news. Study 2 replicated these findings, further showing that participants were less likely to update beliefs when presented with positive evidence via visuo-textual information. In both studies, higher domain-general metacognitive sensitivity (better insight into the correctness of perceptual judgments) predicted greater belief updating to bad news, controlling for task performance. However, higher confidence in perceptual judgments negatively predicted updating to bad news. Political conservatism and dogmatism also predicted less updating to bad news. Study 2 further showed that higher accuracy in the perceptual task positively predicted updating to bad news when presented with visuo-textual information. Regarding domain-specific metacognition, Study 1 revealed that higher confidence in climate change knowledge negatively predicted updating to bad news. There was an interaction showing that more accurate climate change knowledge was associated with greater belief updating to good news. Study 2 replicated the overconfidence effect. Domain-specific metacognitive sensitivity predicted greater belief updating to bad news. Path analyses (SEM) indicated that both domain-general and domain-specific metacognitive sensitivity predicted climate change attitudes. Lower metacognitive sensitivity predicted greater skepticism, which in turn predicted less belief updating. Confidence bias negatively predicted belief updating. The mode of communication (textual vs. visuo-textual) did not significantly impact belief updating.

The findings demonstrate a significant link between metacognitive abilities (both domain-general and domain-specific) and climate change skepticism, influencing how individuals update their beliefs in response to new evidence. Lower metacognitive sensitivity—the inability to accurately assess the correctness of one's own judgments—leads to greater climate change skepticism and less willingness to update beliefs in the face of contradictory information. Confidence bias, independent of metacognitive sensitivity, also plays a role; high confidence, even if misplaced, can hinder belief updating. The study highlights the importance of considering both domain-general and domain-specific metacognitive abilities when understanding and attempting to modify climate change-related beliefs. The relatively large variance explained (39% in Study 1, 34% in Study 2) by metacognitive characteristics underscores their crucial role in shaping socio-political attitudes.

This research confirms and extends previous findings on the interplay between climate change skepticism and belief updating, highlighting the crucial role of domain-general metacognition. Individuals with lower metacognitive sensitivity, particularly when overconfident, are less likely to update their beliefs, especially in response to negative evidence. Domain-specific metacognition also plays a significant role. Future research should investigate culturally diverse samples and explore interventions to enhance metacognitive abilities to improve climate-related belief updating. This could potentially aid in fostering mutual understanding and informed decision-making.

While the studies used large samples, they focused exclusively on US participants, limiting the generalizability of the findings to other cultural contexts. The study did not find a significant effect of the mode of communication (textual vs. visuo-textual), however, future research could more thoroughly explore the influence of visual aids in promoting climate change awareness, considering the potentially stronger emotional impact of images compared to text.