📌 What Problem This Paper Tackles:
Political experiments that aim to test how specific emotions shape behavior often produce collateral shifts in other emotions. Those collateral effects make it hard to attribute outcomes to a single emotion. This letter proposes using causal mediation analysis to separate the direct and indirect pathways linking emotion manipulations to political behavior.
🔬 How the Study Was Designed:
- An experiment manipulated emotions and measured dyadic trust using the standard trust game.
- Causal mediation analysis was applied to partition the total effect of emotion manipulations into effects transmitted through specific emotional states (mediators) versus other channels.
📈 Key Findings:
- Negative emotions can reduce trusting behavior, but this decline occurs only when the negative emotion lowers people’s certainty about their situation.
- Anxiety, characterized by low certainty, produced a negative impact on trust.
- Anger and guilt, which differ in control appraisals but both involve high certainty, showed no effect on trusting behavior.
- Importantly, analyses that do not use causal mediation would have mistakenly ascribed a positive effect of anxiety on trust, illustrating how standard approaches can mischaracterize emotional effects.
💡 Why This Matters:
- Demonstrates that emotional valence alone is insufficient to predict political behavior; the certainty dimension of emotions can be decisive.
- Shows causal mediation analysis as a practical tool for disentangling targeted emotional effects from collateral emotional changes in experiments, improving causal inference about how emotions shape trust and other political behaviors.
