The COVID-19 pandemic triggered not only a global health crisis but also an infodemic, where exposure to heterogeneous information sources influenced public emotional responses. In this work, we investigate the determinants of self-reported fear of infection using data from the Delphi US CTIS survey. In particular, we analyze how demographic variables, epidemiological conditions, and exposure to different information sources shape fear levels. We introduce a Probabilistic Causal Model to estimate causal relationship strengths, identifying the variables that most strongly influence fear. Our results indicate that exposure to information sources accounts for a greater proportion of the variance in fear than demographic and epidemiological variables do. We further compute the Average Treatment Effect to quantify the impact of different information sources on fear. After causal adjustment, institutional and expert-driven sources are associated with increased fear levels, whereas politicians, religious leaders, and alternative information channels are associated with reduced fear. These findings highlight both the central role of the information ecosystem in shaping emotional responses during public health crises and the value of causal inference approaches for studying behavioral responses to pandemics.

翻译：新冠肺炎疫情不仅引发了全球健康危机，还导致了信息疫情，其中暴露于异质性信息源影响了公众的情绪反应。本研究利用德尔菲美国社区趋势调查（CTIS）数据，探讨了自报感染恐惧的决定因素。具体而言，我们分析了人口变量、流行病学状况以及不同信息源的暴露如何塑造恐惧水平。我们引入了一个概率因果模型来估计因果关系的强度，识别出对恐惧影响最强的变量。结果表明，信息源暴露对恐惧差异的解释比例高于人口变量和流行病学变量。我们进一步计算了平均治疗效果，以量化不同信息源对恐惧的影响。经过因果调整后，机构及专家驱动的信息源与恐惧水平上升相关，而政治家、宗教领袖及另类信息渠道则与恐惧水平下降相关。这些发现突显了信息生态系统在公共卫生危机期间塑造情绪反应的核心作用，以及因果推断方法在研究疫情下行为反应中的价值。