Supplementary information files for Russian meddling in U.S. elections: How news of disinformation’s impact can affect trust in electoral outcomes and satisfaction with democracy

Supplementary files for article Russian meddling in U.S. elections: How news of disinformation’s impact can affect trust in electoral outcomes and satisfaction with democracy

Russia’s Internet Research Agency (R-IRA) has been a key focus of disinformation research due to its attempts to use social media to influence the outcome of the 2016 United States presidential election. However, questions remain about the extent to which news coverage of the R-IRA’s efforts may have shaped public perceptions of US democracy. To assess its impact, we ran an experiment involving US social media users (N = 916). We tested whether reading news reports about the R-IRA’s activities heightened perceptions that the R-IRA influenced the public’s vote choices, and whether this influence in turn reduced confidence in the outcomes of the 2016 and 2020 elections and broader satisfaction with democracy. Specifically, we tested if these indirect effects differ depending on whether the R-IRA activity was presented via news frames conveying certainty or uncertainty about the R-IRA’s impact on the US public’s behavior. While the news frames did not significantly influence perceptions that the R-IRA had influenced the US public in general, the degree of certainty with which they described the effects of the R-IRA differently affected perceptions that Republicans and Democrats had been influenced. This, in turn, influenced participants’ confidence in elections and satisfaction with democracy.

附加文件：关于俄罗斯干预美国选举的研究——虚假信息影响新闻如何影响对选举结果及对民主满意度之信任

俄罗斯互联网研究机构（R-IRA）因其试图利用社交媒体影响2016年美国总统选举结果而成为虚假信息研究的焦点。然而，关于对R-IRA行动的新闻报道在多大程度上塑造了公众对美国民主的印象，仍存在疑问。为了评估其影响，我们进行了一项涉及美国社交媒体用户（N = 916）的实验。我们测试了阅读关于R-IRA行动的新闻报道是否加剧了公众认为R-IRA影响了投票选择的认识，以及这种影响是否反过来降低了对于2016年和2020年选举结果及对民主的总体满意度的信心。具体而言，我们测试了这些间接效应是否因R-IRA活动是通过传递对R-IRA对美国公众行为影响的确定性或不确定性新闻框架呈现而有所不同。尽管新闻框架并未显著影响公众认为R-IRA影响了美国公众的一般印象，但它们描述R-IRA影响的确定性程度不同地影响了共和党和民主党是否受到影响的感知。这进而影响了参与者对选举的信心和对民主的满意度。

其中，'Transformer'指Transformer，'Token'指Token，'LLM/Large Language Model'指大语言模型，'Zero-shot'指零样本，'Few-shot'指少样本，'AI Agent'指AI智能体，'AGI'指通用人工智能。