This article focuses on message credibility and detection accuracy of fake and real news as represented on social media. We developed a deception detection paradigm for news headlines and conducted two online experiments to examine the extent to which people (1) perceive news headlines as credible, and (2) accurately distinguish fake and real news across three general topics (i.e., politics, science, and health). Both studies revealed that people often judged news headlines as fake, suggesting a deception-bias for news in social media. Across studies, we observed an average detection accuracy of approximately 51%, a level consistent with most research using this deception detection paradigm with equal lie-truth base-rates. Study 2 evaluated the effects of endorsement cues in social media (e.g., Facebook likes) on message credibility and detection accuracy. Results showed that headlines associated with a high number of Facebook likes increased credibility, thereby enhancing detection accuracy for real news but undermining accuracy for fake news. These studies introduce truth-default theory to the context of news credibility and advance our understanding of how biased processing of news information can impact detection accuracy with social media endorsement cues.

本文重点介绍社交媒体上代表的虚假和真实新闻的消息可信性和检测准确性。我们开发了一种用于新闻标题的欺骗检测范例，并进行了两个在线实验来检验人们（1）认为新闻标题可信的程度，以及（2）在三个一般主题（即政治，科学）中准确地区分假新闻和真实新闻和健康）。两项研究均表明，人们经常将新闻头条视为伪造，这暗示着社交媒体中新闻的欺骗性偏见。在所有研究中，我们观察到平均检测准确度约为51％，这一水平与大多数使用等骗子基准率的欺骗检测范式的研究一致。研究2评估了社交媒体认可代言的效果（例如，Facebook喜欢）关于消息的可信度和检测准确性。结果表明，与大量Facebook相关的头条新闻增加了公信力，从而提高了对真实新闻的检测准确性，但破坏了对虚假新闻的准确性。这些研究将真相默认理论引入了新闻可信度的背景，并加深了我们对有偏见的新闻信息处理如何通过社交媒体认可线索影响检测准确性的理解。