CHS: Small: Assessing the Role of Platform Algorithms in Shaping News Attention

这项研究将对主导人类对新闻信息关注的两大媒体平台进行一系列算法审计：谷歌和Facebook。超过一半的美国人使用这些在线平台接收新闻信息。但人们对这些信息中介的算法策展过程如何推动高质量新闻信息的公开曝光和突出性知之甚少。这个问题的普遍性和严重性还不清楚，真实的用户对这些信息的注意力模式也不清楚。更一般地说：提供哪些类型和来源的新闻并进行优先排序，这些主要平台的算法策划中是否代表了不同的观点？这个项目试图回答的首要问题是：平台算法如何塑造人类对新闻媒体关注的可用性、规模和质量？探讨这些问题将有助于深入了解这些平台在提供多样化和高质量信息方面的作用。这些贡献将推进计算新闻和通信领域，以及信息科学的方法论和信息内容传播，消费和调解的社会科学问题。 该项目将提供知识，这些知识将影响算法策展系统的使用、设计甚至治理，以调解和塑造对个人和公民重要性的关注，从而帮助改善公众获得多样化和高质量的新闻信息。在Google的“Top stories”部分和Facebook的“Trends”部分专门出现的新闻文章URL将被自动收集并使用众包方法，以便可以测量新闻来源多样性，质量和关注度等因素，因为它们与个性化，地点和时间性等中介因素有关并随其变化。将得出各种衡量标准，以便对所出现的信息进行基准测试，并将其纳入更广泛的媒体可用性范围。此外，来自行业新闻指标合作者的数据的创新婚姻将使人们能够彻底了解算法策展如何大规模影响注意力模式。从这些研究中获得的见解可能会导致为平台或最终用户确定设计机会，以改善和填补围绕个人或公民重要性问题的新闻信息的多样性或质量方面的差距。 该项目的主要智力贡献是（1）开发可重复的算法审计方法，可以在现在和未来部署，以及（2）应用这些方法来开发算法驱动的信息中介如何影响新闻信息曝光和注意力模式的新知识。

This research will conduct a series of algorithm audits of two major media platforms that dominate human attention to news information: Google and Facebook. Online platforms such as these are used by more than half of Americans for receiving news information. But little is known about how the algorithmic curation processes of these information intermediaries serve to drive public exposure and salience of quality news information. The prevalence and magnitude of the problem is unclear as are the attention patterns of real users around such information. More generally: What types and sources of news are made available and prioritized, and are there diverse perspectives represented in the algorithmic curation of these major platforms? The overriding question this project seeks to answer is: How do platform algorithms shape the availability, magnitude, and quality of human attention to news media? Answering these questions will provide key insights into the role of these platforms in providing diverse and quality information. These contributions will advance the fields of computational journalism and communication, and information science both methodologically and in relation to social scientific questions of information content dissemination, consumption, and mediation. This project will provide knowledge that will impact the use, design, and perhaps even governance of algorithmic curation systems in mediating and shaping attention of personal and civic importance, thus helping to improve the public's access to diverse and quality news information. News article URLs surfaced specifically on the "Top stories" section of Google and the "Trends" section of Facebook will be collected both automatically and using crowdsourcing methods so that factors such as news source diversity, quality, and attention can be measured as they relate to and vary with mediating considerations such as personalization, locality, and temporality. A variety of metrics will be derived in order to benchmark the information surfaced and to put it into a broader context of media availability. Moreover, an innovative marriage of data from an industry news metrics collaborator will enable a transformative understanding of how algorithmic curation affects attention patterns at scale. Insights from these studies may lead to the identification of design opportunities for platforms or end-users to improve and fill gaps in the diversity or quality of news information available around issues of personal or civic importance. The main intellectual contributions of this project are (1) to develop repeatable algorithm audit methods that can be deployed now and in the future, and (2) to apply those methods to develop new knowledge of how algorithmically-driven information intermediaries affect news information exposure and attention patterns.