Questioning Vaccination Discourse (Quo VaDis): A Corpus-Based Study

Quo VaDis项目将最新的大规模计算机辅助语言分析技术应用于公共话语中关于疫苗接种的讨论，特别是在英语社交媒体讨论，英国议会辩论和英国国家新闻报道中。目标是更好地了解支持和反对接种疫苗的观点，以及尚未决定的观点，这将为未来的公共卫生运动提供信息。该项目将设在兰开斯特大学世界著名的ESRC社会科学语料库方法中心（CASS），该中心在2015年获得了女王周年纪念高等和继续教育奖。一个跨学科项目团队将与三个主要项目合作伙伴进行互动：英格兰公共卫生部、卫生和社会保障部以及数字、文化、媒体和体育部。世界卫生组织（WHO）列出的全球十大健康威胁包括“疫苗犹豫”——“尽管有疫苗接种服务，但在接受或拒绝疫苗方面出现延误”。疫苗接种规划目前估计每年可在全世界预防200万至300万人死亡。然而，据报告，90%的国家接种疫苗受到疫苗犹豫的影响。在英格兰，所有常规儿童疫苗接种的覆盖率正在下降，导致以前已被根除的传染病重新出现。2019年8月，英国失去了世卫组织消除麻疹的地位。疫苗犹豫的原因很复杂，但需要了解这些原因才能有效地加以解决。本项目侧重于话语，因为讨论疫苗接种等有争议话题的方式既反映又塑造了信念和态度，而信念和态度反过来又可能影响行为。更具体地说，自1840年第一个疫苗接种法案出台以来，疫苗接种一直是英国议会辩论的主题；自上世纪90年代初以来，英国媒体对它们的讨论越来越多；尤其是反对接种疫苗的观点，被描述为一个复杂的“反公共话语”网络的一部分，近年来，这一网络在社交媒体上传播和争议。该项目将涉及三个数百万字的数据集的分析：(1)三个社交媒体平台：Mumsnet， Reddit和Twitter自每个平台成立以来的英语贡献-分别为2000年，2005年和2006年；(2) 1990年以来的英国全国性报纸；(3) 1830年以来的英国议会辩论。这些数据集将通过与语料库语言学相关的计算机辅助方法以数据驱动的方式进行分析。语料库语言学是语言学的一个分支，涉及构建自然文本的大型数字集合（称为“语料库”），并通过量身定制的软件进行分析。语料库语言学方法可以将包含数百万单词的语料库的定量分析与单个文本、模式和相互作用的定性分析有机地结合起来。通过这种方式，我们将识别和调查关于疫苗接种的观点在我们的数据中表达的不同方式，例如，通过选择词汇、代词、否定、评价、隐喻、叙述、证据来源和论证的模式。我们将揭示随着时间的推移和不同人群之间支持和反对疫苗接种观点的差异和相似之处，特别是当他们在社交媒体上形成和互动时。我们的研究结果将为理解英国（通过我们的议会和新闻数据集）和国际（通过我们的社交媒体数据集）对疫苗接种的看法做出重大贡献。通过我们项目合作伙伴的参与，以及更广泛的参与活动，这些发现将被用作设计未来有关疫苗接种的公共卫生运动的证据。

The Quo VaDis project applies the latest techniques for large-scale computer-aided linguistic analysis to discussions about vaccinations in public discourse, and specifically in: social media discussions in English, UK Parliamentary debates and UK national press reports. The goal is to arrive at a better understanding of pro- and anti-vaccination views, as well as undecided views, which will inform future public health campaigns.The project will be based in the world-renowned ESRC Centre for Corpus Approaches to Social Science (CASS) at Lancaster University, which was awarded a Queen's Anniversary Prize for Higher and Further Education in 2015. An interdisciplinary project team will work in interaction with three main project partners: Public Health England, the Department of Health and Social Care and the Department for Digital, Culture, Media & Sport.The World Health Organization's (WHO) list of top ten global health threats includes 'vaccine hesitancy' - 'a delay in acceptance or refusal of vaccines despite availability of vaccination services'. Vaccination programmes are currently estimated to prevent between 2 and 3 million deaths a year worldwide. However, uptake of vaccinations in 90% of countries has been reported to be affected by vaccine hesitancy. In England, coverage for all routine childhood vaccinations is in decline, resulting in the resurgence of communicable diseases that had previously been eradicated. In August 2019, the UK lost its WHO measles elimination status. The reasons for vaccine hesitancy are complex, but they need to be understood in order to be addressed effectively. This project focuses on discourse because the ways in which controversial topics such as vaccinations are talked about both reflect and shape beliefs and attitudes, which may in turn influence behaviour. More specifically, vaccinations have been the topic of UK parliamentary debates since before the first Vaccination Act of 1840; they have been increasingly discussed in the UK press since the early 1990s; and anti-vaccination views in particular have been described as part of a complex network of 'anti-public discourses' which, in recent years, are known to be both spread and contested on social media.This project will involve the analysis of three multi-million-word datasets: (1) English-language contributions to three social media platforms: Mumsnet, Reddit and Twitter since the inception of each platform - respectively, 2000, 2005 and 2006; (2) UK national newspapers since 1990; and (3) UK parliamentary debates since 1830. These datasets will be analysed in a data-driven fashion by means of the computer-aided methods associated with Corpus Linguistics - a branch of Linguistics that involves the construction of large digital collections of naturally-occurring texts (known as 'corpora') and their analysis through tailor-made software. A corpus linguistic approach makes it possible to combine in a principled way the quantitative analysis of corpora containing millions of words with the qualitative analysis of individual texts, patterns and interactions. In this way, we will identify and investigate the different ways in which views about vaccinations are expressed in our data, for example, through patterns in choices of vocabulary, pronouns, negation, evaluation, metaphors, narratives, sources of evidence, and argumentation. We will reveal both differences and similarities in pro- and anti-vaccination views over time and across different groups of people, particularly as they form and interact on social media.Our findings will make a major contribution to an understanding of views about vaccinations both in the UK (via our parliamentary and news datasets) and internationally (via our social media datasets). Through the involvement of our Project Partners, as well as more general engagement activities, these findings will be used as evidence for the design of future public health campaigns about vaccinations.