How is scientific data used to represent phenomena and as evidence for claims about phenomena? In this paper, I propose that a specific type of claims – evidential claims – is involved in data practices to define and restrict the representational and evidential content of a dataset. I present an account of data practices in the epidemiology of the exposome based on the notion of evidential claims, which helps unpack the approaches, assumptions and warrants that connect different stages of research. I identify three different strategies to generate different types of evidential claims in this case. The macro strategy, which individuates the dataset that serves as the initial evidential space for research. The micro strategy, which is used to generate evidential claims about the microscopic and individual component of target phenomena. The association strategy, that uses evidence from the other strategies to identify a dataset as representation of the different levels and relations of exposure and disease. Differentiating between these strategies sheds light on the multi-faceted landscape of biomedical research on environment and health; and the roles of data and evidence in the process of inquiry.

科学数据如何用于表示现象和作为对现象的主张的证据？在本文中，我建议在数据实践中涉及一种特定类型的声明（证据声明），以定义和限制数据集的表示形式和证据内容。我将根据证据主张的概念介绍暴露流行病学中的数据实践，这有助于解开连接不同研究阶段的方法，假设和保证。在这种情况下，我确定了三种不同的策略来生成不同类型的证据主张。宏观策略，可以将用作初始证据空间的数据集个性化。微观策略，用于生成有关目标现象的微观和个别成分的证据。关联策略，使用来自其他策略的证据来识别数据集，以表示暴露水平与疾病的不同水平以及关系。这些策略之间的区别为生物医学研究环境与健康的多面性提供了启示。以及数据和证据在询问过程中的作用。