Pandemic communication Tweets from US State Health Agencies with Labels from Health Communication Experts

Pandemic communication texts from US state public health agencies on Twitter, in English, from 2020. Model training data includes binary labels indicating the text passes or fails individual items on the CDC Clear Communication, from the point of view of one trained expert rater. Model validation data includes consensus labels, from the point of view of two separate trained expert raters, with conflicts resolved by a third rater.

本数据集收录2020年美国各州公共卫生机构在Twitter平台发布的英文疫情沟通文本。模型训练数据附带二元标签，由一名经过培训的专业评估员依据美国疾病控制与预防中心（Centers for Disease Control and Prevention, CDC）的清晰沟通规范单项要求，标注对应文本是否达标。模型验证数据采用共识标签：由两名独立的经过培训的专业评估员分别标注，若二者出现标注分歧，则由第三名评估员协调解决。