RAPID: Agenda Generality and Behavior in Social Network Interactions about COVID-19

传播新冠肺炎疾病的同一类型的社交网络可能被用来传播健康规范和积极行为。这项研究收集了重要的、时间敏感的数据，以了解数字社交网络在什么条件下可以影响与新冠肺炎大流行相关的健康行为，以及如何减少数字环境中的负面社会影响。在人们花费史无前例的时间在数字网络上的时候，部署在这些网络上的公共卫生战略可能会在未来12个月内塑造美国人的健康和社会结果。这项研究促进了对这些公共健康战略的理解。该项目的理论是，无论是讨论一般问题还是具体问题(例如，遏制新冠肺炎疾病或戴口罩)都可能对危险态度和行为通过网络传播产生重要影响。这项研究包括(A)对Twitter和Instagram网络进行生态研究，(B)对网络中提倡的健康和危险行为的组合进行实验，并将讨论的重点放在一般或具体问题上。该项目生成公共健康建议和算法，以改善社交媒体上的健康讨论。研究人员使用动态面板数据模型，根据个人的态度和过去的行为以及他们网络中其他成员的行为来预测个人行为。研究团队使用图形卷积网络来捕捉更丰富的网络方面，并对稀疏网络进行建模。这一奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

The same types of social networks that transmit the COVID-19 disease may be leveraged to spread healthy norms and positive behaviors. This research gathers important, time-sensitive data to understand the conditions under which digital social networks can influence health behaviors relevant to the COVID-19 pandemic and how to reduce negative social influences in digital environments. At a time when people spend unprecedented amounts of time on digital networks, public health strategies deployed in these networks may shape the health and social outcomes of Americans in the next 12 months. This research advances understanding of these public health strategies.The project’s theory is that discussing either general or specific issues (e.g., curbing COVID-19 disease or wearing a mask) can have important consequences on the spread of risky attitudes and behaviors through a network. The research entails (a) an ecological study of Twitter and Instagram networks and (b) experiments manipulating the mix of healthy and risky behaviors promoted in the network and the focus of the discussion on either general or specific issues. The project generates public health recommendations and algorithms to improve health discussions on social media. The investigators use a dynamic panel data model to predict individual behavior from the individual’s own attitudes and own past behaviors as well as the behaviors of other members of their network. The research team uses graph convolutional networks both to capture richer network aspects and to model sparse networks.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.