Gleasure Rob, Saigot Maylis, Kanat Irfan
Department of Digitalization, Copenhagen Business School, Copenhagen, Denmark.
UQ Business School, The University of Queensland, Brisbane, Australia.
Affect Sci. 2024 Aug 8;5(4):389-404. doi: 10.1007/s42761-024-00254-0. eCollection 2024 Dec.
This study investigated the role of circadian rhythms in online information sharing. We gathered 416,914 posts from the social media platform X (formerly Twitter). We identified daily patterns of collective positive and negative affect in these posts, consistent with previous research on social media and circadian rhythms. We created predicted values of positive and negative affect for each post, based on the time a post was created. We then used these predicted values for each post to estimate that post's likelihood of being reshared. We controlled for a range of possible confounders, such as the actual positive and negative affect expressed in a specific post and the number of existing followers and previous posts of the user who created the post, as well as whether the post contained hashtags, mentions, and quotes. The results support a strong relationship between the predicted positive and negative affect of a post-based on circadian patterns of collective positive and negative affect-and the likelihood of a post being shared. We further examine seasonal changes and design a natural experiment, in which we compare patterns of positive and negative affect and information sharing before and after the clocks change, i.e., "spring forward" and "fall back." The results suggest that these daily collective patterns of positive and negative affect on social media are influenced, at least partly, by hormonal influences and not only collective daily routines.
本研究调查了昼夜节律在网络信息分享中的作用。我们从社交媒体平台X（前身为推特）收集了416,914条帖子。我们在这些帖子中识别出集体积极和消极情绪的日常模式，这与先前关于社交媒体和昼夜节律的研究一致。我们根据帖子发布的时间为每条帖子创建了积极和消极情绪的预测值。然后，我们使用每条帖子的这些预测值来估计该帖子被转发的可能性。我们控制了一系列可能的混杂因素，例如特定帖子中表达的实际积极和消极情绪、创建该帖子的用户的现有关注者数量和之前的帖子数量，以及该帖子是否包含主题标签、提及和引用。结果支持了基于集体积极和消极情绪的昼夜模式对帖子的预测积极和消极情绪与帖子被分享的可能性之间存在密切关系。我们进一步研究了季节变化，并设计了一项自然实验，在实验中我们比较了时钟变化前后（即“夏令时”和“冬令时”）的积极和消极情绪模式以及信息分享情况。结果表明，社交媒体上这些日常集体积极和消极情绪模式至少部分受到激素影响，而不仅仅是集体日常活动的影响。
