Social Media Reveals Urban-Rural Differences in Stress across China

Abstract

Modelling differential stress expressions in urban and rural regions in China can provide a better understanding of the effects of urbanization on psychological well-being in a country that has rapidly grown economically in the last two decades. Using Weibo posts from over 65,000 users across 329 counties, we build hierarchical mixed-effects models to understand linguistic differences in the experiences and expressions of stress in urban-rural China. We analyze phrases, topical themes, and psycho-linguistic word choices in Weibo posts mentioning stress to better understand appraisal differences surrounding psychological stress in urban and rural communities in China; we then compare them with large-scale polls from Gallup. After controlling for socioeconomic and gender differences, we find that rural communities tend to express stress in emotional and personal themes such as relationships, health, and opportunity while users in urban areas express stress using relative, temporal, and external themes such as work, politics, and economics. We find corroborative trends in physical, financial, and social wellness with urbanization in Gallup polls.

Lady and the Tramp Nextdoor: Online Manifestations of Real-World Inequalities in the Nextdoor Social Network

“…We avoid these problems by reducing the textual embedding with Uniform Manifold Approximation and Projection (UMAP) (McInnes et al 2018), the state-of-the-art method. We reduce the 768 dimensions to 5, as suggested in (Zhang et al 2022;Jimenez Villalonga 2021).…”

“…As prior work indicates that income and self-reported positive sentiment are usually correlated (Easterlin 1974;Giorgi et al 2021;Cui et al 2022), we now look at the relationship between income and post sentiment. We calculate the median of the sentiment scores for the posts of the neighborhoods in the same income group and each % of the duration of the month.…”

“…velop multiple features based on the analysis of the previous sections and investigate their predictive power. We first use UMAP for dimension reduction as prior work shows improvements in the performance (Zhang et al 2022;Jimenez Villalonga 2021): we reduce 768-dimensional mean-pooled textual embeddings to 5 dimensions. For feature selection, we employ a Random Forest (RF) as this is the best-performing method (Chen et al 2020).…”

Social Media Use and College Students Sleep Quality: An Analysis of Influencing Factors

“…In China, one of the world's most densely populated countries, a remarkable surge in social media usage, particularly among Chinese adolescents, has been observed, as Forbes' latest statistics [2] indicate that 4.8 billion people will be involved in the medium by 2023. In contrast to the West, where Facebook, Instagram, and YouTube reign supreme, numerous innovative social platforms are now being developed in China [3]. According to the latest figures from Weiboyi, WeChat is ranked as China's top essential social commerce platform, followed by Weibo, Little Red Book, and Douyin [4].…”