The COVID-19 pandemic presented unprecedented public health challenges, causing not only a global health crisis but also a significant impact on emotional and social well-being. Analyzing public emotional reactions during crises is crucial for effective crisis management. Emotions like anxiety, anger, and sadness influence risk perception and behavior. While negative emotions like anxiety, anger, and sadness are expected during crises, mechanisms promoting positive emotions also exist, including self-efficacy and the pursuit of meaning. Existing research focuses primarily on the micro- or macro-level relationship between emotional responses and social connections during pandemics. This study aims to offer a more nuanced understanding by examining three specific social connections—family, collective, and country—across Chinese and American social media users during the COVID-19 pandemic, accounting for the prolonged and indeterminate nature of the crisis and the need for a longitudinal approach. A cross-cultural comparison is crucial due to the influence of cultural frameworks on emotional responses and emotion regulation strategies. The differing pandemic responses in the US (adaptive norm of "living with" COVID-19) and China (zero-clearance policy) highlight this necessity.

Previous research has established a link between emotional responses during pandemics and social connections. Jo et al. (2021) found that post-pandemic, robust advice and friendship ties led to decreased work-related interactions. Ashokkumar and Pennebaker (2021) observed shifts in social connections among urban U.S. residents, with increased emphasis on family connections. Studies show that self-efficacy fosters positive emotions, and that the search for meaning during crises is linked to positive emotions like hope and satisfaction. Collectivist cultures tend to emphasize communal harmony and belonging, potentially bolstering positive emotions. Existing research, however, largely lacks the cross-cultural, longitudinal investigation of multiple forms of social connection that this study provides.

This study leverages social media data from Weibo (China) and Reddit/X (U.S.) to examine emotional dynamics and social connections during the COVID-19 pandemic. The data were collected and analyzed over a longitudinal period, encompassing four phases of the pandemic: baseline, warning, isolation, and normalization. The researchers analyzed the frequency of emotion-related words and social connection terms (family, collective, country) in posts, employing a three-day rolling average to smooth the data. Topic analysis was conducted to investigate the predominant themes associated with different social connections in both countries. Statistical analyses (regression models) were used to examine the relationships between the number of new COVID-19 cases, emotional expressions (negative and positive), and the different types of social connections. Moderation analyses were performed to assess the interactive effects of new cases and social connections on emotions. Robustness checks involved comparing results from analyses of data from the largest 19 Chinese cities and the 18 largest US cities in the dataset.

The study revealed significant cross-cultural differences in emotional responses to the pandemic. In China, anxiety levels were consistently high, but anger and sadness were relatively low. Conversely, the U.S. showed peaks in anxiety, anger, and sadness during the pandemic's various phases. Positive emotions were high in China and low in the United States. The topic analysis showed distinct thematic differences in social connection discussions. In the U.S., family discussions often focused on relationships, education, and healthcare, while in China, family discussions centered on public health and the pandemic. Regarding collective connections, the U.S. focused on state politics and global affairs, whereas China emphasized encouragement and public health. Country-level discussions in the U.S. revolved around state politics and elections, while in China, they centered on encouragement and pandemic-related issues. Regression analyses demonstrated significant relationships between new COVID-19 cases and emotional expressions. Moderation analysis revealed that family connections in the U.S. and country connections in China buffered negative emotions during surges in cases. Interestingly, the effect of collective connections on emotions differed across cultures. In China, stronger collective connections amplified positive emotions; in the U.S., they lessened them. These findings were robust across different samples of cities.

The findings suggest that cultural values significantly shape emotional responses and social connection priorities during crises. The collectivist culture of China fostered higher anxiety but also a sense of collective responsibility and support, potentially leading to less pronounced sadness and anger. Conversely, the individualistic culture of the U.S. may have facilitated the expression of a wider range of negative emotions, with greater emphasis on individual experiences and struggles. The divergent responses to collective connections reflect cultural differences in trust and reliance on collective entities during crises. The findings underscore the importance of considering cultural contexts when designing crisis response strategies. Tailoring interventions to specific cultural values and social structures could significantly improve their effectiveness.

This study provides valuable insights into the cross-cultural differences in emotional responses and social connections during the COVID-19 pandemic. The findings highlight the crucial role of cultural values in shaping emotional experiences and prioritizing various social connections during crises. Future research should explore these dynamics further using larger datasets, incorporating additional variables like governmental trust and information transparency, and focusing on the nuances of positive emotions by applying advanced text analysis techniques such as machine learning and natural language processing. Moreover, incorporating other data sources, such as surveys, would enrich the understanding of emotional experiences and social interactions during the pandemic and help to mitigate the biases inherent in social media data alone.

The study primarily uses social media data, which may not represent the complete emotional landscape of the entire population. The sample may be biased towards certain demographics more active on social media. The analysis of positive emotions is relatively broad, not accounting for the diverse range of positive emotions that individuals might have experienced. Future studies could address these limitations by incorporating more diverse datasets, including survey data, and employing more sophisticated methods to analyze the nuances of emotional experiences.