Sage Journals: Discover world-class research

Abstract

The global COVID-19 pandemic affected numerous aspects of individual lives, and researchers are now turning to study the potential impacts of the pandemic on culture and values. Previous research has shown that infectious disease outbreaks tend to make groups more collectivist. The present study drew on micro-blog big data from Weibo to verify whether COVID-19 and related factors changed collectivism and individualism in China, while explaining abnormal data. Time-series models verified that COVID-19 significantly affected collectivism and individualism among the Chinese people, and the severity of the epidemic, epidemic-related policies, and discussion about the epidemic among Weibo users significantly explained variation in these levels. Analysis of Weibo text one day before and one day after the dates of statistical outliers in collectivism and individualism suggested that the changes in these levels were associated with significant events in the epidemic.

Keywords

COVID-19 collectivism individualism

Introduction

In April 2022—nearly three years after the outbreak of coronavirus disease 2019 (COVID-19)—the cumulative number of confirmed global cases of COVID-19 exceeded 490 million, and the total number of registered deaths from the virus surpassed six million. As an unprecedented global event, the COVID-19 pandemic greatly influenced peoples’ lifestyles, finances, politics, and cultural values. Much research on COVID-19 has focused on the psychological consequences of the pandemic (Lwin et al., 2020; Wang et al., 2020). However, only a few studies have examined the relationship between the pandemic and cultural values (Huang et al., 2020; Lu et al., 2021; Na et al., 2021).

The complementary cultural syndromes of collectivism and individualism are among the most significant indicators of cultural difference (Greenfield, 2000; Triandis, 1996, 2001). Collectivism prioritizes relationships and interdependence between people, emphasizes loyalty and obligation to in-groups (e.g., family, class, country), and views individuals as part of a collective. In contrast, individualism values autonomy, uniqueness, independence, and freedom of choice (Grossmann & Na, 2014; Oyserman et al., 2002a).

Culture is not a static entity; rather, it is dynamic and changeable (Ogihara, 2018), and it evolves in response to social transformations (e.g., changes in gender roles, urbanization, globalization, and technology; Lansford et al., 2021; Morris et al., 2015). Recent studies have examined factors (i.e., ecological, economic, historical) contributing to cross-cultural variation in individualism and collectivism. These include economic development (Hofstede, 1984), residential mobility (Oishi, 2010), language use (Kashima & Kashima, 2003), and the prevalence of pathogens (Fincher et al., 2008). In particular, pathogen prevalence has been proposed as an ecological foundation of collectivism (Fincher et al., 2008). Numerous studies have shown that societies with a higher prevalence of pathogens exhibit more pronounced pathogen-resistant psychological tendencies and customs, including an embrace of collectivism (Schaller & Murray, 2008; van Leeuwen et al., 2012). Indeed, research has shown that the threat of pathogens leads to an increase in collectivistic behaviors (Wu & Chang, 2012). However, only a few studies have explored changes in collectivism and individualism following the outbreak of an infectious disease. The COVID-19 pandemic provides a rare opportunity for such research. Accordingly, the present study examined changes in collectivism and individualism after the outbreak of COVID-19, and proposed possible explanatory factors. The results provide important information for the formulation of outbreak prevention measures, as well as the prediction of public sentiment and the management of public opinion.

Collectivism and individualism

The collectivism and individualism constructs

The collectivism–individualism spectrum has been a cornerstone of cross-cultural research for more than three decades (Oyserman & Lee, 2008). In a foundational work, Hofstede (1984) argued that an individual's placement on the spectrum depends on their closeness to the group. He further claimed that the placement of a population along this dimension represents one of the most important indicators of cultural variation. Since that time, the constructs of collectivism and individualism have undergone two significant modifications (Beilmann et al., 2018): first, they are now explored on an individual, as well as a cultural, level (Realo, 2003); and second, they are no longer exclusively considered opposite poles of a unidimensional factor, but increasingly understood as relatively independent factors (Realo et al., 2002; Trommsdorff et al., 2004). In fact, Triandis and Gelfand (1998) argued that collectivism and individualism may be horizontal (emphasizing equality) or vertical (emphasizing hierarchy).

Changes in collectivism and individualism

Collectivism and individualism, as values or cultural orientations, are not stable and unchangeable (Morris et al., 2015). Rather, the degree to which individuals and cultures embrace either of these values can change over time. Growth in the global economy, the development of science and technology, the continuous advancement of globalization, and extensive social change since the 1980s has brought about profound and extensive changes to cultural and individual values. Overall, research has shown that individualism has become more widespread, while collectivism has declined (Huang et al., 2018). These trends have been observed in both macro-social and -cultural indicators, as well as data referring to individual/group psychology and behavior.

In terms of macro-social and -cultural indicators, researchers in the United States (Grossmann & Varnum, 2015), Japan (Hamamura, 2012), and China (Huang et al., 2016) have found that changes in divorce, single-living, and multi-generation cohabitation rates, as well as changes in family size, reflect larger social changes in collectivism and individualism. In terms of individual/group psychology and behavior, changes in collectivism and individualism have been measured through the use of first-person plural versus first-person singular pronouns. Individualists tend to focus on the self, appreciate difference from others, and emphasize individual assertion; collectivists, in contrast, tend to construe the self as interdependent and emphasize harmonious interactions with others (Markus & Kitayama, 1991). Theoretically, first-person singular pronouns (i.e., “I,” “my”) are more related to individualism, while first-person plural pronouns (i.e., “we,” “our”) are more related to collectivism (Oyserman & Lee, 2008).

Changes in collectivism and individualism may also be observed in naming trends. Specifically, as individualistic cultures tend to pursue uniqueness as a core value, the uniqueness of names may be used as a measure of individualism. A study that analyzed the naming habits of Americans from 1880 to 2007 found that the proportion of newborns receiving popular names declined over the period, suggesting a rise in individualism (Twenge et al., 2010). Finally, changes in collectivism and individualism may be observed in cultural products such as music and television programs, which reflect consumer tastes and values. In a study of television programs aimed at 9- to 11-year-olds in the United States between 1967 and 2007, researchers found an increasing emphasis on personal achievement and fame and a decreasing emphasis on community belonging (Uhls & Greenfield, 2011).

Factors that influence collectivism and individualism

Several factors may increase an individual's proclivity toward either individualism or collectivism (Triandis, 2018), including both proximal and distal factors (Oyserman et al., 2002b; Xu et al., 2016). Proximal factors have a direct effect on cognition, emotion, and behavior (e.g., social institutions, social situations), and are thus closer to individuals’ daily lives. Economic development is an example of a proximal factor associated with significant change in values and beliefs (Inglehart & Baker, 2000). Farming practices may also be significant: researchers studied 1,162 Han Chinese participants across six sites and found that rice growers in southern China were more interdependent and holistic in their thinking than wheat growers in northern China (Talhelm et al., 2014). This may be because rice cultivation requires significant group cooperation and communication.

Distal factors include culture, the climate, and infectious diseases. Such factors not only directly affect collectivism and individualism, but they also indirectly affect these values through social systems and social contexts. Individualism is most often a consequence of looseness and cultural complexity, while collectivism most frequently results from tightness and cultural simplicity (Triandis, 2018). With respect to the climate, Van de Vliert et al. (2013) conducted a province-level analysis of survey data from 15 Chinese provinces to verify the climato-economic theory (i.e., that the interaction of climatic and economic hardships leads to mutual collectivism). Studies have also shown that natural disasters (Grossmann & Varnum, 2015) and epidemics (Fincher et al., 2008; Fincher & Thornhill, 2012) impact collectivism and individualism.

Impact of infectious disease on collectivism and individualism

The pathogen prevalence hypothesis suggests that the historical prevalence of infectious diseases could contribute to the emergence of cultural variations in collectivism and individualism (Fincher et al., 2008; Murray et al., 2011). This perspective posits that collectivism and individualism may represent evolutionary adaptations to environmental conditions (Jiang et al., 2022). Specifically, the risk of infectious diseases within a given environment shapes the development of behaviors and societal norms aimed at mitigating pathogen transmission. Societies experiencing high pathogenic stress tend to cultivate collectivist cultural norms, serving as a social defense mechanism to reduce the spread of infectious diseases. In contrast, societies with lower pathogenic stress tend to favor individualistic value systems, emphasizing inclusivity, rights, and liberties (Fincher et al., 2008).

In situations of elevated viral infection risk, collectivism tends to increase, irrespective of the culture's typical placement on the individualistic–collectivist spectrum (Fincher et al., 2008; Fincher & Thornhill, 2012). Relative to individualism, collectivism is more likely to prevent epidemics (Fincher et al., 2008; Murray et al., 2011; Thornhill et al., 2010). This is because collectivist cultures, by emphasizing interdependence and in-group coherence, promote social restraint and disease control. Collectivism also emphasizes the importance of adherence to collective behaviors, including those that reduce the risk of infection (Liu et al., 2019; Murray et al., 2013). Faulkner et al. (2004) found that chronic disease anxiety predicted both implicit cognitions associating foreign outgroups with danger, as well as fewer positive attitudes toward foreign (and unfamiliar) immigrant groups. Additionally, subjective perceptions of infection risk increase conformity, preferences for conformity and obedience, and negative responses to those who refuse to conform (Murray & Schaller, 2012). Some researchers have also experimentally demonstrated that pathogen threats tend to increase collectivist behavior (Wu & Chang, 2012). Nevertheless, the experimental investigations could not guarantee ecological validity. In response, some researchers have employed ecological methods to explore whether a one-time outbreak of an infectious disease (e.g., COVID-19) can influence collectivism and individualism in real settings, establishing a theoretical basis for psychological protection against the threat (Han et al., 2021; Na et al., 2021; Ren et al., 2020; Zhao et al., 2021). Han et al. (2021) found that the outbreak of COVID-19 led to an increase in collectivist behavior by analyzing social media posts. In Na et al.'s (2021) study of 9,322 Koreans spanning 14 weeks of the COVID-19 pandemic, they observed an increase in collectivism. Interestingly, individualism remained relatively stable during this period. In contrast, Ren et al. (2020) analyzed Sina Weibo posts by active users and noted that Chinese individuals exhibited reduced individualism and increased collectivism during the COVID-19 phase compared to the pre-outbreak period. Zhao et al. (2021), utilizing a People's Daily database, also reported an increase in collectivism and a decrease in individualism during the COVID-19 pandemic compared to the pre- and post-pandemic periods.

Measurement of collectivism and individualism

The variety of measurement tools for collectivism and individualism rely on numerous operational definitions of the terms. Most of the available tools are Likert self-report scales, including: (1) the Values Survey Module 2013 (VSM; Hofstede & Minkov, 2013), (2) the Horizontal and Vertical Individualism and Collectivism scale (HVIC; Triandis & Gelfand, 1998), and (3) the Individualism–Collectivism scale (INDCOL; Hui & Triandis, 1986). The VSM is a 30-item paper-and-pencil questionnaire that was designed to compare the culturally influenced values and sentiments of similar respondents from two or more countries (or regions within a country). The HVIC assesses cultural orientation across four dimensions, using 16 items: four items assess vertical individualism (e.g., “Winning is everything”), four items assess vertical collectivism (e.g., “Parents and children must stay together as much as possible”), four items assess horizontal individualism (e.g., “I often do ‘my own thing’”), and four items assess horizontal collectivism (e.g., “To me, pleasure is spending time with others”). Higher scores indicate stronger ideological alignment with the relevant dimension. Finally, the INDCOL takes into account the collectivist tendency to vary across target groups. It identifies six target groups (i.e., spouse, parents, kin, neighbors, friends, co-workers) and measures the target-specific construct of individualism–collectivism using 63 items (Hui, 1988).

Since inner traits may also manifest in cultural products and language, some researchers have used these objects to measure collectivism and individualism. Such measurement is culturally specific, and it can support longitudinal research and identify historical trends. Many researchers have used big data from books and other corpora to measure collectivism and individualism, and the feasibility of this method has been repeatedly verified (Xu & Hamamura, 2014). Nafstad et al. (2007) conducted a longitudinal analysis (1984–2005) of Norwegian public discourse, finding that the use of “I/me” equivalents increased considerably over the study period, whereas the use of “we/us” equivalents was stable; moreover, the use of words such as “solidarity,” “common/communal/shared,” “welfare society,” “duty/obligation,” and “equality” decreased, whereas the use of “right/entitlement,” “optional,” and “freedom to choose” increased. On this basis, the researchers concluded that individualism had increased at the expense of communal values. Greenfield (2013) computed the relative frequencies of words indexing individualistic and collectivistic values in American English books published between 1800 and 2000, finding that the use of individualistic words increased, while the use of collectivistic words decreased. Zeng and Greenfield (2015) tracked the use of 16 words (including eight individualistic words, such as “compete”; and eight collectivist words, such as “assign”) to investigate changing cultural values in China. They found that words indexing adaptive individualistic values increased in frequency between 1970 and 2008. In contrast, words indexing less adaptive collectivistic values either decreased in frequency over the same period or increased in frequency at a slower rate than words indexing individualistic values. Cai et al. (2018) sampled the names of Chinese newborns between 1950 and 2010, finding an increasing number of unique names over the period, suggesting a rise in Chinese people's pursuit of uniqueness (i.e., individualism). Zhao et al. (2021) (following Xu & Hamamura, 2014; and Zeng & Greenfield, 2015) analyzed 17 individualistic words and 17 collectivistic words within articles published in the People's Daily of China between 2019 and 2020, finding an increase in the use of collectivistic words and a decrease in the use of individualistic words compared with the pre- and post-pandemic period.

Advantages and feasibility of research based on social media data

Recent studies have used self-report questionnaires to examine psychological factors during the COVID-19 pandemic in different countries (Rossi et al., 2020; Wang et al., 2020). However, these studies have relied on retrospective and time-lagged surveys and interviews, which are limited in their ability to capture accurate psychological data from the pre-pandemic era. This is due to recall bias, which is inevitable when respondents are asked to retrospectively describe a previous mental state.

Social media may contribute to verifying historical mindsets by recording users’ reactions, opinions, and feelings in real time. Previous studies have suggested that the language and psychosocial expressions used on social media may reflect psychological traits (Li et al., 2020; Su et al., 2020). Accordingly, many researchers have used social media data to explore trends in emotional reactions to the COVID-19 outbreak and to trace the public's response to the outbreak in the early stages. Lwin et al. (2020) analyzed more than 20 million Twitter posts during the early phases of the COVID-19 outbreak to examine global trends in four emotions (i.e., fear, anger, sadness, joy) and the narratives underlying those emotions during the pandemic. In China, Weibo is the most widely used social media platform (Huang et al., 2018; Li et al., 2020). Su et al. (2020) collected data from Weibo and Twitter and used the Simplified Chinese and Italian versions of the Language Inquiry and Word Count (LIWC) dictionary to examine and compare the impact of the COVID-19 lockdowns on individuals’ psychological states in China and Italy.

Dictionary-based methods are often suitable for tasks focused on a specific research question or topic (Guo et al., 2016). Accordingly, some researchers have employed relevant dictionaries to measure expressions of collectivism and individualism within social media data. For instance, Huang et al. (2020) applied the Individualism-Collectivism Dictionary (Ren et al., 2017) to Weibo data, finding an association between collectivism and preventive intention in mainland China. Han et al. (2021) used pronouns, group-related words, and relationship-related words to measure collectivism, finding that exposure to COVID-19 (i.e., a pathogen) increased the usage of collectivist words, especially once COVID-19 was revealed to be infectious.

Text analysis methods are particularly well suited for use with large amounts of text data. For large-scale corpora, the calculation of word frequency is not suitable, due to the prevalence of high-frequency, meaningless words. Thus, the frequency-inverse document frequency (TF-IDF) method is typically used to measure the importance of words in a text set (Ramos, 2003). Term frequency (TF) refers to the number of times a word appears in a document, while inverse document frequency (IDF) refers to the relative rarity or uniqueness of that word in the corpus. The TF-IDF of a certain word can be found by multiplying two numbers—the TF and the IDF—indicating the relative importance of the word within a specific document in the corpus. A high TF-IDF score indicates that a word is more important to a specific document and less important to other documents in the corpus. The TF-IDF method is commonly used in information retrieval and text mining, including keyword extraction.

Additionally, the N-Gram method, which predicts word occurrence based on the occurrence of an N–1 (i.e., previous) word (Brown et al., 1992), is extensively used in text-mining and natural-language processing tasks. It involves breaking down a sequence of words into groups of N words, where N is a number chosen by the user, indicating the N-gram size. The most commonly used N-Grams are unigrams (N = 1), bigrams (N = 2), and trigrams (N = 3), which represent single words, word pairs, and word trios, respectively. Thus, the N-Gram method is capable of capturing a word's context within a text. This is important for many natural-language processing tasks, as the meaning of a word can change, depending on the preceding and proceeding words. A combination of TF-IDF and N-Gram approaches is particularly effective at identifying high-frequency unique and effective phrases within a text.

Overview and hypotheses

Most previous studies exploring the relationship between infectious disease threats and collectivism and individualism have either referred to historical records of epidemics/pandemics or generated pathogenic threats experimentally, outside of a natural setting. However, the COVID-19 pandemic provided a natural context in which to study these factors. Several studies have investigated the COVID-19 pandemic’s association with collectivism and individualism, revealing varying conclusions, particularly regarding individualism (Han et al., 2021; Na et al., 2021; Ren et al., 2020; Zhao et al., 2021). Therefore, it is imperative to conduct further exploration of the relationship.

Government responses to the health crisis may also have influenced levels of individualism and collectivism during the pandemic. Luo et al. (2021) found that different COVID-19 prevention policies affected compliance. However, further research is needed to determine whether different policies also affected cultural values, such as collectivism and individualism. Additionally, it would be reasonable to hypothesize that the major events during COVID-19 phrase have a significant impact on collectivism and individualism, as value systems are conditioned by life events, especially those that are important to human life and rooted in autobiographical memory (Czerniawska et al., 2021).

Although the value of dictionary-based approaches has been confirmed, the dictionaries that are applied in such research must be relevant to their target constructs (e.g., collectivism, individualism); otherwise, their validity is diminished. Previous research has studied large-scale corpora (e.g., book publications) over several years, using dictionaries with a limited selection of words representing collectivist and individualist values (Greenfield, 2013; Zeng & Greenfield, 2015; Zhao et al., 2021). However, these limited dictionaries are unable to generate fine-grained depictions of trends in collectivism and individualism over time.

The present study aimed at measuring Weibo users’ collectivism and individualism at the start of the COVID-19 epidemic in China. Considering that the language used on Weibo tends to differ from traditional language expression (i.e., through the inclusion of “online” words that reflect psychological features and update rapidly), a new collectivism and individualism dictionary was constructed to suit the data. The research therefore contributes to filling the gap in the literature on the direct relationship between infectious disease outbreaks and levels of collectivism and individualism, based on social media data. Furthermore, it provides insight into whether the COVID-19 epidemic and individual and collective actions taken in response to the crisis strengthened the collectivism tendency in China.

It was hypothesized that:

H1: The collectivism and individualism among Weibo users during the COVID-19 epidemic would differ from pre-epidemic levels (i.e., the COVID-19 epidemic would impact their collectivism and individualism), showing an increase in collectivism and a decrease in individualism;

H2: The severity of the COVID-19 epidemic would be positively associated with collectivism and negatively associated with individualism among Weibo users;

H3: The COVID-19 public health measures would exhibit a positive correlation with collectivism and a negative association with individualism among Weibo users; and

H4: Fluctuations in the collectivism and individualism among Weibo users would relate to significant events in the COVID-19 epidemic.

The first hypothesis was investigated and confirmed in prior research, and the present study sought to verify these findings. The latter three hypotheses referred to new research questions.

Two studies were conducted. Study 1 involved the construction and validity testing of the Chinese Collectivism and Individualism Dictionary. Study 2 used the Chinese Collectivism and Individualism Dictionary and social media data to explore changes in Weibo users’ collectivism and individualism during the COVID-19 epidemic, as well as the associations between epidemic severity, health policies, and these values.

Study 1

Study 1 aimed at constructing a new Chinese Collectivism and Individualism Dictionary applicable to the online context, and developing a new measurement tool on the basis of verifying its validity.

Method

Participants and data collection

Subjects were recruited online, in two waves.

Thirty-seven subjects (17 men, 20 women, M_age= 26.43) were recruited for the construction of the Chinese Collectivism and Individualism Dictionary. These subjects expanded a previous dictionary, using free association. In addition, for the identification of online words for the dictionary, 498 active Weibo users were recruited who met the following criteria: (a) not newly registered users or users who had abandoned their account, (b) users who consistently used Weibo and had posted content within the prior six months, (c) users who had not deleted any content, and (d) users who posted at least four original Tweets and Retweets each month, with an average number of 10 or more original words per Tweet. The Weibo content of all subjects who met these criteria from October 1, 2019 to October 1, 2020 was crawled.

Four hundred ninety-eight active Weibo users were recruited for the verification of the Chinese Collectivism and Individualism Dictionary, using the HVIC (Triandis & Gelfand, 1998). Ultimately, 460 valid subjects (181 men, 279 women, M_age= 28.7) were retained.

Procedures

Initially, the Chinese Collectivism and Individualism Dictionary was developed and its validity was tested. This laid the foundation for the calculation of collectivism and individualism within the social media data.

To construct the Chinese Collectivism and Individualism Dictionary, dictionaries used in prior research (Greenfield, 2013; Zeng & Greenfield, 2015) were integrated into a preliminary dictionary. Second, subjects were asked to write at least 10 words they associated with the stimuli words “collectivism” and “individualism,” respectively. They were given unlimited time to complete this task, and were instructed to prioritize commonly used words and not to repeat words. All high-frequency words (i.e., those associated by more than 15% of subjects) were included in the dictionary. Subsequently, the dictionary was expanded to include key elements of common collectivism and individualism constructs and metrics. Finally, high-frequency words in Weibo text expressing collectivism and individualism were included.

In the verification phase, subjects completed the HVIC (Triandis & Gelfand, 1998). The HVIC is a Likert self-report scale comprised of 32 items: 16 items measure individualism and 16 items measure collectivism. In the present study, Cronbach's alphas for horizontal individualism, vertical individualism, horizontal collectivism, and vertical collectivism were 0.730, 0.735, 0.808, and 0.738, respectively. Cronbach's alphas for the Individualism and Collectivism subscales were 0.748 and 0.838, respectively.

To verify that the constructed Chinese Collectivism and Individualism Dictionary measured subjects’ collectivism and individualism, a t-test was conducted to determine any significant differences in the frequency of dictionary words between subjects scoring high and low on the HVIC. In addition, two further validity tests were conducted, based on Nicolas et al. (2021). In the first validity test, 32 students majoring in psychology (13 men, 19 women, M_age= 25.94) were invited to rate each word in the dictionary on a scale ranging from 1 (Not at all) to 5 (Extremely), with respect to the degree of accuracy in describing the constructs of collectivism and individualism. A paired sample t-test was conducted to test whether the words in the collectivism dictionary were more consistent with the constructs of collectivism than individualism, and the words in the individualism dictionary were more consistent with the constructs of individualism than collectivism. In the second validity test, 1,000 Weibo posts were randomly selected from the data collected in this study. Subsequently, two students majoring in psychology were invited to rate each Weibo text on a scale ranging from 1 (Not at all) to 5 (Extremely), with respect to the degree of accuracy in reflecting collectivism and individualism, respectively. At the same time, the dictionary was used to analyze and calculate collectivism and individualism scores for each Weibo post. The correlations were then calculated between the mean manual score and the dictionary-based score respectively.

Results

The constructed Chinese Collectivism and Individualism Dictionary contained 200 words: 110 words representing collectivism and 90 words representing individualism (see Table A1, Appendix). To ensure that the corpus was sufficiently large, only subjects posting more than 1,000 words were selected for the analysis. Of the 460 subjects, 117 were separated into groups reflecting high and low collectivism, according to their HVIC scores: 60 subjects scored higher than one standard deviation above the mean (reflecting high collectivism) and 57 subjects scored lower than one standard deviation beneath the mean (reflecting low collectivism). The original Weibo text data of these subjects over the study period were analyzed to extract the collectivism word-frequency index (i.e., the number of collectivism words divided by the total number of words), and an independent sample t-test was conducted. The results indicated a significant difference in the word-frequency index between the high collectivism group and the low collectivism group, t(115) = 2.24, p = 0.027, Cohen's d = 0.42, 95% CI[0.03, 0.45], with the high collectivism group demonstrating higher collectivism word frequency. This suggested that the collectivism words in the dictionary related to collectivism characteristics.

Similarly, 134 of the 460 subjects were separated into groups reflecting high and low individualism, according to their HVIC scores: 67 subjects scored higher than one standard deviation above the mean (reflecting high individualism) and 67 subjects scored lower than one standard deviation beneath the mean (reflecting low individualism). The original Weibo text data of these subjects over the study period were analyzed to extract the individualism word-frequency index (i.e., the number of individualism words divided by the total number of words), and an independent sample t-test was conducted. No significant difference emerged between groups, t(132) = 1.20, p = 0.233, Cohen's d = 0.21, 95% CI[−1.72, 0.42]. Independent sample t-tests were also conducted, using the individualistic dictionaries and first-person singular pronouns used in previous studies. No significant differences emerged between groups, for either the individualistic dictionaries used in previous studies, t(132) = 0.96, p = 0.340, Cohen's d = 0.17, 95% CI[−0.81, 0.28], or the first-person singular pronouns, t(132) = 1.39, p = 0.166, Cohen's d = 0.24, 95% CI[−0.009, 0.016].

For the first validity test, words in the collectivism dictionary were rated as significantly more consistent with the construct of collectivism than the construct of individualism, t(31) = 13.837, p < 0.001, Cohen's d = 2.44, 95% CI[2.02, 2.73]. Similarly, words in the individualism dictionary were rated as significantly more consistent with the construct of individualism than the construct of collectivism, t(31) = 14.118, p < 0.001, Cohen's d = 2.50, 95% CI[2.03, 2.72].

For the second validity test, a significant positive correlation between the mean manual score and dictionary-based score for collectivism was found, r = .36, p < .001, and a significant positive correlation between the mean manual score and dictionary-based score for individualism was found, r = .29, p < .001.

Study 2

Study 2 aimed at using the dictionary and Weibo data to form a continuous fine-grained picture of collectivism and individualism before and after the epidemic, in order to determine whether the epidemic affected collectivism and individualism (H1), the potential association between epidemic-related variables and these levels (H2 and H3), and the relationship between special events in the epidemic and fluctuations in levels of collectivism and individualism (H4).

Method

Participants and data collection

For the formal research, the Study 1 filtering rules were applied to recruit 20,153 active Weibo users. Subsequently, the Weibo content of these users between January 1, 2020 and April 30, 2020 was crawled. The research spanned four months, starting on January 1, 2020, about three weeks before Dr. Zhong Nanshan’s inspection of the epidemic situation in Wuhan, and ending on April 30, approximately two weeks after the official closure of Wuhan’s Leishenshan Hospital. Therefore, the research timeframe comprehensively encompasses the diverse phases of epidemic progression. Fifty subjects were recruited (21 men, 29 women, M_age= 24.94) for the evaluation of significant events during the COVID-19 epidemic.

Procedures

Social media data were analyzed using text analysis techniques, to explore changes in Weibo users’ collectivism and individualism during the COVID-19 epidemic, as well as the associations between epidemic severity, health policies, and these values. First, epidemic-related variables (e.g., severity, public health measures, significant events) were identified and an epidemic-related dictionary was constructed. Second, based on the Chinese Collectivism and Individualism Dictionary, the analyzed Weibo users were scored for collectivism and individualism, trends in these variables were mapped, and epidemic-related factors of influence were explored.

Collection of epidemic-related variables and construction of the epidemic-related dictionary

Variables related to the COVID-19 epidemic included:

Epidemic severity: This was measured using five indicators corresponding to the number of new diagnoses, the number of existing diagnoses, the number of new deaths, the cumulative number of deaths, and the number of suspected cases, daily. These data were obtained from the National Health Commission of the People's Republic of China (National Health Commission of the People's Republic of China, 2020).

COVID-19 policies: These were measured according to an Oxford University dataset that tracked government responses to the pandemic (Hale et al., 2021). Policies were coded using 24 indicators and a miscellaneous field, and organized into five groups: (a) containment and closure policies (indicators C1–C8), such as school closures and movement restrictions; (b) economic policies (indicators E1–E4), such as domestic income support and foreign aid; (c) health system policies (indicators H1–H8), such as COVID-19 testing, emergency investment in healthcare, and vaccination regimes; (d) vaccine policies (indicators V1–V4), relating to prioritization, eligibility, individual cost, and mandates; and (e) miscellaneous policies (M1), for those that did not fit into the other categories. Of note, the coding of policies was updated continuously. Vaccine policies were coded as H7 over the study period (i.e., January 1, 2020 to April 30, 2020). Miscellaneous policies were attributed on the basis of free text note fields. The present study used ordinal data from categories C, E, and H for the analysis.

Significant events in the COVID-19 epidemic: These were initially comprised of the event list published on the “COVID-19 epidemic memorabilia” website, supplemented by 50 subjects’ impressions of influential events, which were sorted into 21 significant events.

The epidemic-related dictionary was constructed by collecting epidemic-related words from existing studies and screening high-frequency words related to the “COVID-19 epidemic” in the collected Weibo data. The Chinese and English names and abbreviations for the COVID-19 epidemic were supplemented. Finally, 119 words were obtained.

Trends in collectivism and individualism as expressed on Weibo, and the mining of influential factors

The Weibo content of 20,153 active users from January 1, 2020 to April 30, 2020 was crawled, and all content was organized according to the user ID, date, and text. Using the Chinese Collectivism and Individualism Dictionary, word frequency and the total number of words were extracted from each user's daily text, and daily collectivism and individualism scores were synthesized at the group level, according to the following calculation: $D a i l y c o l l e c t i v i s m o r i n d i v i d u a l i s m s c o r e = \frac{N u m b e r o f c o l l e c t i v i s m / i n d i v i d u a l i s m w o r d s t h a t d a y}{T o t a l n u m b e r o f w o r d s i n W e i b o t e x t s t h a t d a y} * 100$ First, changes in the number of Weibo posts and levels of collectivism and individualism over the study period were described. Subsequently, a time-series model was fit to the collectivism and individualism scores, respectively. Using the auto.arima() function in R, the period was set to 7 (i.e., 1 week) and the model results (AIC) were combined with the list of significant events in the COVID-19 epidemic to determine the best segmentation point (i.e., the cutting point before and after the epidemic). Differences in collectivism and individualism before and after this point were tested.

For the segmented collectivism and individualism data, the auto.arima() function in R was used to identify the optimal time-series models, with the period set to 7 and the tendency retained. Once sequences during the second epidemic period were set, residues were obtained after fitting. These residues were fitted using hierarchical multiple regression, to investigate potential links between epidemic severity, related policies, and the levels of collectivism and individualism among Weibo users. In the analysis, the residues of collectivism and individualism were used as dependent variables; the periodicity of Weibo content (i.e., with respect to differing levels of engagement on weekends) was used as a control variable; and the severity of the epidemic, epidemic-related policies, and degree of attention to the epidemic (i.e., frequency of posts about the epidemic) were considered independent variables. Given the multiplicity of variables relating to epidemic severity and epidemic-related policies, the correlation between variables was tested to determine whether a collinear relationship was present. When a correlation between variables exceeded 0.6, one of the variables was entered into the model and the other was replaced with a constant or missing value.

Finally, the residue sequence after hierarchical multiple regression fitting was observed, to identify large outliers. TF-IDF and the N-Gram algorithm was used to analyze Weibo text before and after the segmentation and outlier points, to identify the main content of Weibo user discussions before and after the abnormal date, and to explore significant events associated with changes in collectivism and individualism.

Results

Description of trends

The number of Weibo users (of the 20,153 analyzed) who posted daily and the total number of daily Weibo posts were counted, to identify trends over the study period (i.e., January 1, 2020 to April 30, 2020; Figure 1). After January 20 (i.e., after “Zhong Nanshan inspected the epidemic situation in Wuhan and confirmed that the new type of pneumonia can be transmitted from person to person”), there was a large increase in the number of posts, and the longer trend included several peaks (e.g., on February 7, when “Li Wenliang died, and the State Supervision Commission launched an investigation on related issues”). The overall peak appeared around March 19–20, which was the first time that newly confirmed cases in Hubei dropped to 0. At this point, the situation began to stabilize and the number of daily Weibo posts declined.

Figure 1.

Trends in the number of daily Weibo posts and users during the epidemic.

Figure 2 displays trends in Weibo users’ collectivism and individualism over the study period. Collectivism showed an obvious upward trend from the start of the epidemic on January 20 (i.e., when Zhong Nanshan announced the existence of human-to-human transmission of COVID-19), reaching its highest point on January 23 (i.e., when Wuhan entered lockdown). After the situation stabilized (i.e., January 23–March 20, over which period the number of new cases in Hubei declined to 0), levels of collectivism gradually declined. However, collectivism was high on April 4 (i.e., Tomb-Sweeping Day, reflecting a national day of sacrifice). Individualism showed a slight downward trend from the start of the epidemic, then increased once the epidemic stabilized.

Figure 2.

Trends in Weibo users’ collectivism and individualism during the epidemic.

Collectivism and individualism time-series models

The study period was divided into two sections, according to the overall trends in collectivism and individualism. Considering the second to the penultimate day of the study period, five segmentation points with minimum AIC values were output for collectivism (see Appendix, Table A2). The four best segmentation points were concentrated around the period January 19–January 23, reflecting the official outbreak of the epidemic. On each of the identified days, the AIC value was lower than that of the non-segmented fitting (i.e., −296.617), indicating differences in collectivism relative to the start and end of the study period.

Considering that the April 16 AIC output was small but belonged to the late stage of the epidemic, the time-series data were divided into three segments: the first segmentation point was between the 2nd through the 30th day; and the second segmentation point was between the 31st through the penultimate day. Finally, the segmentation points with the smallest AIC values were output for each segment. The first and second segmentation points were combined and the AIC and rationality of different combinations were compared (see Appendix, Table A3). When the 20th and 107th days were used as segmentation points, the AIC reached a minimum value of −326.909; and when the 20th and 95th days were used as segmentation points, the AIC was −322.311 (i.e., there was little difference between the two models). On January 20, Zhong Nanshan announced the human-to-human transmission of COVID-19; and April 4 (i.e., Tomb-Sweeping Day, a national day of mourning) was considered a more reasonable segmentation point than April 16. Therefore, January 20 and April 4 were used as segmentation points.

Between January 1 and January 19, collectivism was low and relatively stable. Between January 20 and April 4, collectivism rose sharply and fluctuated frequently. After April 4, collectivism decreased yet continued to fluctuate.

Individualism was likewise divided into three sections. The best segmentation points of the first segment were January 3, January 19, January 20, and January 21; and the best segmentation points of the second segment were April 15 and April 16. When January 21 and April 16 were used as segmentation points, the AIC reached a minimum value of −257.0156. January 21 fell in a period of increasing COVID-19 cases, and April 16 fell in a period of relative stability in the number of new cases, but significantly more deaths. After segmentation, the AIC value was smaller than that of the non-segmented fitting, indicating differences in individualism relative to the start and end of the study period. Between January 1 and January 21, individualism gradually declined from a high level; between January 22 and April 15, individualism was relatively stable; and from April 16 onward, individualism gradually increased.

Influential factors in collectivism and individualism

Scores for collectivism and individualism were analyzed using traditional ARIMA models. For collectivism, Model 1 spanned January 1 to January 19; Model 2 spanned January 20 to April 4; and Model 3 spanned April 5 to April 30. For individualism, Model 1 spanned January 1 to January 21; Model 2 spanned January 22 to April 15; and Model 3 spanned April 16 to April 30.

For collectivism, the best-fitting model for the first stage was ARIMA (0,0,0), for which the AIC was −71.6624. That is, the autoregression order (AR), moving average order (MA), and integral order (I) of the model were all 0, indicating a purely random sequence. For the second stage, the best-fitting model was ARIMA (1,0,1), for which the AIC was −200.5065, indicating a first-order autoregressive first-order moving average model. The best-fitting model for the third stage was ARIMA (0,0,1), for which the AIC was −57.9443, indicating a first-order moving average model.

For individualism, the best-fitting model for the first stage was ARIMA (1,0,0), for which the AIC was −24.3916, indicating a first-order autoregressive sequence. For the second stage, the best-fitting model was ARIMA (2,0,0), for which the AIC was −122.024, indicating a second-order autoregressive sequence. The best-fitting model for the third stage was ARIMA (1,0,0), for which the AIC was −103.1792, indicating a first-order autoregressive model.

Table 1 displays the results of the analysis of collectivism. The final model accounted for 30.5% of the variance in scores (R²= 0.305). Adding epidemic severity variables in the second step and epidemic-related policy variables in the third step resulted in a significant change in ΔR². In Model 2, the number of confirmed cases significantly predicted the residual (β= −0.27, p < 0.05), suggesting that levels of collectivism were related to epidemic severity. In Model 3, workplace closure (C2) significantly predicted the residual (β= −0.57, p < 0.001), as did restricted internal movement (C7) (β= 0.43, p < 0.01). These findings are consistent with the rejection of out-groups by the behavioral immune system, as demonstrated in previous studies.

Table 2 displays the results of the analysis of individualism. The final model accounted for 31.2% of the variance in scores (R²= 0.312). The addition of attention to the epidemic, epidemic severity, and epidemic-related policy variables to the model resulted in a significant change in ΔR². In Model 4, attention to the epidemic significantly predicted the residual (β= −0.74, p < 0.001), suggesting that discussion of the epidemic on Weibo was related to individualism. The logarithm of the number of suspected infections also significantly predicted the residual (β= 0.56, p < 0.01), indicating that growth in the disease itself, rather than the number of reported cases, was associated with individualism. In terms of epidemic-related policies, school closure (C1) significantly predicted the residual (β= 0.57, p < 0.01) and public information campaigns (H1) significantly predicted the residual negatively (β= −0.75, p < 0.01) (Tables 1 and 2).

Table 1.

Hierarchical multiple linear regression models of collectivism.

Model	Predictors		b	SE	95%CI	β	t	p	R ²	ΔR ²
Model 1									0.026
		Intercept	−0.01	0.02	[−0.04,0.02]		−0.67	0.508
	Control variable	Periodicity	0.04	0.03	[−0.02,0.10]	0.16	1.40	0.166
Model 2									0.101	0.074*
		Intercept	0.05	0.03	[−0.01,0.12]		1.78	0.115
	Control variable	Periodicity	0.04	0.03	[−0.01,0.10]	0.17	1.50	0.100
	Epidemic severity	Number of confirmed cases	−1.08E-06	0	[0.00,0.00]	−0.27	−2.44	0.000***
Model 3									0.290	0.189*
		Intercept	−0.04	0.07	[−0.18,0.10]		−0.55	0.587
	Control variable	Periodicity	0.04	0.03	[−0.02,0.10]	0.14	1.35	0.180
	Epidemic severity	Number of confirmed cases	−5.07E-07	0	[0.00,0.00]	−0.13	−1.10	0.276
	Epidemic-related policies	C2	−0.09	0.02	[−0.13, −0.04]	−0.57	−4.09	0.000***
	Epidemic-related policies	C7	0.15	0.05	[0.06,0.25]	0.43	3.38	0.001**
Model 4									0.305	0.015
		Intercept	−0.18	0.14	[−0.46,0.09]		−1.33	0.188
	Control variable	Periodicity	0.03	0.03	[−0.02,0.08]	0.12	1.20	0.234
	Epidemic severity	Number of confirmed cases	2.34E-08	0	[0.00,0.00]	0.01	0.04	0.971
	Epidemic-related policies	C2	−0.09	0.02	[−0.13, −0.05]	−0.58	−4.16	0.000***
	Epidemic-related policies	C7	0.14	0.05	[0.04,0.23]	0.38	2.88	0.005**
	Attention to the epidemic	Frequency of epidemic words	0.10	0.08	[−0.06,0.25]	0.18	1.22	0.226

Note: * p < 0.05, ** p < 0.01, *** p < 0.001.

Table 2.

Hierarchical multiple linear regression models of individualism.

Model	Predictors		b	SE	95%CI	β	t	p	R ²	ΔR ²
Model 1									0.004
		Intercept	−0.004	0.01	[−0.03,0.02]		−0.30	0.768
	Control variable	Periodicity	−0.01	0.02	[−0.06,0.03]	−0.06	−0.55	0.586
Model 2									0.111	0.108**
		Intercept	0.16	0.05	[0.06,0.27]		3.05	0.003**
	Control variable	Periodicity	−0.01	0.02	[−0.05,0.04]	−0.03	−0.27	0.785
	Attention to the epidemic	Frequency of epidemic words	−0.11	0.04	[−0.18,−0.04]	−0.33	−3.19	0.002**
Model 3									0.208	0.096**
		Intercept	0.22	0.05	[0.11,0.33]		4.10	0.000***
	Control variable	Periodicity	−0.002	0.02	[−0.04,0.04]	−0.01	−0.10	0.917
	Attention to the epidemic	Frequency of epidemic words	−0.25	0.06	[−0.36,−0.14]	−0.74	−4.56	0.000***
	Epidemic severity	Log (number of suspected infections)	0.05	0.02	[0.02,0.08]	0.51	3.17	0.002**
Model 4									0.312	0.105*
		Intercept	0.29	0.07	[0.16,0.43]		4.47	0.000***
	Control variable	Periodicity	0.01	0.02	[−0.03,0.05]	0.07	0.68	0.499
	Attention to the epidemic	Frequency of epidemic words	−0.25	0.05	[−0.36,−0.14]	−0.74	−4.68	0.000***
	Epidemic severity	Log (number of suspected infections)	0.06	0.02	[0.03,0.09]	0.56	3.56	0.001**
	Epidemic-related policies	C1	0.07	0.03	[0.02,0.13]	0.57	2.51	0.014*
	Epidemic-related policies	H1	−0.30	0.09	[−0.49,−0.12]	−0.75	−3.31	0.001**

Note: * p < 0.05, ** p < 0.01, *** p < 0.001.

Significant events associated with changes in collectivism and individualism

From the residue series of the multiple regression of collectivism and individualism, larger points outside the 95% confidence interval were found. Figures 3 and 4 display the abnormal values of the collectivism and individualism residuals, respectively. For collectivism, these values were recorded for January 31, February 1, February 8, and April 4; for individualism, they were recorded for February 8, April 4, and April 5.

Figure 3.

Residual series plot for collectivism.

Figure 4.

Residual series plot for individualism.

Using the TF-IDF and N-Gram algorithms to analyze Weibo text content published on the outlier dates and the days immediately preceding and following, an explanatory tri-gram was produced.

As collectivism was a first-order autoregressive sequence over the study period, each day was affected by the previous day. Thus, data were extracted from the day before and the day after the date on which abnormal values were registered. The first analysis examined text content published on January 30, January 31, and February 1 (see Appendix, Table A4). On January 31 and February 1, the TF-IDF value for the “novel, coronavirus, infection” greatly improved, as did that of attention to the epidemic. Additionally, General Secretary Xi Jinping noted that “The virus is a devil, and we must hunt it down.” Thus, “devil” emerged as a new high-frequency word, exhibiting a correlation with collectivism. Finally, some high-frequency words (i.e., “Shuanghuanglian”) arose related to two significant events: the announcement from the Shanghai Institute of Materia Medica and the Wuhan Institute of Virology that Shuanghuanglian Oral Liquid could inhibit the novel coronavirus, and dispute over the distribution of materials by the Wuhan Red Cross. These events also related to responsibility and collective interests, which may have promoted collectivism.

The results of the TF-IDF analysis of February 8 are shown in the Appendix, Table A5. On February 6 and February 7, words related to “Dr. Li Wenliang's death” were most frequent, surpassing even discussion of the “novel coronavirus infection.” This shows that Weibo users were concerned about the event. Furthermore, expressions such as “Come on, China, come on,” and “Wuhan, come on, China” began to appear, peaking on February 8, reflecting recognition of and preference for the in-group. February 8 was also the date of the traditional Chinese Lantern Festival, which likely affected individualism and collectivism, beyond the COVID-19 epidemic.

The results of the TF-IDF analysis of April 4 are shown in the Appendix, Table A6. As April 4 was Tomb-Sweeping Day (i.e., a national day of mourning to commemorate those who died from COVID-19), discussion of this event dominated. Commemoration is a common behavior of in-groups, as well as a social norm. On the other hand, the words “sacrifice” and “compatriot” are collectivistic, and they may have influenced the rise in collectivism, and the corresponding decrease in individualism on this day. On April 6, the frequency of words about the COVID-19 epidemic returned to their prior level; at the same time, discussion of international mutual assistance and foreign aid increased.

Discussion

The values of collectivism and individualism are critical for differentiating between cultural groups. However, they are not necessarily stable across time (Morris et al., 2015; Ogihara, 2018). Rather, they may be impacted by, among other variables, an infectious disease outbreak (Huang et al., 2020). While there have been many studies on emotional reactions to COVID-19 and the influence of the pandemic on mental health, there have been relatively fewer studies on the impact of the pandemic on cultural values. Thus, the present study retrospectively analyzed trends in collectivism and individualism in China during the early stage of the COVID-19 epidemic, and the factors that influenced these trends, using an original dictionary to analyze social media data.

Use of the dictionary to measure collectivism and individualism

Previous research has used the word frequencies of personal pronouns or representative words (e.g., “cooperation”) to examine changes in collectivism and individualism (Greenfield, 2013; Zeng & Greenfield, 2015). However, the dictionaries used in these studies contain fewer parts of speech, and are thus not suitable for intensive tracking. Second, previous studies have relied on data from books and other publications in different languages; thus, their dictionaries are not necessarily generalizable to social media data. Considering the rapid updating of online vocabulary, a bespoke dictionary is needed to analyze social media text.

The present study employed various methods (i.e., free association, construct analysis, scale-sensitive words) to revise and expand the collectivism and individualism dictionaries used in previous research, based on Weibo data. The validity of the new dictionary was verified using Weibo text and questionnaire data from Weibo users. The results showed that users with high and low collectivism significantly differed in the frequencies with which they used words from the dictionary. However, there was no significant difference in the frequencies between high and low individualism groups.

The reason for this difference may be that people tend to conduct more self-disclosure and personalized display online (Sima & Pugsley, 2010; Wellman et al., 2003). Indeed, research has shown that there are some differences between an individual's internet personality and personality in real life (Liu & Zhu, 2017; Park et al., 2004; Sung et al., 2011). Furthermore, questionnaire investigations are susceptible to many factors, such as social pressure and pressure from measurement scenarios. In China, excessive individualism is not appreciated; thus, while subjects may have intentionally or unintentionally concealed this value when answering the questionnaire, their online behavior may have revealed more of their true thoughts and embrace of individualism. There is no such problem with collectivism. The results of two further validity tests also verified the validity of the dictionary. At the individual level, the correlation between the dictionary and measured constructs of individualism and collectivism was verified, suggesting that the dictionary represents an effective tool for big data measurement of collectivism and individualism.

Use of time-series models to analyze the influence of COVID-19 on collectivism and individualism

To date, there has been scant research on the impact of COVID-19 on cultural values. The present study obtained intensive time-series data about collectivism and individualism from social media data. After trends in Weibo users’ collectivism and individualism over the study period were described, ARIMA time-series models were used (as proposed by Box and Jenkins (1976)) to identify segmentation points. The AIC criterion showed that the segmented models demonstrated the best fit, suggesting that Weibo users’ collectivism and individualism changed over the study period (likely due to the epidemic). In the traditionally collectivist country of China, the overall trend in collectivism increased significantly among Weibo users during the early stages of the epidemic, and declined after the situation stabilized. This finding is consistent with the results of previous studies (Han et al., 2021; Na et al., 2021; Ren et al., 2020; Zhao et al., 2021) showing that collectivism rose significantly after the COVID-19 outbreak. In the present research, individualism declined at the beginning of the epidemic, then gradually rose after the situation stabilized, in line with the findings of some studies (Ren et al., 2020; Zhao et al., 2021). However, Na et al. (2021) found no significant changes in individualism during the health emergency. These contradictory findings may be due to the different periods of analysis, which included different events in the course of the epidemic. Na et al. (2021) examined the period of January 1 to April 7, 2020, whereas the present study identified a split point in individualism on April 15, 2020. In addition, the present study found that, over the longer term, collectivism and individualism returned to their original levels. This suggests that collectivism and individualism, as values, are not only affected by the immediate situation, but also have a certain stability (Grossmann & Na, 2014). The present findings support and enrich the pathogen prevalence hypothesis in a real longitudinal setting. Considering the butterfly effect, the fluctuating levels of collectivism and individualism during the health emergency should attract the attention of the government and people. The present findings suggest that the government may find greater success with strict and effective measures at the beginning of an outbreak, when collectivism is at a peak, and thereby achieve better prevention of viral spread (Lu et al., 2021).

Influential factors in collectivism and individualism

As expected, the COVID-19 outbreak was associated with a rise in collectivism and a decline in individualism. After the situation stabilized, collectivism fell and individualism rose. Epidemic severity, epidemic-related policies, and attention to the epidemic were found to explain these trends, consistent with the results of previous research.

Specifically, the interregional movement prohibitions positively predicted collectivism. A likely explanation for this is the tendency to exclude out-groups and stay among familiar and safe in-groups in the context of an infectious disease outbreak (Murray & Schaller, 2012; Wu & Chang, 2012). Furthermore, workplace closure significantly negatively predicted collectivism. One potential explanation lies in workplace closures, which may engender social isolation and financial insecurity (Brodeur et al., 2020). Consequently, this can diminish the chances for the cultivation of collectivist sentiments, given that social connections represent a pivotal component of collectivism. Similarly, the present study found a significant positive association between school closures and individualism. Due to school closures, the loss of social connections (Hawrilenko et al., 2021) and the emphasis on individual self-directed learning may provide an explanation for the increase in individualism. After controlling for epidemic severity and attention to the epidemic, school closure positively predicted individualism, while individualism was negatively predicted by policy propaganda. Public information campaigns, referring to government-sponsored communication efforts, typically aiming at shaping beliefs, attitudes, social norms, and actual behaviors in the mass public (Solovei & van den Putte, 2020; Weiss & Tschirhart, 1994), foster shared goals and promotes social responsibility among the public, potentially explaining its negative predictive association with individualism. Previous studies have found that different types of policies have a differential impact on compliance, and during the COVID-19 pandemic, attitudes of individualism (emphasizing individual freedom) made it harder for governments to implement certain policies (Bazzi et al., 2021; Chen et al., 2021). And the present findings suggest that policies may also play a role in explaining the changes in cultural values. Thus, the findings may enrich relevant theories.

Attention to the epidemic significantly negatively predicted individualism. Participation in or observation of pandemic-related discussions on social media platforms like Weibo heightens individuals’ awareness of the prevailing collective public health crisis, potentially correlating with diminished individualism. Furthermore, the logarithm of the number of suspected infections significantly predicted individualism. This suggests that the growth rate of the disease itself (rather than the actual number of reported cases) is associated with individualism. The public may have difficulty intuitively perceiving the growth rate; therefore, statistical significance may indicate a relationship between the growth rate and individualism. This may be related to the pathogen prevalence hypothesis, where individualism is considered less conducive to stopping the spread of infectious diseases (Fincher et al., 2008).

Use of Tf-IDF and N-Gram analysis to identify significant events associated with changes in collectivism and individualism

Since word frequency analysis may be affected by high-frequency meaningless words, TF-IDF and N-Gram analysis may assist in highlighting important text features within large corpora. In the present study, TF-IDF and N-Gram analysis were used to identify three categories of events associated with changes in collectivism and individualism: (1) conflict related to responsibilities and collective interests, which tended to make collectivism more pronounced; (2) traditional Chinese festivals, which also emphasized collectivism (Drury, 2018) (i.e., the Lantern Festival, which celebrates reunion within families and in-groups); and (3) collective behaviors, which also tended to increase collectivism (e.g., celebration of National Memorial Day, which commemorates individuals who have sacrificed themselves for collective interests). On this note, previous studies have also used words such as “sacrifice” and “dedication” to measure collectivism (Greenfield, 2013). The traditional questionnaire method requires subjects to recall a certain period of their history. However, this method is limited by the unreliability of memory recall. In contrast, the present study relied on Weibo data, which records users’ feelings and experiences in real time. This data, analyzed using a revised collectivism and individualism word dictionary, was able to generate rich insight into the factors affecting collectivism and individualism in China during the COVID-19 epidemic.

Given the complementary relationship between collectivism and individualism, changes in the former are typically reflected in the latter. In the present study, after the epidemic situation stabilized, collectivism declined while individualism rose. The text analysis showed that the global response to COVID-19 first appeared on April 5. At that time, China adopted strict policies to contain the viral spread, and it was believed that the threat of the virus may not last forever. Over recent decades, collectivism has gradually declined, while individualism has gradually risen (Santos et al., 2017); this may explain why collectivism began to decline after the situation stabilized in April, while individualism began to increase.

Limitations and future directions

In terms of data collection, both the number of study participants and the period of analysis were limited. Future research should examine data over a longer period and include a larger and richer sample of Weibo users of different ages and from different geographical regions.

With respect to methodology, the study relied mainly on a dictionary method to determine daily levels of collectivism and individualism. The dictionary method has certain requirements for the number of words in a text. If data are extracted over a short time (e.g., one day) and at an individual level, the frequency of many words may be 0; thus, the method is unsuitable for measuring short-term data at an individual level. Future research should aim at constructing a prediction model for questionnaire scores. Additionally, more text features should be extracted, including the original forwarded Weibo data and time-series data, to resolve the problem of sparse features. In this way, a short-term prediction model could be built to measure individual trends in collectivism and individualism.

In the present study, individualism scored higher than collectivism, in line with the results of Ren et al. (2020) and Zhao et al. (2021). This is a somewhat surprising finding, given that China has traditionally been considered a collectivist country (Oyserman, 2017). However, this inconsistency may be explained by the bias of online social media, as users tend to express more personal opinions on such platforms. Additionally, research has shown that individualism has become more widespread overall, while collectivism has declined (Huang et al., 2018). More research is needed to explain this trend.

Conclusion

The present research identified turning points in collectivism and individualism among Weibo users, aligned with significant events in the COVID-19 epidemic. Overall, Weibo users’ collectivism significantly increased in the early stages of the epidemic and declined after the situation stabilized. In contrast, individualism declined at the beginning of the epidemic and rose after the situation stabilized. This suggests that the epidemic had a significant impact on these cultural attitudes.

Epidemic-related variables were also found to be correlated with collectivism and individualism. In more detail, epidemic severity, epidemic-related policies, and attention to the epidemic had a significant effect on levels of collectivism and individualism. However, collectivism and individualism exhibited associations with different factors. Specifically, interregional movement prohibitions significantly positively predicted collectivism, while workplace closure significantly negatively predicted collectivism. Additionally, the number of suspected infections and school closures significantly positively predicted individualism, while attention to the epidemic and policy propaganda significantly negatively predicted individualism. Finally, both collectivism and individualism were influenced by significant events in the epidemic.

Footnotes

Data availability statement

Analytic data of the present study are available at

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research,authorship,and/or publication of this article.

Funding

The authors received no financial support for the research,authorship,and/or publication of this article.

ORCID iD

Xinyi Wang

References

Bazzi

Fiszbein

Gebresilasse

(2021). “Rugged individualism” and collective (in)action during the COVID-19 pandemic. Journal of Public Economics, 195(3), 104357. https://doi.org/10.1016/j.jpubeco.2020.104357

Beilmann

Kööts-Ausmees

Realo

(2018). The relationship between social capital and individualism–collectivism in Europe. Social Indicators Research, 137(2), 641–664. https://doi.org/10.1007/s11205-017-1614-4

Box

G. E. P.

Jenkins

G. M.

(1976). Time series analysis. Forecasting and control. Holden-Day Series in Time Series Analysis.

Brodeur

Clark

A. E.

Fleche

Powdthavee

(2020). Assessing the impact of the coronavirus lockdown on unhappiness, loneliness, and boredom using Google Trends. ArXiv, abs/2004.12129.

Brown

P. F.

Della Pietra

V. J.

Desouza

P. V.

Lai

J. C.

Mercer

R. L.

(1992). Class-based N-Gram models of natural language. Computational Linguistics, 18(4), 467–480. https://aclanthology.org/J92-4003

Cai

Zou

Feng

Liu

Jing

(2018). Increasing need for uniqueness in contemporary China: Empirical evidence. Frontiers in Psychology, 9, 554. https://doi.org/10.3389/fpsyg.2018.00554

Chen

Frey

C. B.

Presidente

(2021). Culture and contagion: Individualism and compliance with COVID-19 policy. Journal of Economic Behavior & Organization, 190(10), 191–200. https://doi.org/10.1016/j.jebo.2021.07.026

Czerniawska

Szydło

(2021). Between collectivism and individualism—Analysis of changes in value systems of students in the period of 15 years. Psychology Research and Behavior Management, 14, 2015–2033. https://doi.org/10.2147/PRBM.S330038

Drury

(2018). The role of social identity processes in mass emergency behaviour: An integrative review. European Review of Social Psychology, 29(1), 38–81. https://doi.org/10.1080/10463283.2018.1471948

10.

Faulkner

Schaller

Park

J. H.

Duncan

L. A.

(2004). Evolved disease-avoidance mechanisms and contemporary xenophobic attitudes. Group Processes & Intergroup Relations, 7(4), 333–353. https://doi.org/10.1177/1368430204046142

11.

Fincher

C. L.

Thornhill

(2012). Parasite-stress promotes in-group assortative sociality: The cases of strong family ties and heightened religiosity. Behavioral and Brain Sciences, 35(2), 61–79. https://doi.org/10.1017/S0140525X11000021

12.

Fincher

C. L.

Thornhill

Murray

D. R.

Schaller

(2008). Pathogen prevalence predicts human cross-cultural variability in individualism/collectivism. Proceedings of the Royal Society B: Biological Sciences, 275(1640), 1279–1285. https://doi.org/10.1098/rspb.2008.0094

13.

Greenfield

P. M.

(2000). Three approaches to the psychology of culture: Where do they come from? Where can they go? Asian Journal of Social Psychology, 3(3), 223–240. https://doi.org/10.1111/1467-839X.00066

14.

Greenfield

P. M.

(2013). The changing psychology of culture from 1800 through 2000. Psychological Science, 24(9), 1722–1731. https://doi.org/10.1177/0956797613479387

15.

Grossmann

(2014). Research in culture and psychology: Past lessons and future challenges. WIREs Cognitive Science, 5(1), 1–14. https://doi.org/10.1002/wcs.1267

16.

Grossmann

Varnum

M. E. W.

(2015). Social structure, infectious diseases, disasters, secularism, and cultural change in America. Psychological Science, 26(3), 311–324. https://doi.org/10.1177/0956797614563765

17.

Guo

Vargo

C. J.

Pan

Ding

Ishwar

(2016). Big social data analytics in journalism and mass communication: Comparing dictionary-based text analysis and unsupervised topic modeling. Journalism & Mass Communication Quarterly, 93(2), 332–359. https://doi.org/10.1177/1077699016639231

18.

Hale

Angrist

Goldszmidt

Kira

Petherick

Phillips

Webster

Cameron-Blake

Hallas

Majumdar

Tatlow

(2021). A global panel database of pandemic policies (Oxford COVID-19 Government Response Tracker). Nature Human Behaviour, 5(4), 529–538. https://doi.org/10.1038/s41562-021-01079-8

19.

Hamamura

(2012). Are cultures becoming individualistic? A cross-temporal comparison of individualism–collectivism in the United States and Japan. Personality and Social Psychology Review, 16(1), 3–24. https://doi.org/10.1177/1088868311411587

20.

Han

Ren

Liu

Zhu

(2021). Increase of collectivistic expression in China during the COVID-19 outbreak: An empirical study on online social networks. Frontiers in Psychology, 12, 632204. https://doi.org/10.3389/fpsyg.2021.632204

21.

Hawrilenko

Kroshus

Tandon

Christakis

(2021). The association between school closures and child mental health during COVID-19. JAMA Network Open, 4(9), e2124092–e2124092. https://doi.org/10.1001/jamanetworkopen.2021.24092

22.

Hofstede

(1984). Culture’s consequences: International differences in work-related values (Vol. 5). Sage.

23.

Hofstede

Minkov

(2013). VSM 2013. Values survey module.

24.

Huang

Ding

Liu

Zhu

(2020). How fear and collectivism influence public’s preventive intention towards COVID-19 infection: A study based on big data from the social media. BMC Public Health, 20(1), 1707. https://doi.org/10.1186/s12889-020-09674-6

25.

Huang

Jing

Zhou

Zhang

Cai

(2018). Increasing individualism and decreasing collectivism? Cultural and psychological change around the globe. Advances in Psychological Science, 26(11), 2068–2080. https://doi.org/10.3724/SP.J.1042.2018.02068

26.

Huang

Z. H.

Wang

Cai

H. J.

(2016). Rising individualism in China: Convergent evidence from diverse perspectives. Poster Presented at the 23rd Conference of International Association for Cross-Cultural Psychology.

27.

Hui

C. H.

(1988). Measurement of individualism–collectivism. Journal of Research in Personality, 22(1), 17–36. https://doi.org/10.1016/0092-6566(88)90022-0

28.

Hui

C. H.

Triandis

H. C.

(1986). Individualism–collectivism: A study of cross-cultural researchers. Journal of Cross-Cultural Psychology, 17(2), 225–248. https://doi.org/10.1177/0022002186017002006

29.

Inglehart

Baker

W. E.

(2000). Modernization, cultural change, and the persistence of traditional values. American Sociological Review, 65(1), 19–51. https://doi.org/10.2307/2657288

30.

Jiang

Wei

Zhang

(2022). Individualism versus collectivism and the early-stage transmission of COVID-19. Social Indicators Research, 164(2), 791–821. https://doi.org/10.1007/s11205-022-02972-z

31.

Kashima

E. S.

(2003). Individualism, GNP, climate, and pronoun drop: Is individualism determined by affluence and climate, or does language use play a role? Journal of Cross-Cultural Psychology, 34(1), 125–134. https://doi.org/10.1177/0022022102239159

32.

Lansford

J. E.

Zietz

Al-Hassan

S. M.

Bacchini

Bornstein

M. H.

Chang

Deater-Deckard

Di Giunta

Dodge

K. A.

Gurdal

Liu

Long

Oburu

Pastorelli

Skinner

A. T.

Sorbring

Tapanya

Steinberg

Uribe Tirado

L. M.

Yotanyamaneewong

Alampay

L. P.

(2021). Culture and social change in mothers’ and fathers’ individualism, collectivism and parenting attitudes. Social Sciences, 10(12), 459. https://doi.org/10.3390/socsci10120459

33.

Wang

Xue

Zhao

Zhu

(2020). The impact of COVID-19 epidemic declaration on psychological consequences: A study on active Weibo users. International Journal of Environmental Research and Public Health, 17(6), 2032. https://doi.org/10.3390/ijerph17062032

34.

Liu

S. S.

Morris

M. W.

Talhelm

Yang

(2019). Ingroup vigilance in collectivistic cultures. Proceedings of the National Academy of Sciences, 116(29), 14538–14546. https://doi.org/10.1073/pnas.1817588116

35.

Liu

Zhu

(2017). Comparing online personality of Americans and Chinese. In Information Resources Management Association (Ed.), Gaming and Technology Addiction: Breakthroughs in Research and Practice (pp. 339–351). IGI Global. https://doi.org/10.4018/978-1-5225-0778-9.ch016

36.

J. G.

Jin

English

A. S.

(2021). Collectivism predicts mask use during COVID-19. Proceedings of the National Academy of Sciences, 118(23), e2021793118. https://doi.org/10.1073/pnas.2021793118

37.

Luo

Xue

Mattila

A. S.

(2021). Appreciation vs. apology: When and why does face covering requirement increase revisit intention? Journal of Retailing and Consumer Services, 63(6), 102705. https://doi.org/10.1016/j.jretconser.2021.102705

38.

Lwin

M. O.

Sheldenkar

Schulz

P. J.

Shin

Gupta

Yang

(2020). Global sentiments surrounding the COVID-19 pandemic on Twitter: Analysis of Twitter trends. JMIR Public Health and Surveillance, 6(2), e19447–e19447. https://doi.org/10.2196/19447

39.

Markus

H. R.

Kitayama

(1991). Culture and the self: Implications for cognition, emotion, and motivation. Psychological Review, 98(2), 224–253. https://doi.org/10.1037/0033-295X.98.2.224

40.

Morris

Chiu

C. Y.

Liu

(2015). Polycultural psychology. Annual Review of Psychology, 66(1), 631–659. https://doi.org/10.1146/annurev-psych-010814-015001

41.

Murray

D. R.

Schaller

(2012). Threat(s) and conformity deconstructed: Perceived threat of infectious disease and its implications for conformist attitudes and behavior. European Journal of Social Psychology, 42(2), 180–188. https://doi.org/10.1002/ejsp.863

42.

Murray

D. R.

Schaller

Suedfeld

(2013). Pathogens and politics: Further evidence that parasite prevalence predicts authoritarianism. Plos One, 8(5), e62275. https://doi.org/10.1371/journal.pone.0062275

43.

Murray

D. R.

Trudeau

Schaller

(2011). On the origins of cultural differences in conformity: Four tests of the pathogen prevalence hypothesis. Personality and Social Psychology Bulletin, 37(3), 318–329. https://doi.org/10.1177/0146167210394451

44.

Kim

Suk

H. W.

Choi

J. A.

Kim

J. H.

Kim

Choi

(2021). Individualism–collectivism during the COVID-19 pandemic: A field study testing the pathogen stress hypothesis of individualism-collectivism in Korea. Personality and Individual Differences, 183(16), 111127. https://doi.org/10.1016/j.paid.2021.111127

45.

Nafstad

H. E.

Blakar

R. M.

Carlquist

Phelps

J. M.

Rand-Hendriksen

(2007). Ideology and power: The influence of current neo-liberalism in society. Journal of Community & Applied Social Psychology, 17(4), 313–327. https://doi.org/10.1002/casp.931

46.

National Health Commission of the People’s Republic of China (2020). Daily briefing on COVID-19. http://www.nhc.gov.cn/xcs/yqtb/list_gzbd.shtml .

47.

Nicolas

Bai

Fiske

S. T.

(2021). Comprehensive stereotype content dictionaries using a semi-automated method. European Journal of Social Psychology, 51(1), 178–196. https://doi.org/10.1002/ejsp.2724

48.

Ogihara

(2018). The rise in individualism in Japan: Temporal changes in family structure, 1947–2015. Journal of Cross-Cultural Psychology, 49(8), 1219–1226. https://doi.org/10.1177/0022022118781504

49.

Oishi

(2010). The psychology of residential mobility: Implications for the self, social relationships, and well-being. Perspectives on Psychological Science, 5(1), 5–21. https://doi.org/10.1177/1745691609356781

50.

Oyserman

(2017). Culture three ways: Culture and subcultures within countries. Annual Review of Psychology, 68(1), 435–463. https://doi.org/10.1146/annurev-psych-122414-033617

51.

Oyserman

Coon

H. M.

Kemmelmeier

(2002a). Rethinking individualism and collectivism: Evaluation of theoretical assumptions and meta-analyses. Psychological Bulletin, 128(1), 3–72. https://doi.org/10.1037/0033-2909.128.1.3

52.

Oyserman

Kemmelmeier

Coon

(2002b). Cultural psychology, a new look: Reply to Bond (2002), Fiske (2002), Kitayama (2002), and Miller (2002). Psychological Bulletin, 128(1), 110–117. https://doi.org/10.1037/0033-2909.128.1.110

53.

Oyserman

Lee

S. W. S.

(2008). Does culture influence what and how we think? Effects of priming individualism and collectivism. Psychological Bulletin, 134(2), 311–342. https://doi.org/10.1037/0033-2909.134.2.311

54.

Park

S. E.

Kim

(2004). How to visually create clear personalities with Blogs?. In ACM SIGGRAPH 2004 Posters (p. 8). https://doi.org/10.1145/1186415.1186425.

55.

Ramos

(2003). Using TF-IDF to determine word relevance in document queries. Proceedings of the First Instructional Conference on Machine Learning, 242(1), 29–48. https://api.semanticscholar.org/CorpusID:14638345

56.

Realo

(2003). Comparison of public and academic discourses: Estonian individualism and collectivism revisited. Culture & Psychology, 9(1), 47–77. https://doi.org/10.1177/1354067X03009001004

57.

Realo

Koido

Ceulemans

Allik

(2002). Three components of individualism. European Journal of Personality, 16(3), 163–184. https://doi.org/10.1002/per.437

58.

Ren

Han

Zhu

(2020). COVID-19 increased collectivistic expression on Sina Weibo. In 2020 7th International Conference on Behavioural and Social Computing (BESC), 1–4.

59.

Ren

Xiang

Zhou

Zhu

(2017). Individualism/collectivism map of China based on Weibo. Journal of Inner Mongolia Normal University (Philosophy & Social Science), 46(6). https://doi.org/10.3969/j.issn.1001-7623.2017.06.011

60.

Rossi

Socci

Talevi

Mensi

Niolu

Pacitti

Di Marco

Rossi

Siracusano

Di Lorenzo

(2020). COVID-19 pandemic and lockdown measures impact on mental health among the general population in Italy. Frontiers in Psychiatry, 11, 790. https://doi.org/10.3389/fpsyt.2020.00790

61.

Santos

H. C.

Varnum

M. E. W.

Grossmann

(2017). Global increases in individualism. Psychological Science, 28(9), 1228–1239. https://doi.org/10.1177/0956797617700622

62.

Schaller

Murray

(2008). Pathogens, personality, and culture: Disease prevalence predicts worldwide variability in sociosexuality, extraversion, and openness to experience. Journal of Personality and Social Psychology, 95(1), 212–221. https://doi.org/10.1037/0022-3514.95.1.212

63.

Sima

Pugsley

P. C.

(2010). The rise of a “me culture” in postsocialist China: Youth, individualism and identity creation in the blogosphere. International Communication Gazette, 72(3), 287–306. https://doi.org/10.1177/1748048509356952

64.

Solovei

van den Putte

(2020). The effects of five public information campaigns: The role of interpersonal communication. Communications, 45(s1), 586–602. https://doi.org/10.1515/commun-2020-2089

65.

Xue

Liu

Chen

Liu

Gong

Zhu

(2020). Examining the impact of COVID-19 lockdown in Wuhan and Lombardy: A psycholinguistic analysis on Weibo and Twitter. International Journal of Environmental Research and Public Health, 17(12), 4552. https://doi.org/10.3390/ijerph17124552

66.

Sung

Moon

Kang

Lin

J.-S.

(2011). Actual self vs. avatar self: The effect of online social situation on self-expression. Journal of Virtual Worlds Research, 4(1), 4–21. https://doi.org/10.4101/jvwr.v4i1.1927

67.

Talhelm

Zhang

Oishi

Shimin

Duan

Lan

Kitayama

(2014). Large-scale psychological differences within China explained by rice versus wheat agriculture. Science, 344(6184), 603–608. https://doi.org/10.1126/science.1246850

68.

Thornhill

Fincher

C. L.

Murray

D. R.

Schaller

(2010). Zoonotic and non-zoonotic diseases in relation to human personality and societal values: Support for the parasite-stress model. Evolutionary Psychology, 8(2), 147470491000800201. https://doi.org/10.1177/147470491000800201

69.

Triandis

H. C.

(1996). The psychological measurement of cultural syndromes. American Psychologist, 51(4), 407–415. https://doi.org/10.1037/0003-066X.51.4.407

70.

Triandis

H. C.

(2001). Individualism-collectivism and personality. Journal of Personality, 69(6), 907–924. https://doi.org/10.1111/1467-6494.696169

71.

Triandis

H. C.

(2018). Individualism and collectivism. Routledge.

72.

Triandis

H. C.

Gelfand

M. J.

(1998). Converging measurement of horizontal and vertical individualism and collectivism. Journal of Personality and Social Psychology, 74(1), 118–128. https://doi.org/10.1037/0022-3514.74.1.118

73.

Trommsdorff

Mayer

Albert

(2004). Dimensions of culture in intra-cultural comparisons: Individualism/collectivism and family-related values in three generations. In Comparing cultures (pp. 157–179). Brill.

74.

Twenge

J. M.

Abebe

E. M.

Campbell

W. K.

(2010). Fitting in or standing out: Trends in American parents’ choices for children’s names, 1880–2007. Social Psychological and Personality Science, 1(1), 19–25. https://doi.org/10.1177/1948550609349515

75.

Uhls

Y. T.

Greenfield

P. M.

(2011). The rise of fame: An historical content analysis. Cyberpsychology: Journal of Psychosocial Research on Cyberspace, 5(1), 1. https://cyberpsychology.eu/article/view/4243

76.

Van de Vliert

Yang

Wang

Ren

X.-p.

(2013). Climato-economic imprints on Chinese collectivism. Journal of Cross-Cultural Psychology, 44(4), 589–605. https://doi.org/10.1177/0022022112463605

77.

van Leeuwen

Park

J. H.

Koenig

B. L.

Graham

(2012). Regional variation in pathogen prevalence predicts endorsement of group-focused moral concerns. Evolution and Human Behavior, 33(5), 429–437. https://doi.org/10.1016/j.evolhumbehav.2011.12.005

78.

Wang

Pan

Wan

Tan

C. S.

R. C.

(2020). Immediate psychological responses and associated factors during the initial stage of the 2019 coronavirus disease (COVID-19) epidemic among the general population in China. International Journal of Environmental Research and Public Health, 17(5), 1729. https://doi.org/10.3390/ijerph17051729

79.

Weiss

J. A.

Tschirhart

(1994). Public information campaigns as policy instruments. Journal of Policy Analysis and Management, 13(1), 82–119. https://doi.org/10.2307/3325092

80.

Wellman

Quan-Haase

Boase

Chen

Hampton

Díaz

Miyata

(2003). The social affordances of the internet for networked individualism. Journal of Computer-Mediated Communication, 8(3), JCMC834. https://doi.org/10.1111/j.1083-6101.2003.tb00216.x

81.

B.-P.

Chang

(2012). The social impact of pathogen threat: How disease salience influences conformity. Personality and Individual Differences, 53(1), 50–54. https://doi.org/10.1016/j.paid.2012.02.023

82.

Ren

(2016). The influencing factors of individualism/collectivism: A perspective of ecology. Advances in Psychological Science, 24(8), 1309. https://doi.org/10.3724/SP.J.1042.2016.01309

83.

Hamamura

(2014). Folk beliefs of cultural changes in China [original research]. Frontiers in Psychology, 5(8), 1066. https://doi.org/10.3389/fpsyg.2014.01066

84.

Zeng

Greenfield

P. M.

(2015). Cultural evolution over the last 40 years in China: Using the Google Ngram Viewer to study implications of social and political change for cultural values. International Journal of Psychology, 50(1), 47–55. https://doi.org/10.1002/ijop.12125

85.

Zhao

Ren

(2021). SARS-CoV-2 increased collectivistic expressions in People’s Dailys. In 2021 8th International Conference on Behavioral and Social Computing (BESC), 1–4.

Impact of the COVID-19 epidemic on collectivism and individualism in China: A study of Weibo users

Abstract

Keywords

Introduction

Collectivism and individualism

The collectivism and individualism constructs

Changes in collectivism and individualism

Factors that influence collectivism and individualism

Impact of infectious disease on collectivism and individualism

Measurement of collectivism and individualism

Advantages and feasibility of research based on social media data

Overview and hypotheses

Study 1

Method

Participants and data collection

Procedures

Results

Study 2

Method

Participants and data collection

Procedures

Collection of epidemic-related variables and construction of the epidemic-related dictionary

Trends in collectivism and individualism as expressed on Weibo, and the mining of influential factors

Results

Description of trends

Collectivism and individualism time-series models

Influential factors in collectivism and individualism

Significant events associated with changes in collectivism and individualism

Discussion

Use of the dictionary to measure collectivism and individualism

Use of time-series models to analyze the influence of COVID-19 on collectivism and individualism

Influential factors in collectivism and individualism

Use of Tf-IDF and N-Gram analysis to identify significant events associated with changes in collectivism and individualism

Limitations and future directions

Conclusion

Footnotes

Data availability statement

Declaration of conflicting interests

Funding

ORCID iD

References