The COVID-19 pandemic revealed significant cross-national differences in infection rates. While initial models focused on government responses and travel restrictions, this study explores the role of social norms in shaping these variations. The researchers question whether individual willingness to compromise freedoms, societal compliance, sociability, or accustomed freedoms influence the effectiveness of social distancing measures. Previous research on pandemics has linked factors like low education, ethnicity, overcrowding, individualistic values, and poor socio-environmental conditions to increased infection rates. However, the connection between social capital, personal freedom, and virus spread under extreme restrictions remained unexplored. This study aimed to analyze the association between social characteristics and COVID-19 doubling rates, controlling for climate, population, government effectiveness, testing rates, healthcare access, preventative interventions, GDP, and tourism.

Existing literature highlights the impact of various social and cultural factors on pandemic spread. Studies on H1N1, zoonotic diseases, and tuberculosis showed correlations between low education levels, non-Caucasian ethnicity, overcrowding, material deprivation, individualistic values, and poor socio-environmental conditions with increased hospitalization rates and outbreaks. However, these studies lacked a comprehensive assessment of the interplay between social capital, personal freedom, and virus spread dynamics during periods of stringent restrictions. This study bridges this gap by investigating the influence of social norms and values on the early transmission dynamics of COVID-19.

The study used a multi-faceted approach combining various datasets. Data selection prioritized relevance (worldwide studies with nationally representative samples), recency (data close to the COVID-19 outbreak), and reproducibility (studies with repeated annual sampling). Datasets included the 2019 Legatum Institute's Prosperity Index, the 6th wave of the World Values Survey (WVS), and the 2015 Healthcare Access and Quality index. COVID-19 incidence data for a 71-day period (January 23 to March 27, 2020) was sourced from the Johns Hopkins University CSSE database. Doubling times were calculated from log-linear models of incidence data. The study included several covariates: government effectiveness, individual relationships, personal freedom, climate zone, population size and density, government stringency index (descriptive purposes only), testing rates (descriptive purposes only), healthcare access and quality, preventative interventions, GDP per capita, and tourism contribution to GDP. Model-based clustering (mclust package) with the Bayesian Information Criterion (BIC) identified distinct groups of societies. Random forest regression (RandomForestSRC package) analyzed the predictors of COVID-19 doubling time. Continuous variables were log-transformed, and node size and variable selection were optimized to minimize out-of-bag error. The analysis used 100 trees with sampling without replacement (swor). Minimal depth and permutation variable importance measures (VIMPs) were calculated.

Clustering revealed four distinct societal groups: 1. **Reserved societies:** Characterized by high population, low population density, low government effectiveness, low GDP per capita, low HAQ index, high doubling time, low testing rates, moderate government stringency, low agency, low freedom of assembly and association, low family relationships, dry climates (e.g., Iran, Pakistan, Egypt). Citizens showed less interest in societal good and proper behavior but higher confidence in government and obedience to rulers. 2. **Drifting societies:** Featured temperate climates, low population and density, high government effectiveness, moderate freedom of assembly and association and agency, moderate family relationships, low civic participation, low GDP and HAQ index, low doubling time, and high government stringency (e.g., Czechia, Greece, Italy, Hungary). Citizens showed less importance of societal good, proper behavior, and lower confidence in government and obedience to rulers. 3. **Assertive societies:** Included temperate climates, high government effectiveness, strong social networks, high personal and family relationships, high freedom of assembly and association, high GDP per capita, high HAQ index, and low doubling time (e.g., Australia, USA, UK, Germany). Citizens emphasized societal good and proper behavior but reported lower confidence in government and obedience. 4. **Compliant societies:** Showed dry climates, high population density, low social networks and family relationships, low freedom of assembly and association, high HAQ index, high GDP per capita, high doubling time, high testing rates, and high government stringency (e.g., South Korea, China, UAE). Citizens prioritized societal good and proper behavior, exhibiting high confidence in government and importance of obedience. High doubling times despite high population densities suggest regulatory compliance. Random forest analysis revealed that population density, freedom of assembly and association, and agency were the most important predictors of doubling time. Population density, GDP per capita, and temperate climates showed positive associations with doubling time, while freedom of assembly and association and agency showed negative associations. The model's error rate was 0.25.

The findings emphasize the significant influence of public attitudes and responses on COVID-19 spread. Accustomed freedom of assembly and agency were stronger predictors of transmission speed than economic or healthcare indicators. Societies with high freedom of assembly and agency (drifting and assertive societies) found it more challenging to comply with strict regulations, indicating a trade-off between protecting human rights and controlling the pandemic. The research highlights the ethical dilemma between effectively combating the virus and upholding fundamental rights. The study touches upon the role of public risk interpretation (“I am” vs. “we are” at risk) and the potential link between individualistic values and increased transmission rates. While some previous research associated collectivistic cultures with lower transmission in hotter climates, this study’s findings are more nuanced, particularly distinguishing between Cluster 4 (compliant) and Cluster 1 (reserved) societies.

This study underscores the crucial role of social factors in shaping COVID-19 transmission. Freedom of assembly and agency emerged as significant predictors of transmission rates, highlighting the need to consider the societal context and public response when developing pandemic control strategies. Future research should explore the ethical and psychological aspects of the freedom versus security dilemma, particularly examining the effectiveness of libertarian paternalism in promoting social distancing. Further investigation into the replicability of the cluster typology across other infectious diseases is warranted.

The study acknowledges limitations such as the potential obsolescence of some datasets (WVS and HAQ), the unconfirmed replicability of cluster assignments, and the limited number of observations. The reliance on existing datasets also limits the ability to directly address specific causal mechanisms.