Factors driving the implementation of the 'Local New Year' policy to prevent COVID-19 in China

The Spring Festival Transportation (SFT) in China is the world’s largest periodic human migration, typically moving nearly 3 billion passenger trips annually prior to COVID-19. With the outbreak of COVID-19 in early 2020, SFT volumes dropped drastically due to stringent prevention measures, including advocacy for a ‘local new year’ (staying in the city of work rather than returning to hometown). While effective for disease control, repeated advocacy of ‘local new year’ under normalized prevention and control sparked social tension by conflicting with strong cultural and familial reunion norms tied to the Spring Festival. This study addresses the core research question: what factors drive individuals’ willingness to support and comply with the ‘local new year’ policy? By uncovering the key determinants of behavior, the study seeks to inform more human-centered, operationally feasible policies that both curb transmission and better accommodate emotional and social needs during the festival period.

Prior research shows population mobility substantially accelerates infectious spread; a one-standard-deviation increase in mobility is associated with a 12.8% rise in confirmed cases, and reducing mobility lowers infection rates. Behavioral adaptations during COVID-19 included more home cooking, online shopping, remote work, reduced travel and social visits, and shifts from public transport to walking, cycling, and private vehicles—changes that reduced contact opportunities. However, most studies emphasize descriptive behavior changes, with less analysis of the underlying decision logic and determinants. Policy responses in China during early 2020 included lockdowns, intensive monitoring, rapid diagnosis, contact tracing, and data-driven adjustments. On the eve of the 2022 SFT, at least 55 regions advocated ‘local new year’, with heterogeneous implementations ranging from strict enforcement without supporting measures, to moral persuasion coupled with material and cultural support, to flexible, needs-based approaches including subsidies. This context motivates examining the determinant structure behind individuals’ support for ‘local new year’ to reconcile pandemic control with cultural-emotional needs.

Design: A two-stage survey was conducted to identify and analyze determinants of willingness to celebrate the ‘local new year’.

• Stage 1 (exploratory): An initial online survey (n=100) asked migrant workers in large cities to list factors influencing their decision. Frequently mentioned items (>10%) formed a candidate factor list (11 items; see Table 2 in paper).
• Stage 2 (confirmatory): A structured online questionnaire was disseminated across East, North, Southwest, and Southeast China in December 2021 (eve of 2022 SFT). Of 600 distributed, 584 valid responses were collected (~97.3% effective rate). The instrument captured demographics and coded factor variables (binary and discrete) per Table 3 (e.g., income satisfaction, relatives in hometown, government call for ‘local new year’, ticket availability, subsidies, perceived infection risk during travel, traditional reunion sense, activities arranged at hometown/work city, travel convenience).

Factor screening: Mallows’ Cp method was applied to select a parsimonious subset of predictors with good fit, balancing model simplicity and mean squared error. The top-fitting model (lowest Cp) included four variables: COL10 (relatives in hometown), COL12 (can buy ticket), COL14 (risk of contracting COVID-19 while traveling), COL18 (activities arranged in city of work).

Modeling: A binary Logit model (maximum-likelihood estimation) was specified with the dependent variable indicating support for ‘local new year’ (i.e., staying rather than returning home). The log-odds of the outcome were modeled as a linear function of selected predictors. To prevent complete separation, multiple predictors were included: Logit = β0 + β1X1 + β2X2 + ... . Significance was assessed via Wald Chi-Square tests; effect sizes are presented as odds ratios with 95% Wald CIs.

Sample profile (Stage 2 overview): Gender: 52.7% male, 47.3% female; Age: 20–30 (29.17%), 30–40 (32.5%), 40–50 (36.67%), 50–60 (1.67%); Occupations included migrant/construction (33.33%), service (25%), official (10%), company employees (10%), professional/technical (5%), other (16.67%); Monthly income: <2000 (8.33%), 2000–5000 (49.17%), 5000–10,000 (32.5%), >10,000 (1.67%); Children: 75% yes; Family around: 71.67% yes; Hometown: rural (78.33%), county (10%), city (11.67%); Distance to hometown: inter-province (67.5%), neighboring province (7.5%), intra-province (25%).

• Cp screening identified four key candidates: COL10 (relatives in hometown), COL12 (can buy ticket), COL14 (risk of contracting COVID-19 during travel), COL18 (activities arranged in city of work).
• In the Logit model (MLE), COL12 (ticket availability) was not significant (p=0.3736), leaving three significant determinants: • Relatives in hometown (COL10): Estimate 2.0164 (SE 0.9028), Wald χ²=4.9881, p=0.0255; OR=7.511 (95% CI 1.280–44.079). Individuals with elderly/children at home had 7.5 times higher odds of returning home (i.e., lower support for staying) than those without. • Infection risk during travel (COL14): Estimate −2.1320 (SE 0.9530), Wald χ²=5.0051, p=0.0253; OR=0.119 (95% CI 0.018–0.768). Concern about contracting COVID-19 while traveling reduced odds of going home; inversely, those unconcerned were ~8.4 times more likely to go home than those concerned. • Activities arranged in city of work (COL18): Estimate −4.5826 (SE 1.7591), Wald χ²=6.7866, p=0.0092; OR=0.010 (95% CI <0.001–0.321). Having activities arranged locally greatly reduced the odds of going home; inversely, those without activities were about 100 times more likely to go home than those with activities.
• Primary driver: Arranged activities in city of work (COL18) is the strongest determinant encouraging support for ‘local new year’.
• Secondary drivers: Having relatives (elderly/children) in hometown and perceived infection risk during travel have similar magnitudes and act as secondary determinants.
• Policy relevance: Cash subsidies and ticket availability (though highlighted by Cp) did not retain significance in the final model, suggesting material compensation is less influential than fulfilling spiritual/cultural needs via local activities and addressing family care concerns.

The findings directly address the tension between pandemic control needs and cultural-emotional imperatives during the Spring Festival. By quantifying determinants, the study shows that enriching local cultural and social life (arranging activities in the work city) most strongly increases acceptance of staying, while familial obligations and perceived travel infection risk shape the decision in secondary ways. Consequently, effective policy should: (1) foreground human-centered measures that meet spiritual and social needs for those staying; (2) mitigate concerns about elderly/children left in hometowns through supportive services; and (3) tailor policy intensity to local epidemic conditions. The authors propose a tiered framework: strict implementation with essential services in high-severity areas; encouragement with strong cultural-life support in medium-risk areas; and flexible, non-coercive approaches in controlled/low-risk areas with guidance based on destination risk. Implementation should prioritize cultural/recreational offerings and humanistic care over cash subsidies, and provide mechanisms (e.g., online reunions, community support, targeted transport/isolation solutions) to reduce the perceived need to return home while maintaining public health safeguards.

Using a two-stage design with Cp screening and Logit modeling under normalized COVID-19 control, the study identifies the key drivers of willingness to support the ‘local new year’ policy. The strongest driver is the provision of activities in the city of work, with familial ties (elderly/children at home) and perceived infection risk during travel as secondary determinants. These results imply that current policies emphasizing material compensation are misaligned with actual motivations; instead, policies should be regionally flexible and prioritize spiritual-cultural fulfillment and care for dependents in hometowns. The work contributes a behavior-determinant lens for policy design during mass-mobility festivals under epidemic conditions, offering a theoretical and practical basis for more acceptable and effective interventions.

• Data are from online questionnaires conducted in December 2021 among migrant workers across selected regions; generalizability may be limited by sampling scope and self-report bias.
• Original datasets contain personal information and cannot be fully disclosed, constraining external validation and replication.
• The cross-sectional design near the 2022 SFT captures intentions under specific epidemic conditions, which may differ across time and locales.
• Some variables identified by Cp did not retain significance in the final model (e.g., ticket availability), indicating sensitivity to model specification and sample context.