Experimental Design
This study implements a randomized controlled trial (RCT) to examine how different types of informational nudges influence influenza vaccination intentions among food delivery workers in urban China—a high-exposure, high-mobility group with persistently low vaccination rates.
Target Population and Recruitment: Participants are active food delivery workers recruited via online rider platforms and local delivery worker networks across selected cities. The study is embedded within a custom-designed online survey platform optimized for smartphone use. Respondents complete a baseline module capturing demographics, work patterns, prior vaccination behavior, and health attitudes.
Intervention Structure: Immediately following the baseline module, participants are randomly assigned to one of four treatment arms, each receiving a distinct informational message embedded within the survey. The randomization is at the individual level and stratified by city to ensure geographic balance.
Information Treatments: There are four different information intervention.
1. Income Risk Frame: participants will read a message from the online platform that says "Data from the Chinese CDC showing that influenza infection leads to 3–8 days of virus shedding (average of 5 days). Based on average rider income in second- and third-tier cities, a single flu episode is estimated to cause a loss of ~1,000 RMB due to missed workdays." This helps to emphasize financial consequences of illness-related absenteeism.
2. Health Risk to Self and Family Frame: participants will read a message from the online platform that says "WHO estimates that influenza causes 3–5 million severe cases and 290,000–650,000 deaths annually. Describes higher risk to children, the elderly, pregnant women, and those with chronic conditions." This message emphasizes protective benefits of vaccination for family and vulnerable groups.
3. Side Effect Reassurance Frame: participants will see "According to Chinese CDC, flu vaccines are 40–60% effective and generally safe. Lists typical side effects (e.g., mild fever, soreness) as rare, short-lived (1–2 days), and self-limiting. Severe reactions are extremely rare.", this message reduces vaccine hesitancy driven by fear of side effects.
4. Social Norm (Peer Behavior) Frame: participants will see "In your city, over 300 community residents including delivery riders have already voluntarily received the flu vaccine this season.” This intervention higlights the role of social norms or peer effect in changing the perceived vaccination behavior among others.
Control group will receive no information intervention when they enter the platform.
The primary outcome is the respondent's stated willingness to receive a flu vaccine, measured immediately after the information exposure using a Likert scale and a binary intention question. We also collect secondary outcomes including stated timing of vaccination, willingness to pay, and perceived risks. We estimate treatment effects by comparing average outcomes across the four groups using OLS regressions, controlling for baseline covariates (e.g., age, gender, prior vaccine experience and vaccine willingness etc.) to improve precision. Heterogeneous treatment effects will be explored by income level, prior vaccine history, and household composition. Robustness checks include placebo tests using unrelated attitudinal questions and balance checks across treatment arms.