COVID-19 health certification reduces outgroup bias: evidence from a conjoint experiment in Japan

This study examines whether COVID-19 health certifications (vaccination proof or a recent negative test) can reduce outgroup bias that arises under health threat. Psychological and behavioral immune system (BIS) theories predict that infectious threats heighten negative attitudes toward outgroups. During COVID-19, increased prejudice and xenophobia toward racial/ethnic minorities, immigrants, and tourists have been widely documented. The research question is whether providing health certification to inbound travelers reduces this bias among host-country residents. Focusing on Japan—a context with documented aversion to foreigners and evolving policy debates over vaccine passports—the study evaluates if certificates mitigate perceived infection threat and, consequently, outgroup bias toward inbound foreigners (including immigrants and tourists). The work has policy relevance for balancing border controls with public acceptance during pandemics.

The BIS framework links pathogen threats to avoidance and negative attitudes toward outgroups, mediated by perceived vulnerability to disease, germ aversion, and disgust sensitivity. Numerous studies during COVID-19 documented increased prejudice toward racial/ethnic groups, immigrants, and tourists. Fewer studies explore mitigating factors. Relevant work shows: (1) cultural values (high individualism/low collectivism) condition threat effects on xenophobia (Kim et al., 2016); (2) avoidance of immigrants due to perceived non-adherence to local norms (Karinen et al., 2019); and (3) avoidance is triggered by high pathogen risk irrespective of group membership (van Leeuwen & Petersen, 2018). While health certification has been studied regarding public responses and behavior, its role in reducing outgroup bias remains underexamined. The authors hypothesize that health certification will reduce outgroup bias by lowering perceived infection risk associated with certified individuals, thereby attenuating nationality-based bias.

Design: A pre-registered, population-based discrete choice experiment using a randomized conjoint design was conducted in Japan (Feb 22–24, 2021) during the second State of Emergency and in the run-up to the Tokyo Olympics/Paralympics. The study framed inbound travelers from abroad (foreigners, immigrants, tourists) as the outgroup during a pandemic and tested whether health certification mitigates outgroup bias toward them. Sample and procedure: Participants were recruited online via Rakuten Insight Inc. with quota sampling for national representativeness by gender, age, and prefecture. Of 2061 who started, 1066 completed (female 51.2%; mean age 48.6, SD 16.51; 12.6% Tokyo residents). Exclusions included failed attention checks, dropouts, and certain professions/stakeholders. Respondents evaluated two paired traveler profiles across two tasks (yielding ~4000 observations planned). Compensation was via points. IRB approvals were obtained; informed consent collected. Power analysis (cjpowR; α=0.05, β=0.2) pre-registered (Supplementary Note 3). Conjoint attributes: Ten attributes with multiple levels formed 864,000 potential profile combinations (3×6×5×5×5×2×4×3×4×4), randomized per respondent with minor constraints for clarity. Key attributes: (1) Nationality (Japanese; foreigner with permanent residency; foreigner without permanent residency); (2) Quarantine_Certificate (vaccination certificate; negative COVID-19 test; moving plan; self-isolation; nothing); (3) Region (Taiwan vs. regions with widespread infection: China, South Korea, US, Brazil); (4) Duration of stay (1 month; 3 months; 1 year; 4 years; indefinite); (5) Purpose (Tokyo Olympics; migration; study; business; tourism); (6) Sex; (7) Age (e.g., 65 years); (8) Speaking Japanese (fluently; somehow; with interpreter); (9) Education (graduate; university; high school; junior high); (10) Income (¥9M; ¥6M; ¥3M; none). The analysis focused on Nationality and Quarantine_Certificate and their interaction. Outcome variables: (a) Choice (3 options): choose Traveler A/B/Neither; (b) Rating: 1–7 admission scale recoded to 0–1 for each profile; (c) Conditional forced choice (2 options) asked only if “Neither” was chosen in (a). Subgroups (pre-registered): Risk perceptions—Infection Risk (High ≥ median 30 vs Low), Serious Risk (High ≥ median 17 vs Low), Discomfort near sneezing masked person (High ≥ median vs Low); Associations with foreigners—Foreigners’ Contacts (Yes/No) and Foreigners’ Favorability (High ≥ median vs Low); Political partisanship—Conservative Support (LDP/Komeito/JIP vs Others) and Conservative Temperature (High vs Low). Estimation: Linear regression models estimated marginal means (MMs) and uniform average marginal component effects (uAMCEs) with standard errors clustered by respondent. Significance threshold p<0.05 (two-tailed); no multiple-testing corrections (per pre-registration). Analyses used R (cregg). Data and code: OSF doi:10.17605/OSF.IO/342UH.

- Evidence of outgroup bias by nationality: Compared to Japanese nationals, foreigners were less likely to be admitted. In uAMCE terms, respondents were 12% less likely to admit foreigners without permanent residency and 5% less likely to admit foreigners with permanent residency, relative to Japanese (choice, 3-option). - Strong positive effects of health certification: Relative to having “Nothing,” self-isolation (+27%), vaccination certificate (+31%), and negative test certificate (+27%) all significantly increased admission probability (choice, 3-option). In rating outcomes, effects were similar. In forced choice (2-option), vaccination certificate was as effective as self-isolation, while negative test resembled moving plan in effect size. - Absolute probabilities (Fig. 1, Panel a): Travelers with a vaccination certificate had the highest admission probability at 44%; negative certificate and self-isolation followed; travelers with “Nothing” were admitted only 10% of the time. - Interaction mitigating outgroup bias (Fig. 3): Health certification reduced gaps between Japanese and foreigners with permanent residency. Example (Panel a, 3-option): Japanese with self-isolation 51%; Japanese with vaccination 49%; foreigners with PR + vaccination 47%; foreigners with PR + negative 46%; Japanese with negative 41%. In forced choice (2-option), vaccination certificates equalized effects for Japanese and foreigners with PR; negative certificates showed weaker or no equalizing effect. - Other attribute effects: Region—Taiwan (no major outbreak at survey time) had highest acceptance (42%), China lowest (26%) in 3-option; in 2-option, US most likely (59%). Duration—shorter stays (1 month) more acceptable than indefinite (contrary to preregistration). Purpose—Tokyo Olympics travelers (40%) more accepted than tourists (23%). Lower Japanese proficiency, education, or income reduced acceptance in 3-option outcomes (not robust in rating or 2-option). - Heterogeneity: Risk perceptions did not condition vaccination certificate effects, but higher Serious Risk and Discomfort reduced the effect of negative certificates (i.e., negative certificates were less effective among those with higher perceived severity or discomfort). Associations with foreigners showed little conditioning of certification effects; higher general favorability to foreigners raised acceptance overall. Conservative partisanship: conservative supporters and those with higher conservative temperature were more likely to admit travelers with negative certificates; higher conservative temperature also associated with greater acceptance of vaccination certificates.

The study directly addresses whether health certifications reduce outgroup bias during a pandemic. Conjoint results show that vaccination and negative test certificates substantially increase willingness to admit inbound travelers and attenuate nationality-based disparities, particularly between Japanese and foreigners with permanent residency. These findings align with BIS theory emphasizing avoidance of high-pathogen-risk individuals rather than outgroups per se. Policy implications are notable: vaccine passports can enhance public acceptance of inbound travelers and reduce nationality-based bias; negative test certificates can serve as an alternative pathway, though their effect is more context-sensitive (e.g., conditioned by perceived severity and political conservatism). Overall, health certification functions as a psychological signal mitigating perceived infection threat and, consequently, outgroup bias in border admission preferences.

Health certification—especially vaccination proof—reduces negative bias against inbound travelers during COVID-19 and can bring acceptance of foreigners with permanent residency closer to that of nationals. The results support implementing vaccine passports alongside negative test options to reopen borders while addressing equity concerns. For future research, the authors suggest: examining mechanisms (e.g., trust in government, science, and healthcare systems), assessing policy context effects, conducting longitudinal and cross-national studies, and improving external validity by aligning conjoint profile distributions with real-world prevalence.

- Mechanisms not fully examined: variables such as trust in government, science, or healthcare systems were not modeled; recent policy changes may have influenced attitudes. - Context specificity: data collected during Japan’s second State of Emergency and pre-Tokyo Olympics/Paralympics; results may vary over time or across countries; longitudinal/cross-sectional validation needed. - Conjoint profile distribution: uniform level distribution may limit external validity relative to proportional (real-world) distributions; replication or modified designs advised.