The survey recruited a total of 2,155 respondents, each of whom was given four hypothetical scenarios and asked if they would accept the implementation of policies described in them. Subsequently, we extracted and analyzed 2,155 × 4 = 8,620 results. The respondents represented a panel composed of members of the public (N = 2,155). Table 2 shows the demographic distributions of the study samples compared with the Japanese population. The average age was 45.0 ± 14.1 years.
Table 2
Demographics of respondents
| | Number of individuals | Weighted (%) | Japanese population (%) |
Number | 2155 | | |
Sex | | | |
| Male | 1069 | 49.6 | 48.6 |
| Female | 1086 | 50.4 | 51.4 |
Age (years) | | | | |
| 20–29 | 420 | 19.5 | 10.1 |
| 30–39 | 429 | 19.9 | 11 |
| 40–49 | 429 | 19.9 | 14.2 |
| 50–59 | 436 | 20.2 | 13.5 |
| 60–69 | 441 | 20.5 | 12.2 |
Occupation1 | | | |
| Full-time worker (regular employment) | 852 | 39.5 | 35.7 |
| Part-time/non-regular employment worker | 532 | 24.7 | 18.6 |
| Self-employed | 150 | 7 | 6.1 |
| Household worker/unemployed/student | 621 | 28.9 | 39.6 |
Household income (JPY)2 | | | |
| < 2 million | 252 | 11.7 | 19 |
| 2 to 6 million | 1099 | 51 | 45.6 |
| 6 to 10 million | 560 | 26 | 23.2 |
| 10 million or more | 244 | 11.3 | 12.1 |
Routine visits to health care facility (more than monthly) | | | |
| Yes | 520 | 24.1 | |
| No | 1635 | 75.9 | |
Impact of COVID-19 on household income | | | |
| Higher income in 2020 than in 2019 | 137 | 6.4 | |
| Unchanged | 1157 | 53.7 | |
| Lower income in 2020 than in 2019 | 861 | 40 | |
Fear of unknown infectious disease | | | |
| Very scared | 1065 | 49.4 | |
| Fairly scared but currently not very scared | 724 | 33.6 | |
| Not scared | 268 | 12.4 | |
| Not scared at all | 98 | 4.5 | |
1 The demographics regarding occupation were included for those aged over 15 years (Ref) |
2 Income levels were classified by every million JPY. Income data were merged based on the annual average income in Japan, which was JPY 5.5 million. |
[Table 2 here]
Table 3 shows the results of the public preferences for policy measures. Most of the respondents affirmed the tendency to accept a decrease in income to control the spread of infection. However, in situations where the number of infected persons was low, they might be more likely to accept some increase in infection as a trade-off against an increase in income.
Table 3
Probability of respondents answering "YES (Accept)" for 12 scenarios
Attributes | Scenario NO. |
1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 |
Overall | 24.1% | 20.3% | 19.9% | 26.4% | 25.3% | 22.8% | 17.8% | 23.3% | 38.0% | 21.1% | 30.8% | 42.8% |
Pattern 1 (100,000 patients/5,000 deaths) | 26.0% | 19.2% | 21.6% | 31.1% | 29.0% | 24.9% | 15.7% | 19.2% | 37.7% | 21.7% | 24.9% | 41.3% |
Pattern 2 (1,000,000 patients/50,000 deaths) | 26.6% | 20.7% | 19.3% | 26.4% | 23.3% | 22.5% | 17.9% | 27.4% | 37.2% | 20.3% | 33.0% | 39.5% |
Pattern 3 (10,000,000 patients/500,000 deaths) | 19.9% | 21.3% | 19.1% | 22.0% | 23.3% | 20.9% | 20.2% | 23.5% | 39.1% | 21.5% | 34.5% | 47.3% |
Table 4 shows the results of the conjoint analysis using various models. Both the income change (positive value for acceptance) and multiplier for the number of infected individuals (negative value for acceptance) had a statistically significant impact on the probability of acceptance. According to the simplest model (Table 4, Model A), the coefficient for income increase was 0.544 (for JPY 1 million, 95%CI: 0.460–0.627) and that for the multiplier for infected individuals was − 0.346 (natural log scale, 95%CI: -0.391 – -0.301). A comparison of the magnitudes of the two attributes revealed that the public was 4.81 times (95%CI: 3.24 to 8.01) more likely to accept an increase in infections/deaths in exchange for an increase of JPY 1 million in their annual income.
The attitude toward the pandemic status, or fear levels, had a significant impact on the preference between income change and number of infections (Tables 5 and 6). The coefficients for the infection multiplier for those with (N = 1,065) and without (N = 1,090) extreme fear were − 0.444 (95%CI: -0.510 to -0.378) and − 0.256 (95%CI: -0.318 to -0.194), respectively (Tables 5 and 6). Thus, those with extreme fear were less likely to accept a policy that could lead to an increase in infections. The coefficients for the change in income were similar for the two subgroups.
We conducted another analysis (Tables 4, 5, and 6; Models B, B1, and B2), including an interaction factor, to assess the impact of the baseline prevalence rate of the infectious disease. This dummy variable was set to 1 in the case of the proposed policy leading to an increase in the number of patients in the scenario of an already high prevalence (10 million patients infected). According to the analysis with Model B, respondents were less likely to accept an increase in infections/deaths under the high prevalence scenario. The coefficients for overall respondents, those who with extreme fear, and those without, were − 0.465 (95%CI: -0.687 - -0.243), -0.251 (95%CI: -0.572–0.070), and − 0.634 (95%CI: -0.946 - -0.322), respectively. The impact of change in the baseline risk, or initial number of infectants/deaths was higher for those without fear of pandemic.
Table 4
Results on the general panel (N = 2,155)
| Model A (original) | | Model B |
| Coefficient | Lower 95%CI | Upper 95%CI | p value | | Coefficient | Lower 95%CI | Upper 95%CI | p value |
Income change (JPY 1 million) | 0.544 | 0.460 | 0.627 | 0 | | 0.564 | 0.480 | 0.649 | 0 |
Multiplier for number of infections (Ln) | -0.346 | -0.391 | -0.301 | 0 | | -0.327 | -0.373 | -0.282 | 0 |
High prevalence (10 million patients infected) scenario | | | | | | -0.465 | -0.686 | -0.243 | 0.008 |
Constant | -1.47 | -1.60 | -1.35 | 0 | | -1.38 | -1.51 | -1.25 | 0 |
Table 5
Results of the respondents with extreme fear for the unknown virus (N = 1,065)
| Model A1 | | Model B1 |
| Coefficient | Lower 95%CI | Upper 95%CI | p value | | Coefficient | Lower 95%CI | Upper 95%CI | p value |
Income change (JPY 1 million) | 0.537 | 0.417 | 0.656 | 0 | | 0.549 | 0.428 | 0.670 | 0 |
Multiplier for number of infection (Ln) | -0.444 | -0.510 | -0.378 | 0 | | -0.433 | -0.500 | -0.366 | 0 |
High prevalence (10 million patients infected) scenario | | | | | | -0.251 | -0.572 | 0.070 | 0.126 |
Constant | -1.62 | -1.81 | -1.44 | 0 | | -1.57 | -1.76 | -1.37 | 0 |
Table 6
Results of the respondents without fear (N = 1,090)
| Model A2 | | Model B2 |
| Coefficient | Lower 95%CI | Upper 95%CI | p value | | Coefficient | Lower 95%CI | Upper 95%CI | p value |
Income change (JPY 1 million) | 0.549 | 0.431 | 0.666 | 0 | | 0.575 | 0.457 | 0.694 | 0 |
Multiplier for number of infection (Ln) | -0.256 | -0.318 | -0.194 | 0 | | -0.230 | -0.294 | -0.167 | 0 |
High prevalence (10 million patients infected) scenario | | | | | | -0.634 | -0.946 | -0.322 | 0 |
Constant | -1.35 | -1.53 | -1.18 | 0 | | -1.23 | -1.41 | -1.05 | 0 |