Correction to: Nature Medicine 10.1038/s41591-020-1124-9, published online 20 October 2020.
In the version of this article initially published, due to a human error in data processing, the data for responses to “You would accept a vaccine if it were recommended by your employer and was approved safe and effective by the government” were reported incorrectly.
In the third sentence of the Abstract and in the second sentence of the first paragraph of the Discussion section, the value ‘61.4%’ for “would accept their employer’s recommendation” (Abstract) or “would get vaccinated if their employer recommended it” (Discussion) was incorrect. The correct value is 48.1% for each.
In the first sentence of the third paragraph of Results, the statements “31.9% (4,286 of 13,426) completely agreed” and “17.9% (2,411 of 13,426) somewhat or completely disagreed” were incorrect (for “You would accept a vaccine if it were recommended by your employer and was approved safe and effective by the government”). The correct text is “14.0% (1,881 of 13,426) completely agreed” and “25.9% (3,478 of 13,426 somewhat or completely disagreed” (respectively).
The fourth sentence of the fourth paragraph of Results (“The opposite trend was observed in regard to acceptance of the vaccine if one’s employer required it”) was incorrect. The correct statement is as follows: “A similar trend was observed in regard to acceptance of the vaccine if one’s employer required it, except that responses from the oldest age cohort were similar to the youngest age cohort.” At the end of the final sentence of that paragraph, the values for the business-related question (an OR of 0.87 (95% CI (0.81, 0.93))) were incorrect. The correct values are as follows: an OR of 0.93 (95% CI (0.87, 1.00)).
The final sentence of the fifth paragraph of Results (“Cases and mortality per million of a nation’s population were associated with a higher likelihood of vaccine acceptance in countries with medium and high disease incidence and mortality”) has been removed, because the corrected values do not support a uniform association between these two parameters, with variations between countries being observed.
The values at the end of the final sentence of the sixth paragraph of Results (OR = 2.34; 95% CI (2.20, 2.56)) were incorrect (for “if an individual trusted their government, they were more likely to respond positively to their employer’s vaccine recommendation than someone who did not”). The correct values are as follows: OR = 4.35; 95% CI (4.01, 4.72)).
In the second sentence of the fourth paragraph of the Discussion section, “younger respondents” was insufficiently precise. The correct description is “respondents 25–54 and 55–64 years of age.”
In Table 1, the values in the bottom five rows (for ‘Accept COVID-19 vaccine if employer recommended it (%)’) were incorrect. The corrections are as follows: for ‘Completely agree’, 1,881 (14.0) instead of 4,286 (31.9); for ‘Somewhat agree’, 4,579 (34.1) instead of 3,957 (29.5); for ‘Neutral/no opinion’, 3,488 (26.0) instead of 2,772 (20.6); for ‘Somewhat disagree’, 2,299 (17.1) instead of 1,090 (8.1); and for ‘Completely disagree’, 1,179 (8.8) instead of 1,321 (9.8).
In Table 2, the beta-coefficients and 95% CI values for the top five rows and bottom five rows in the far right column (business question) were incorrect. The corrections are as follows:
For ‘Age (years)’: 1.22 (1.10, 1.34) instead of 0.95 (0.86, 1.05), for ‘25–54 vs 18–24’; 1.17 (1.02, 1.33) instead of 0.84 (0.73, 0.96), for ‘55–64 vs 18–24’; and 1.13 (0.99, 1.29) instead of 0.78 (0.68, 0.89), for ‘65+ vs 18–24’.
For ‘Sex’: 0.93 (0.87, 1.00) instead of 0.87 (0.81, 0.93), for ‘Male vs female’; and 0.68 (0.45, 1.03) instead of 0.32 (0.21, 0.49), for ‘Other vs female’.
For ‘Cases per million population’: 1.17 (1.08, 1.27) instead of 1.30 (1.20, 1.42), for ‘Middle vs low’; and 0.74 (0.68, 0.80) instead of 1.62 (1.49, 1.76), for ‘High vs low’.
For ‘Mortality per million population’: 0.61 (0.56, 0.67) instead of 1.25 (1.15, 1.37), for ‘Middle vs low’; and 0.66 (0.61, 0.72) instead of 1.28 (1.18, 1.39), for ‘High vs low’.
For ‘Trust in government’: 4.35 (4.01, 4.72) instead of 2.34 (2.20, 2.56), for ‘Yes vs no’.
All the original source data of the surveys underlying the results of the study are available online at https://osf.io/kzq69/ and have remained unchanged since the initial publication. The corrected manuscript has been peer-reviewed and the main conclusions of the study remain unchanged.
All errors have been corrected in the HTML and PDF versions of the article.