Misleading: Number of confirmed COVID-19 cases at Japan’s airports does not correspond to domestic cases in China
On Sept. 25, 2020, a post on Hong Kong’s popular online forum LIHKG alleged China’s reported number of COVID-19 cases was “magical,” implying that the country has been underreporting it.
The basis of the claim was a comparison between the number of Chinese and American visitors to Japan who tested positive for coronavirus between Sept. 16 and 23 at the airports in relation to the confirmed cases in each country.
The post said during that week, four flight passengers from China and 12 from the U.S. were found positive at Japan’s airports while the locally confirmed cases during the same period were 180 in China and 289,981 in the U.S.
With the assumption that visitors who departed from China were twice as many as the Americans in the said period, surmised from Japan’s tourism figures in July and August, it fallaciously calculated the proportion of infected nationals and claimed 2.2% of Chinese carrying coronavirus got “caught” in Japan while only 0.004% of American tourists were found positive, implying that the discrepancy could only be explained if far more number of Chinese people had actually been infected.
The same claim was posted on the Facebook page “H.A. Secrets,” a page about the Hospital Authority and its staff, on Sept. 25, which received more than 2,400 reactions and 525 shares by Oct. 9. Another Facebook page, “Finance Fellow” also shared the post and recorded more than 8,300 reactions and 1,300 shares.
Annie Lab found the provided calculation has no validity and the post’s claim is not supported by evidence.
Problem with datasets
Three groups of data were used in establishing the claim, namely:
- The number of passengers who tested positive for COVID-19 at Japanese airports
- The record of visitor arrivals to Japan by country
- The number of confirmed cases reported in China and the U.S.
Dataset 1: The number of passengers who tested positive at the airports in Japan seems to have come from the records published by the country’s Ministry of Health, Labour and Welfare.
While the post in question regarded the dataset as indicating the nationality of people infected with coronavirus, the tables by the ministry actually shows “travel history” (行動歴).
During the period of Sep. 16 to Sep. 23, 2020, Japan was denying entry of foreign travelers who had stayed in 159 countries/regions including China and the U.S. “within the past 14 days.”
Visa exemptions and pre-clearances had also been suspended for foreigners from 84 countries/regions including China at the same time.
Considering the extremely tight immigration control in Japan, most people listed in the tables were presumably Japanese passport holders who had been to China, the U.S. and other countries before returning home.
Dataset 2: On the other hand, the visitor figures were taken from another governmental body, the Japan National Tourism Organization. This dataset simply discloses the nationality of visitors with no information about COVID-19.
During this period, the only foreign passport holders who could enter Japan were those who met the special conditions such as being a “spouse or child of a Japanese national/permanent resident,” according to an official document by the Immigration Department.
The two datasets do not complement each other at all.
The first dataset does not show the nationalities of people who tested positive at the airports. The second dataset includes only those foreign visitors who met the special conditions.
Neither of them represents regular travelers from China or the U.S.
Dataset 3: Although the LIHKG post did not cite the sources of the last dataset, the figures of reported cases worldwide compiled by the European Centre for Disease Prevention and Control (ECDC) matched the figures in the claim.
However, ECDC says on its website that their data comes with limitations and advises that a comparison between countries should be done with caution because it may contain inconsistencies.
A case in point would be the coronavirus infections recorded in China from Sept. 16–23, which included imported cases and travelers from abroad. It does mean the origin of transmission was located outside of China, according to the explanation by the World Health Organization (WHO).
The number of cases in the U.S., on the other hand, represents new COVID-19 cases from community transmission, according to WHO reports (here and here).
The third dataset does not provide valid evidence to the comparison between China and the U.S. in the original claim, either.