A research paper by Arnold Barnett, Professor of Statistics at the MIT Sloan School of Management, concludes that empty middle seats may cut the risk of COVID infections in half on US commercial aircraft. The paper, which was published on July 8, 2020, has not been certified by peer review.
A few notable findings in the paper:
- The risk of contracting COVID when all seats are full is 1 in 4,300
- The risk of contracting COVID when middle seats are empty is 1 in 7,700
- The death risk from COVID when all seats are full is 1 in 540,000
- The death risk from COVID when middle seats are blocked is 1 in 770,00
- Both death estimates from COVID are higher than the risk of death in a US air crash (1 in 34 million)
- Barnett’s study focuses on transmission in a narrow-body 737 or A320 aircraft
- Barnett notes that “This paper strives for such an approximation, with an emphasis on the word “rough.”
The paper takes a statistical approach to assessing COVID infection risks. Barnett’s model incorporates:
- The probability that a passenger infected with COVID is onboard
- The risk of infection based on the locations between contagious and uninfected passengers
- The impact that masking has on preventing virus spread
There are a few important points to consider when reading this paper:
- The research has not undergone peer review
- Barnett’s study assumes that infection can only occur between passengers in the same row. Multiple studies show that virus transmission on planes occurs frequently two or more rows away from the infected host.
- A central variable in the paper is estimated transmission risk. Barnett uses a study by Chu to model the likelihood of virus transmission by distance.
- However, the Chu study relies on transmissions primarily in healthcare and household environments where subjects are likely to be seated face-to-face; it does not focus on data in an aircraft environment where subjects are seated forward-facing. This is important, as droplets are propelled forward, not sideways.
- Barnett considers only droplet transmission of COVID. He does not consider microdroplet, or aerosol transmission. His rationale for excluding microdroplets as an infection vector relies on a study of a single flight on a wide-body aircraft.
Based on the data in Barnett’s study, it appears that he underestimates the potential for COVID transmission on planes. His assumption that virus transmission occurs only within a single airplane row is not supported by any existing studies. Furthermore, the reliance on Chu to estimate the probability of transmission by distance is questionable, since Chu’s data is primarily focused on non-aircraft environments.