Masks seem like they couldn’t possibly work. How could wearing a little piece of cloth possibly be effective?
However, there have been a lot of real world studies that masks do actually work at reducing infections. Masks are not perfect, but they do significantly lower the risk of the wearer and the people around the wearer. Here are some studies:
- This study from Norway (2024-05-29) did it, and found that people who were asked to wear a surgical face mask in public places for 14 days were 29% less likely to report symptoms consistent with a respiratory infection than those who were asked to not wear a mask in public.
- This HUGE study in Bangladesh (2021-09-09) sent people out to villages to stand around and remind people to wear masks. The reminder-people had zero enforcement authority, all they did was remind people to mask up. The compliance, even with reminders, still wasn’t great (42% in those villages, vs. 13% in non-reminder villages), but there was a statistically significant difference in how many cases and deaths villages had depending on whether they had reminder-people or not.
- This study from Arizona (2021-10-01) found that schools without mask mandates were 3.5 times more likely to have outbreaks than schools that had mask mandates.
- This study from USA (2021-10-01) found that children’s case rates were 2.1x higher in counties without mask mandates than with.
- The US CDC used to have a page (found in their archive dated 2023-06-28) that listed a ton of studies, but now their web site says almost nothing about masking. @@@ go through and make sure I’ve got all of the studies @@@
- This review article (2021-01-11) cites a lot of studies and concludes that masking works.
- This paper (2020-10-15) reports that 66% of hamsters in a cage adjacent to a cage with COVID-infected hamsters got infected themselves. (There was one infected hamsters in one cage and three uninfected hamsters in the other cage. Two exposed hamsters were examined after five days, and the other at seven days.) When a barrier of surgical mask-material was placed in between the pair of cages, only 12.5% of the exposed hamsters got sick when the fabric was placed to simulate the sick hamster “wearing a mask”, and 37.5% of the exposed hamsters got sick when the fabric placement simulated the exposed hamsters “wearing a mask”; the “masked” hamsters also got less sick than the “unmasked” hamsters in the first part of the study.
- This paper from South Korea (2024-10-14) reports that COVID-19 levels was much much lower in speaking trials when the speakers wore N95 masks than if they did not. However, they still found significant amounts of COVID-19 in the air (although they admit it might have been in the room air before they started testing), and urge people (as much as one can in an academic article) to do two-way masking.
- This article from Missouri (2021-12-01) reports on a study looked at the difference between counties in Missouri which had mask mandates and those which did not, and found that the risk of COVID-19 was 37% higher in counties which did not have mask mandates.. (See graph below.)
From the CDC archive of masks information:
- This study from the US (2020-06-12) of an outbreak aboard the USS Theodore Roosevelt, an environment notable for congregate living quarters and close working environments, found that use of face coverings on-board was associated with a 70% reduced risk of infection.
- This study from China (2020-05) reported that in 124 Beijing households with more than one 1 laboratory-confirmed case of SARS-CoV-2 infection, mask use by the index patient and family contacts before the index patient developed symptoms reduced secondary transmission within the households by 79%.
- This study from Massachusetts (2023-08-04) examining SARS-CoV-2 secondary attack rates among eight public K-12 school districts in Massachusetts (70 schools with >33,000 enrolled students) during the 2020–21 school year found an unadjusted secondary attack rate of 11.7% for unmasked versus 1.7% for masked interactions.40
- This paper from Thailand (2020-11) reports that, among more than 1,000 persons interviewed as part of contact tracing investigations, those who reported having always worn a mask during high-risk exposures experienced a greater than 70% reduced risk of acquiring infection compared with persons who did not wear masks under these circumstances.
- This case study from USA (2020-07-17) found that when 67 clients got haircuts from two symptomatic hairdressers where everybody wore a mask, nobody got sick.
- This paper from Australia (2017-08-11) found that wearing N95s all the time was better than wearing N95s in targeted situations, which was better than wearing medical masks, which was better than not wearing masks. N95-all-the-time cut the risk of viral/bacterial respiratory diseases by about half to three-quarters.
- This paper from Australia and China (2013-01-24) also reported that N95s all the time was better than wearing N95s in targeted situations, which was better than wearing medical masks. People got clinical respiratory illnesses 2.36 more often wearing medical masks than N95-all-the-time and bacterial infections 1.45 times more often. Targeted N95 use wasn’t any better than medical masks; you have to wear them all the time.
Effectiveness of different masks
This paper (2025-04-21) reports on actual measurements of various types of masks. They found that filtration efficiency (FE) of particles 0.02–1 µm in size, without fit testing, was as follows:
| Mask type | FE |
| cloth mask with earloops | 47 ± 5% |
| double cloth masks with both using earloops | 47 ± 5% |
| cloth mask with overhead ties | 55 ± 6% |
| double cloth mask with overmask using overhead ties | 66 ± 6% |
| level 1 masks | 56 ± 9% |
| level 1 masks with cloth-overhead-ties overmask | 84 ± 10% |
| level 1 masks with silicone or Fix-the-Mask brace | 87–91% |
| level 3 masks | 70 ± 13% |
| level 3 masks with cloth-earloops overmask | 84 ± 13% |
| level 3 masks with cloth-overhead-ties overmask | 92 ± 3% |
| level 3 mask with brace | 92–94% |
| non-certified KN95 look-alike | 57-77% |
| two different KF94s | 57-77% |
| three KN95s | 57-77% |
| N95 | 97-98% |
| CaN99 | 97-98% |
Level 1 and 3 are different quality levels of what are commonly called “baggy blues”.
The cloth mask they tested was the “Essex” pleated mask type.
Wearing ear guards (which I think are these things), scrub cap, or by knot-and-tuck didn’t change the FE for L1 or L3 masks (“baggy blues”).
The authors specifically called attention to the fact that a well-fitting cloth mask has filtration comparable to an L1 mask. Apparently the better edge seal of the cloth masks makes up for the leakier fabric.
The authors also called attention to how the N95s and CaN99s were significantly better than the KN94/5s. They tested the material and the material was fine, so it must be the fit. KN94-style have earloops while the N95-type has overhead straps. This is consistent with this paper (2021-05-18) which I talked about in the 2025-03-15 General post which said that overhead straps were way better than earloops (for beared men, true but it probably tracks).
I was surprised at how effective the L1 and L3 masks were. 56 ± 9% isn’t that great, but it’s a whole lot better than zero.