There are various ways to analyze the controversial question of whether wearing a mask helps prevent transmission of the Wuhan coronavirus. One way is to examine what happened after mask mandates were imposed in various jurisdictions and compare it to what happened in jurisdictions that declined to impose that mandate.
This study used that approach. It considered the top 15 states in terms of coronavirus infections in the period from March 1 through May 18, 2020. Nine of these states imposed a mask-wearing requirement. Six did not.
The study found that seven of the states with that requirement saw a flattening of the infection curve afterwards. The six holdout states did not fare as well.
The study estimates, perhaps ambitiously, that “the total infections prevented by face covering reache[d] ~252,000 on May 18 in seven states.” The authors conclude that “face covering play[ed] the dominant role in. . .flattening the total infection curve.”
Other studies using a similar approach have reached the same conclusion. The CDC studied differences in Kansas between counties that imposed a mask-wearing requirement and those that didn’t. It found that after July 3, 2020, when the governor mandated the wearing of masks, coronavirus incidence decreased in the 24 counties that followed the mandate, but continued to increase in the 81 counties that didn’t.
There are, I’m sure, imperfections in these studies. However, this type of before/after study seems much more probative than studies that simply compare overall rates of infections and/or death counts in mask-mandate states and states with no such mandate.
Comparing counts in states like New York with those in states like South Dakota, for example, doesn’t seem illuminating, given the differences in demography, exposure to foreign travelers, etc. Comparing what happened within states or counties after mask-wearing requirements were imposed (or rejected) seems more scientifically sound.
Studies like the two discussed above indicate to me that masks, and mask mandates, probably do help prevent the spread of the coronavirus to some degree, or did in 2020. Whether masks still serve this purpose to an appreciable degree now that a goodly percentage of the population has been vaccinated is another question. Whether people who have been vaccinated need to wear masks is yet another.
UPDATE: Donald Luskin, one of my favorite analysts going back to the days when he regularly debunked Paul Krugman, has studied the effectiveness of wearing masks (as opposed to mask mandates) in preventing infections. He employed the same basic methodology he uses to dispute the effectiveness of lockdowns.
Luskin found strong evidence that, unlike lockdowns, masks are effective. You can listen to his nuanced explanation (about 7 minutes) here.
Regarding Kansas, another study, this one by the University of Kansas from last October, found a big drop in cases in counties that adopted a mass-wearing requirement, but not in counties that didn’t. However, a reader has done his own analysis of the current county-by-county data and based on his review, tells me that “the evidence is now clear that mask mandates had no impact” in Kansas. (He distinguishes mask mandates from mask wearing, which he believes probably does prevent transmission in some cases.)