For the ninth day of mask mess, we got these:

Lyu and Wehby

Community Use of Face Masks And COVID-19: Evidence From A Natural Experiment Of State Mandates In The US.Health Affairs (2020).

Published on August 20 but with a “Fast Track Ahead of Print Version” posted much earlier, this study is very important to the Cooper administration. Cooper cited it first among the five “examples of studies studying the effectiveness of face coverings” that he included in his Nov. 23 executive order tightening his mask order against people. It was among the three presented before the General Assembly by state health bureaucrat Mandy Cohen on June 17, a week before Cooper first leveled his mask order against people. In August I wrote about that study in the context of North Carolina.

Lyu and Wehby identified 15 states and DC with executive orders or directives issued between April 8 and May 15 that mandate mask-wearing. They used the preceding five days before announcing the mask mandates as the baseline period for the trends in those states in order to build a predictive model for cases numbers if those states did not mandate mask-wearing. They then compared the differences in actual case numbers with the predicted, no-mask case numbers and estimated “daily case rate declines by 0.9, 1.1, 1.4, 1.7, and 2.0 percentage-points within 1-5, 6-10, 11-15, and 16-20, and 21+ days after signing [the mask orders], respectively.” From that they declare, “The main model estimates suggest that as many as 230,000–450,000 cases may have been averted due to these mandates by May 22.”

Note how Lyu and Wehby hedge such a bold-sounding finding. It’s based on their own model, the upper-bound estimates (“as many as”) range widely, and the estimates suggest that cases may have been averted. Lyu and Wehby themselves add that, “Estimates of averted cases should be viewed cautiously and only as general approximations.”

They found that the effects of the mask orders were larger when measured from the dates they were announced and signed, rather than the dates they took effect (typically 2-3 days later), which they said suggested “earlier compliance.” They also state that their “estimates suggest increasing effectiveness and benefits from these mandates over time.”

Time has not been as impressed with those estimates’ suggestions as Cooper has. Lyu and Wehby’s early-mandate states’ graphs (and North Carolina’s) are not showing evidence of “increasing effectiveness over time.”

Lyu and Wehby include an interesting side finding regarding governments ordering employers to mandate mask-wearing among employees: “We find no evidence for effects of states mandating employee face mask use.” They posit that it’s because “many businesses have been requiring their employees to wear masks” without a government order. Rather than conclude that such an order is redundant, however, they say it “may be reinforcing what many businesses are already choosing to do.”

Lyu and Wehby conclude with a might: “requiring use of face masks in public might help in reducing COVID-19 spread.”

Does this study support Cooper’s extreme exercise of power?


Probable effects (estimates “suggest,” “may have been,” “might” reduce, etc.) are insufficient for extreme emergency orders. Studies with model predictions of effects “over time” that turn out to be increasingly elusive over time are no basis for extreme emergency orders, especially not over time.

Estimates that should be “viewed cautiously” cannot be the basis of extreme emergency orders.

Eikenberry et al.

To mask or not to mask: Modeling the potential for face mask use by the general public to curtail the COVID-19 pandemic. Infect Dis Model (2020);5:293-308.

Received on April 6, 2020, this study builds a mathematical model for “assessing the community-wide impact of mask use by the general, asymptomatic public, a portion of which may be asymptomatically infectious.” Eikenberry et al. note from the outset that “Mask use in public spaces has been controversial in the US” but “far more prevalent in many Asian countries.” They project that an 80 percent adoption of wearing “moderately effective” masks “could prevent” some range of projected deaths in Washington and New York. They cite a great amount of uncertainty in the research literature about the effectiveness of masks, especially the cloth masks as required by Cooper’s order, but despite it all they write:

In summary, the benefit to routine face mask use by the general public during the COVID-19 pandemic remains uncertain, but our initial mathematical modeling work suggests a possible strong potentialbenefit to near universal adoption of even weakly effective homemade masks that may synergize with, not replace, other control and mitigation measures.

In one sentence, Eikenberry et al. balance “remains uncertain” with “suggests,” “possible,” “potential,” and “may” — according to their model.

About that uncertainty: they note that homemade masks “may” afford “significant, although variable and generally lesser, protection” than medical-grade masks; that “a range of masks provide at least some protective value”; that “It is largely unknown to what degree homemade masks … may protect against droplets/aerosols and viral transmission”; that an experiment “suggest[s] that while homemade masks were less effective than surgical masks, they were markedly superior to no mask”; and that a clinical trial “showed relatively poor performance for cloth masks relative to medical masks.”

Eikenberry et al. later refer again to “appreciable uncertainty.” Their modeling framework finds little possible effect from masks alone, which leads them to argue for masks as a complement to other public health control measures.

There is so much uncertainty. Eikenberry et al. write:

Our theoretical results still must be interpreted with caution owing to a combination of potentially high rates of noncompliance with mask use in the community, uncertainty with respect to the intrinsic effectiveness of (especially homemade) masks at blocking respiratory droplets and/or aerosols, and even surprising amounts of uncertainty regarding the basic mechanisms for respiratory infection transmission.

Beset with so much uncertainty, the authors resort to the Precautionary Principle for promoting “nearly universal” mask adoption: “Despite uncertainty, the potential for benefit, the lack of obvious harm, and the precautionary principle lead us to strongly recommend as close to universal (homemade, unless medical masks can be used without diverting healthcare supply) mask use by the general public as possible.”

Does this study support Cooper’s extreme exercise of power?


Regardless of how desirable a policymaker may view a particular action, a mathematical model’s “theoretical results” with so much admitted “uncertainty” surrounding them that they should be “interpreted with caution” can never be sufficient for extreme emergency orders. Such orders are the cardinal opposite of interpreting with caution.

It should also be self-evident that the Precautionary Principle is no basis for extreme emergency orders.

Click for more of the Twelve Days of Mask Mess series.