Science and Data? Leading Applied Statistician Dissects Cooper & Cohen’s ‘We Know Masks Work’ Rhetoric

RALEIGH – Masks. Tired of hearing about them yet? Of course, you are. Unfortunately, Governor Roy Cooper has mandated you wear them anywhere in public, that your elementary school kids wear them, and likes to imply people not wearing them enough are selfish and reckless.

Any skepticism of masks or mandates at all is a heretical denial of science, according to Cooper. After all, his administration is fighting the coronavirus using ‘data and science.’

So what does that data and science actually say? An applied statistician, PhD, and a director with the National Association of Scholars wondered the same thing. So, he took a look, and, wouldn’t you know it, he concluded that what Cooper asserts is NOT actually what the data and science actually says.

His name is Dr. S. Stanley Young, an applied statistician, a Fellow of the American Statistical Association and AAAS, Director of the Shifting Sands Project with the National Association of Scholars, and a member of the EPA advisory board, and he wrote about this very question (and answer) in the North State Journal:

“Dr. Mandy Cohen has told us we must wear masks in many kinds of settings. She told us that wearing the masks will help “fight” the COVID virus (named by WHO as SARS-CoV-2). Gov. Cooper has told us they are relying on “data and science.”

I am a scientist. I disagree.

Not long ago, I considered the COVID data our health experts were giving us. If masks were so effective, why were we not seeing improvement in the numbers? I decided to dive into the literature.

I should point out that this line of work uses a statistical technique used by epidemiologists called a meta-analysis. I am a statistician, but it’s really not as hard as it sounds.

A meta-analysis is a study of studies. A question is posed, and studies are gathered that address the question. Do masks work? A computer search of the research that has been reported is made based on key words. The papers that come back are filtered for relevance. The selected papers are then examined in detail in hopes of pooling the data from all of the papers. By pooling the data, conclusions are sometimes possible that would not have been seen in any one of the studies.

Statistics are used to determine if the summary result was simply due to randomness or whether a specific cause is indicated. A simple example can be helpful.

Imagine tossing a coin 30 times. A fair coin will result in heads about 50% of the time. But what if you suspect the coin you have is loaded — that is, it is not a fair coin? You suspect it because you toss the coin 30 times and you get heads only 40% of the time. Statistics can be used to answer your question — is your coin significantly different than the “fair” coin? Through meta-analysis, we can include multiple studies, so instead of 30 tosses, we can include the results for all the studies used in the meta-analysis studies.

The same analysis can be applied to the studies on wearing a mask. The question we are addressing is whether wearing a mask is any different than not wearing a mask? Instead of tossing a coin, of course, we examine results from randomized clinical trials (RCT). Using a meta-analysis allows us to use as much of the available RCT data as possible. […]

[CONTINUE READING]