Saturday, June 27, 2020

Degree of urbanization is at least as important as masks

Here is an article telling us that wearing masks helps defeat the spread of Covid-19.

As somebody who used data to solve problems for much of my working life, I can tell you that the degree of scatter suggests a very, very weak relationship.

One hopes for a narrow "sausage" cloud of data where one can move along the horizontal axis and accurately predict the vertical value of the observed data.

We see the weakness of the assumed relationship when we look at the data around 55% mask wearing. In Illinios, for instance, every new patient can be expected to only infect 0.8 additional patients which should result in the case load to collapse. On the other hand, Nevada shows 1.5 additional patients for the very same degree of mask-wearing. A difference of almost 2X.

One technique for "finding" other variables is to look for the outliers. According to the "masks save lives" hypotheses Nevada, Texas and Florida are outliers. One would expect them to have lower transmission rates. Another set of outliers in the kitty-corner quadrant are Indiana, Kentucky, Kansas and Iowa. One would expect them to have higher transmission rates.

The numbers in bold, white font are the degree of urbanization of those states.

Nevada is as urbanized as New Jersey which is already on the backside of the first peak.

Suggesting that urban areas inherently present a higher risk for the spread of communicable diseases should surprise nobody.


1 comment: