Attached. Missing Arkansas, Connecticut, Florida, Hawaii, Iowa, Kansas, Maryland, Missouri, New York, Pennsylvania, West Virginia, Wyoming.
On 8/10/21 4:43 PM, David Eric Smith wrote: > I am sure it is just dieseling at this point, but I was pleased to see the > following article: > https://www.nytimes.com/interactive/2021/08/10/us/covid-breakthrough-infections-vaccines.html > > <https://www.nytimes.com/interactive/2021/08/10/us/covid-breakthrough-infections-vaccines.html> > (I usually get to these things late; y’all probably have read it already) > > In reading the first table, on hospitalization and death fractions by > vax/unvax, I was thinking “okay, now since we have vaccinated fractions by > date, we could do a covariance plot, and of course could then do more > involved multiple regressions on dummy variables as we could find them.” (No > pun meant on “dummy variable”, though I am unable to miss it myself. Things > like measures of hospital performance, coverage of masking rules or other > public health measures, population density and gathering density, etc. Some > of these to be proxies for fraction exposed, which is hard to get at.) > > But then that is just where the article goes. It’s funny how a pair made of > a careful writer and a lazy reader can be an unhelpful combination. The text > leading to the second table says "people who were not fully vaccinated were > hospitalized with Covid-19 at least five times more often than fully > vaccinated people, according to the analysis, and they died at least eight > times more often.” I remember the nice passage in John Paulos’s book > “Innumeracy”, where (to make some point, which I now forget), he comments on > why a sign over the highway “Entering New York, Population at least 6” is not > particularly informative, though quite true. > > Look then at the distribution of multipliers in the table. For the “at least > five times” column, the first six entries, alphabetically, are 75x, 17x, 47x, > 68x, 22, 148x, 161x, and likewise for the “eight times” column. Ahh, if the > American Public would only tolerate being shown a histogram giving the whole > distribution at a glance…. Of course, if I were not lazy, I could find and > download the data and make my own histogram. > > But, credit to those authors. Within the bounds of what is permitted to > them, this is a useful data digest. > > Eric -- ☤>$ uǝlƃ
- .... . -..-. . -. -.. -..-. .. ... -..-. .... . .-. . FRIAM Applied Complexity Group listserv Zoom Fridays 9:30a-12p Mtn GMT-6 bit.ly/virtualfriam un/subscribe http://redfish.com/mailman/listinfo/friam_redfish.com FRIAM-COMIC http://friam-comic.blogspot.com/ archives: http://friam.471366.n2.nabble.com/
