Source: William and Mary COVID-19 dashboard ,downloaded 10/24/2020.
Is my daughter safer from COVID-19 while at college, than she would be at home? That’s the point of today’s calculation. And the answer is a qualified “yes”.
You’ll see a lot of statistics regarding the number of COVID-19 infections within some defined population. A few weeks back, the news was that nearly 20,000 Amazon.com workers had contracted COVID-19. Back in June, it was 11,500 grocery store workers that had been infected. Separately, you will see ongoing statistics on the large fraction of health care and first responder personnel who have been infected.
These are nice bits of click-bait, and do their job of getting people mindlessly angry at somebody. Throw a big number in your face, get you to respond with your emotions and not your brain, and they’ve done their job.
Nobody ever stops to ask: How many infections would be “normal” for the population in question? Does the stated count of infections represent excess risk, above and beyond what the average American faces? Or does it just match the existing “background” rate of infections in a comparable US population?
Take Amazon.com, for example.
Per the article cited above, Amazon.com had almost 20,000 infections among their line (non-executive) workers. But that’s out of a combined line workforce (Amazon and Whole Foods) of nearly 1.4 million people. And so, as of September 19, 2020, their cumulative COVID-19 infection rate works out to be 1.4 percent.
As of that same date, the cumulative infection count for the US as a whole was about 6.6 million persons, out of a population of 328 million, or just about exactly 2.0 percent.
That’s not an apples-to-apples comparison by any means. The Amazon.com worker population is younger and able-bodied, compared to the US worker population as a whole. There would be, for example, no nursing home residents represented in the Amazon.com workforce. But offsetting that, there would also be no children.
Very roughly speaking, based on the age mix of COVID-19 cases in Virginia, you’d have to eliminate all the US cases over roughly age 45, in order match the US average COVID-19 rate to the Amazon.com employee rate.
In other words, the COVID-19 infection rate among Amazon.com employees is just about equal to the US average for persons age 21 – 45. The only reasonable conclusion you can make from that is that working at Amazon.com isn’t particularly risky, from the standpoint of picking up a case of COVID-19. Or to put it a little stronger, the infection rate in the Amazon.com work force is pretty much average. It’s not materially different from the infection rate for the rest of the US in that age group.
But “Amazon COVID-19 infection rate matches US average” is not click-bait. Nobody can make a buck telling you that. So nobody does.
The Ancient and Honorable College of William and Mary.
The College of William and Mary (W&M) has been, up to now, a clear success story in terms of students returning to college campuses. It is one among many such successes, per Post #852.
W&M shared two things in common with other successfully-reopening universities. They tested everybody, and they cracked down on those who broke the rules. Separately, most student housing is on-campus, including their relatively sane fraternity and sorority houses.
They recently had a cluster of a dozen cases among their student athletes. That’s unfortunate, but not surprising. Athletes are classified as a “high contact” population, meaning, they can’t go about their business without coming into close contact with a lot of people. I was surprised that W&M was not continuously testing their athletes. I suspect that they are, from now on.
And so I got to wondering, now that there’s a few months of data to look at: Even with this last outbreak, how does the cumulative count of cases among W&M students compare to what we would expect, if they had merely gotten infected at the “community” rate, that is the average rate in Virginia for persons their age. How many COVID-19 cases would we have expected to see if everyone had just stayed home?
I had already calculated the “community” infection rate, in Virginia, in an earlier post. Things haven’t changed much since then, so let me use that. Here’s the rate of infections (per month!), recently observed in Virginia:
In my mind, college students are far more like the 20-29 age group than they are like the 10-19 age group. So I’m going to use the rate of infections, for community-resident individuals age 20-29, in Virginia, as my baseline. (Bearing in mind that about 2000 out of the 8000 W&M enrollment is graduate students.) That’s the rate of infections I would have expected, if W&M college students had just stayed home and caught COVID-19 like everybody else in that age group.
And so, when I compare the actual cumulative count of cases (blue) to an expected count, at present, the actual count is below the expected count. (The series starts on 9/28/2020 because that’s the first day the W&M count was high enough that they could report the exact count. The blue bars missing for some days because I only check that dashboard once-in-a-while).
I should be savvy enough to say that because the expected number is highly uncertain, really, there’s no statistically significant difference between the actual and expected infection counts. But click-bait-wise, that’s a snoozer of a headline. So in keeping with our current tradition of making a big deal out of absolutely nothing, I’ll say that it looks like it’s safer for students to be at William and Mary than to be at home.
And, really, there’s good reason to think that’s plausible. Sure, they are living in dorms. But mask use is rigidly enforced, they are tested on a routine basis, any infected individuals discovered by testing are immediately quarantined, and so on. None of which I could do, as a parent. Honestly, I have to wonder whether W&M really is taking better care of my kid than I could.
As a final check, let me work out a baseline rate based on the 20 students who were discovered to have had COVID-19 in the pre-move-in testing. That’s a sort of pre-post analysis of the same student body, rather than a cross-sectional comparison to the Virginia average population of the same age. If they came in with 20 active cases, and active cases have detectable virus for an average of 14 days, that means they were being infected at an actual rate of roughly 1.4 new cases per day, in the period just prior to moving to campus. Whereas my all-Virginia average, per day, for 8000 students, works out to an expected rate of 1.6 new cases per day. So, the actual W&M student body’s historical rate appears to be just slightly lower than the rate I estimated for the cross-section of all Virginians of that age. I think that’s entirely plausible, given the generally high socioeconomic status of the families of the average W&M student.
