TUESDAY, FEBRUARY 23, 2021
Attempts at information: Sunday morning, on its front page, The New York Times reported the latest grim milestone.
High atop the paper's web site, a link was offered to the report. This capsule account was offered:
A Ripple Effect of Loss: U.S. Covid Deaths Approach 500,000
Twice as many Americans have died as early worst-case projections.
We'll consider those "early worst-case projections" below. First, though, here's the way the front-page news report started, once you clicked the link:
BOSMAN (2/21/21): A nation numbed by misery and loss is confronting a number that still has the power to shock: 500,000.
Roughly one year since the first known death by the coronavirus in the United States, an unfathomable toll is nearing—the loss of half a million people.
No other country has counted so many deaths in the pandemic. More Americans have perished from Covid-19 than on the battlefields of World War I, World War II and the Vietnam War combined.
"No other country has counted so many deaths in the pandemic?"
The statement is technically accurate. But one year into the pandemic, on its front page, the New York Times still doesn't adjust for population when it makes such statistical statements.
The United States has endured the most Covid deaths—but we're also the world's third largest nation by population. (Only China and India are larger.) After adjusting for population, some other peer nations have experienced a higher rate of Covid deaths to date.
We're using the current Financial Times numbers. We're omitting a few smaller nations whose death rates are higher than ours:
Total Covid deaths to date, per million population
United Kingdom: 1,808
Czech Republic: 1,805
United States: 1,487
We have the largest number of deaths only by dint of our large population. In our view, it's amazing, and yet not amazing, to see major news orgs and major pundits continue to work outside the boundaries of this amazingly basic type of statistical adjustment.
This is the way it works in Our Town. Having established that basic point, let's flesh out the picture a bit:
After splitting it off from the U.K., England's death rate to date is much higher than ours. Spain is a tick behind the U.S., France a bit farther back.
Here you see the way we stack up against the largest Euro nations, with Canada also thrown in:
Total Covid deaths to date, per million population
United States: 1,487
England has suffered the highest death rate to date. Germany and Canada have done much better than we have.
Among those major peer nations, we pretty land in the middle. The numbers for the Pacific nations are still amazingly low.
"No other country has counted so many deaths?" That's true in the narrowest literal sense. That said, it strikes us as amazing that this continues to be the way basic information gets ladled to us on the front page of our allegedly smartest major newspaper. (And on the Maddow Show.)
Now, how about that claim in the headline—the claim about those "early worst-case projections?" Here's the part of Bosman's report from which that headline was derived:
BOSMAN: One year ago, as the coronavirus took hold in the United States, few public-health experts predicted its death toll would climb to such a terrible height.
At a White House briefing on March 31, Dr. Anthony S. Fauci, the top infectious-disease expert in the country, and Dr. Deborah L. Birx, who was coordinating the coronavirus response at the time, announced a stunning projection: Even with strict stay-at-home orders, the virus might kill as many as 240,000 Americans.
“As sobering a number as that is, we should be prepared for it,” Dr. Fauci said at the time.
Less than a year later, the virus has killed more than twice that number.
In part, that passage is accurate. In other ways, that passage illustrates the confusion that almost always develops whenever our upper-end journalists try to deal with numbers.
On March 31, 2020, Fauci and Birx made a high-profile appearance with the commander himself. On that occasion, they made a new high-profile projection in which, using Bosman's language, they said the virus "might kill as many as 240,000 Americans."
That said, there was a bit of a problem with their presentation. As best we can tell, they never explicitly said what time frame this new projection covered.
Other public health experts were predicting certain numbers of Covid deaths by certain specific dates. As best we can tell, Fauci and Birx offered no such framework at the widely-discussed public event from which that new projection emerged.
That presentation by Fauci and Birx produced extensive press coverage. In its reporting of their projection, the Washington Post was bright enough to mention the lack of this standard framework.
(Rucker and Wan: "[N]o time frames or other details were provided...One key question, for example, is what time period the projection of 100,000 to 240,000 deaths covers. If it is only the few months until summer, as is the case in at least one academic model, the true death toll will probably be larger.")
The Washington Post explicitly noted the absence of a time frame. The Post explicitly noted the possibility that "the true death toll" could exceed the stated number in the fullness of time.
At NBC News, Denise Chow reported that the startling new projection did involve a specific time frame. According to Chow, the new projection by by Fauci and Birx only extended through "mid-June."
In standard fashion, Chow didn't explain the basis for this statement. For all we know, her statement may have been accurate.
Meanwhile, back at the ranch, you knew it had to happen! In its report on the new projection, the New York Times failed to note the lack of a time frame. (The Times did note that the IHME was predicting 84,000 deaths by the beginning of August.) In fairness, this same statistical insouciance was on display almost everywhere else.
Eleven months later, the New York Times is still failing to adjust for population when it makes nation-to-nation comparisons.
Meanwhile, is it accurate to say, as Bosman does, that "few public-health experts predicted its death toll would climb to such a terrible height" as the very large number of Covid deaths we have now experienced? Is it true that "twice as many Americans have died as early worst-case projections?"
In part, it all depends on what the meaning of "few" is! But uh-oh:
Two days before the March 31 event, Fauci himself had floated a very large number—or at least, that's what New York magazine said:
STEIB (3/29/20): Though President Trump continues to downplay the necessity of an all-out federal response to the coronavirus, one of his most senior public health advisers, National Institute of Allergy and Infectious Diseases director Dr. Anthony Fauci, issued a grim projection on Face the Nation on Sunday. Speaking with Jake Tapper, Fauci said he anticipated somewhere between 100,000 and 200,000 coronavirus deaths in the United States, as well as “millions” of cases:
Fauci suggested that the 100,000 to 200,000 death range is a moderate estimate, and that the possibility of 1 million or more Americans dying from the coronavirus is “almost certainly off the chart”—that “it’s not impossible, but very, very unlikely.”
As you know, Jake Tapper isn't the host of Face the Nation. In reality, Fauci made the quoted remarks on CNN's State of the Union on Sunday, March 29.
On that show, had Fauci really floated the outside possibility of as many as a million Covid deaths? On balance, it looks to us like he did. But confusion developed as to whether Fauci was discussing cases or deaths, and Tapper didn't make a point of clarifying what his guest had actually said.
So it goes, here in Our Town, when we try to conduct public discourse. This seems to be the best we can do here in the streets of Our Town.
In the end, you can be certain of two things:
Thanks to our large population, we have experienced the largest number of Covid deaths.
Of that one fact you can be sure. Also, statistics are hard!
Just a guess: Why didn't Fauci and Birx state a time frame for their projection?
Just a guess—the commander didn't allow it. That would be our first guess.