TUESDAY, SEPTEMBER 22, 2020
As always, statistics were hard: Are we the people up to the task of being self-governing?
Top experts say we are not. As a species, we simply aren't wired for such atask, these despondent top scholars have told us.
Is our upper-end press corps equal to its basic tasks? As always, statistics are hard. Consider this passage from a front-page report about the virus in yesterday's New York Times:
ROMERO ET AL (9/21/20): In the United States, the daily death toll from the virus is down from where it was in early August, when more than 1,200 deaths were occurring every day. Yet even as some of the country’s most populous states report vast improvements, and as Northeastern states have kept new infections low, deaths continue to trend upward in 12 states and two territories.
Really? In early August, "more than 1,200 deaths were occurring every day?"
Below, you see the daily numbers from the relevant data set maintained online, for public consumption. by that very same New York Times (click here, scroll down). Can you see any sign that "more than 1,200 deaths were occurring every day" during early August?
New reported deaths by day in the United States:
August 1: 1056
August 2: 420
August 3: 608
August 4: 1356
August 5: 1252
August 6: 1075
August 7: 1356
August 8: 966
August 9: 539
August 10: 537
First, some basic background:
As we've explained, the Times doesn't report the number of deaths which occur on a given day. According to the Times data set, those are the number of deaths which were officially recorded ("reported") on those particular days.
As we've noted, the recording of deaths tends to drop off as part of the typical weekend. People keep dying of the virus, but the official recording of many deaths is delayed.
(For that reason, averages should be, and generally are, computed over 7- or 14-day periods. Larger distortions tend to occur over three-day holiday weekends.)
At any rate, those are the numbers of deaths reported on the first ten days in August according to the New York Times. Do you see any reason to believe that "more than 1,200 deaths were occurring every day" during this period?
For the first seven-day period in August (August 1-7), recorded deaths averaged 1017.6 per day. That's well under 1200. For the seven days from August 3 through August 9, the average number of deaths per day rose to 1021.7.
The rolling 7-day average began to drop at that point. So where exactly was the period where "more than 1,200 deaths were occurring every day?"
(Warning: Don't be fooled by the statistical blip which resulted from a change in reporting procedures in Texas on July 27. This change produced a statistical blip due to the way the change in procedure was processed by the Times. Through this link, the paper tried to explain.)
Above, we've quoted a front-page report in the New York Times about a deadly public health pandemic. Even in a matter like that, our biggest news orgs, as if by some congenital instinct, will almost always misreport the most elementary data.
According to official reports, there was no period in early August during which "more than 1,200 deaths were occurring every day." Correctly computed, the rolling seven-day average climbed over 1100, though only slightly, for a very brief period at the start of August. But there was no period in early August in which "more than 1,200 deaths were occurring every day."
Absolutely nothing turns on this misstatement. That said, it it helps us understand a basic fact about our upper-end press corps. Simply put, the journalists at our most prestigious newspapers will almost always misstate whatever data they're asked to describe.
We aren't speaking here about a high school newspaper. We aren't speaking about an arcane statistic in paragraph three million of some tedious news report buried on page 300 on some distant past date.
We're talking about a statistical claim in a front-page report about the largest public health disaster of the past hundred years. Even there, the journalists misstated the data, as they routinely will.
Projections suggest that the ongoing death rate will rise to more than 2000 deaths per day over the next three or four months. These projections aren't being widely explained or reported.
On cable, we see journalists saying that total deaths are projected to rise to 300,000 or to 400,000 by some not-too-distant date. We see them reporting such projections without seeming to realize what those numbers mean about the very large increase in daily deaths which is thereby being predicted.
We can only guess that they spend the bulk of their time in wardrobe, makeup and hair. Despite everything you'll tend to assume, they just aren't especially sharp, not even the Harvard-Yale-Fordham types.