This week marked a sobering, significant number: 100,000, the official count of COVID-19 deaths in the U.S. That number is higher than the population of Flint, Michigan; Lawrence, Kansas; or Albany, New York. It is one of many numbers that now define our world: the number of cases, the number of deaths, the number of tests.
Making sense of the constant drum of data can be a struggle.
I spoke with Temple University mathematics professor John Allen Paulos, author of several books, including “Innumeracy” and “A Mathematician Reads the Newspaper,” to get his take on the numbers surrounding the COVID-19 pandemic.
In summary, he said: “We don’t know enough.”
But we act as if we do. We report numbers precisely, down to the digit, as if they were extremely accurate when we know that most of the numbers are undercounts. He wrote in the Feb. 18, New York Times, that such precision can be dangerous.
Paulos offered the hypothetical example of the guard at the Museum of Natural History telling visitors that the skeleton in the lobby is 70 million and 8 years old. He knows because he was told, eight years earlier, it was 70 million.
Such precision isn’t necessary because most people understand estimates, Paulos said. We’re used to seeing polls that have margins of error.
Raw numbers
President Trump continues to say that the U.S. is doing more testing than any other country. But “the relevant number is the percentage, not the raw number,” Paulos said.
But even when computing percentages we face a “nebulous fraction.” Because we don’t always have accurate numerators (the number on top of the fraction) or denominators (the number on the bottom).
Let’s take the percent of positive tests. Is it the percent of those tested? The percent of everyone? We know that testing levels vary widely, depending on where you live. If everyone hasn’t been tested, the figure could be confusing.
Kaiser Health News dissected Trump’s testing claim several ways: “The United States has a far bigger population than many of the ‘major countries’ Trump often mentions. So it could have run far more tests but still have a much larger burden ahead than do nations like Germany, France or Canada.”
Also, the number of tests is not the same as the number of people who have been tested. We know that White House staffers are tested regularly, understandably, but that means that there are more tests conducted than people tested.
Death tolls
Even numbers that often are considered more reliable, such as the number of deaths, are really just estimates.
Paulos suggested that perhaps a better measure is the number of deaths compared to previous years.
The New York Times recently reported that “Nationwide, nearly 64,000 more people have died than usual between March 15 and April 25. That number is more than 16,000 higher than the official count of coronavirus deaths for that period. “
Even the moment we surpassed the 100,000 mark varied from source to source. In the end, the number is surely higher because many deaths were not attributed to COVID-19.
Yet the numbers serve an important role, Paulos said: “The incessant focus on the numbers is helpful. It impresses upon people that this is not over. People are dying in huge numbers.”
As the crisis continues, we become numb to the numbers. “On 9/11, 3,000 people died. Now, that happens every day and people shrug and say we want to go to work or to restaurants,” he said.
Paulos said he hasn’t seen enough reporting on what other countries did – especially those that have managed to tamp down or keep their numbers low.
“We had the longest warning, and we are the richest country, and yet our number of deaths swamps all of Europe together,” he said.
In the end, the professor who has worked to make us all more numerate, thinks we may learn something in this crisis.
“The relentless focus on numbers,” he said, “can help the public realize that math is somehow relevant to their lives.”
