Numbers Illiteracy

John Allen Paulos teaches math at Temple University. He writes books, too. Good books. Books for lay people about math. One of which he titled Innumeracy … which is why this post got titled Numbers Illiteracy, as opposed to … Innumeracy.

Stupid numbers drive me nutso. They abound now, in an era when stupid, vapid creatures rule.

I hear lots of people talking about how the United States has the largest number of COVID-19 cases in the world. These comments come, mostly, from those whose fury focuses first on our Fearless Leader.

Well, duh! More people live in the United States than in any other country, save China and India. Among large nations we have, by far, the highest capita GDP (which means our people have more money, which allows us to travel more.) In addition, we have a greater right to move around than do people from China.

So, it should surprise no one that we have lots more cases than other countries. But, what about China? A really big place, with four of them for every one of us.

Well, there’s the movement thing, and the ability to lockdown a region in ways we can’t imagine. More sensible comparisons? Italy and Spain, where people can move around, and have enough Euros to do so, at least a little. Their numbers? Death rates of 263 and 282 persons per million, while the U.S. death rate is 29 per million.

On the other hand, COVID-19 death and infection rates depend entirely on testing. If you test five people, I promise you this: you will not have more than five confirmed cases. News reports, too often, focus on the low number of cases, without giving the testing limits due consideration. Of course, honest reporting matters too, and that’s another issue.

(Everyone, I hope, should note the mixing of infection and death data, especially in a piece about misusing numbers. Forgive me, please. First, the U.S. will almost surely have more reported deaths than any other country. Second, data from country to country means very, very little, given testing and reporting challenges.)

Then there are COVID-19 deniers. They like to quote annual flu death stats and compare them with where we are now … err, were a couple of weeks ago. Timing matters, so if you want to tell me this coronavirus is not much more than a flu, be quiet until we’ve reached the one-year mark. Or, prorate. Prorate? Yes, take the number of flu deaths in an influenza season – 34,200 in the U.S. in 2018-19 – and divide by 52. Roughly 660 deaths per week. Assume a start date for COVID-19 of February 1, nine weeks ago. About 6000 people would have died by now if flu and COVID-19 matched up, and as I write the COVID-19 total exceeds 11,000. (Again, without adequate testing we can only assume COVID-19 has caused some additional deaths which we can’t blame, conclusively, on the virus.)

I also see and hear plenty of comments about the projected number of U.S. deaths, compared to U.S. deaths from wars since 1945. True these numbers are, but let’s recall the fact that, as of this date, we have lots of good reasons to believe everyone will die, someday. No one who takes air right now will surely die, but the odds seem high that everyone’s day will come. So, what about the 150,000 mostly preventable lung cancer deaths here every year. Or the 40,000 or so mostly preventable gun deaths. Etc. The point? Cherry-picking comparable numbers contributes nothing.

One more thing. Projections represents statistical analyses, based on available date and modeling. They provide a range. No one will likely hit the mark, exactly, and it will move around. I offer this observation for two reasons.

• We’ve been hearing about between one and two hundred thousand deaths. Then, today, the number dropped to less than 90,000. That doesn’t mean people were wrong; it means, simply, that the data points have changed.
• The range allows for many outcomes. By way of example only, pollsters didn’t “get it wrong” in 2016 when our cockamamy Electoral College gave us the Buck-Passer in Chief. A Trump victory was a less than likely outcome, bigly, but it was always a possibility. Sadly, the votes he needed showed in all of the right places.

Please, please, please, remember these words about stats when Mr. Trump takes credit for fewer deaths, or blames others if the number ends up being higher.

I appreciate the challenges we all face here. I’m a little bit freaked out about this thing, but way better than I was a couple of weeks ago. I also know the odds count heavily in my favor, (although my lungs don’t work as well as they might, what with lots of scar tissue from pneumonias, and my advanced age.) I get, as well, the fact that needing to engage from our Isolation Wards explains plenty of commentary. That said, let’s remember one thing about numbers: without context, they represent a series of digits, telling us nothing that means anything.

Be well and stay safe.

Mark