April 2, 2020

1154 words 6 mins read

Why Should I Believe You?

Why Should I Believe You?

Someone sent me a video. The video shows a doctor explaining how coronavirus is worse than the flu. Should we believe him?

I say, no. And the reasons we should not believe this doctor about the relative dangers of two diseases have nothing to do with medicine. They have everything to do with statistics and lying.

First, the statistics.

The Doctor’s Statistical Error

The doctor, a general surgeon, opens the video by stating his conclusion that coronavirus is deadlier than influenza and his proof. His proof is simple: the death rate for influenza is 0.1 percent, while the death rate for coronavirus in Italy is 10 percent. That seems to make coronavirus 100 times deadlier than the flu, which was Dr. YouTube’s conclusion.

Case closed, right? I mean, these are numbers we can all look up. And they come from a doctor.

And there is a modicum of truth in what the doctor tells us. There are published statistics that show only 0.1 percent of influenza cases end in death while (as of today) 11 percent of Italians with coronavirus have died.

Again, seems like an open-and-shut case, right?

But what the doctor isn’t telling you is that he is comparing two different kinds of data and treating them as if they were identical.

In the case of influenza, he’s using the number of presumed cases. In the case of coronavirus, he’s using the number of confirmed cases. Since over 90 percent of people who come down with the flu don’t get tested for the flu unless it’s serious enough to require hospitalization, the number of presumed flu cases is very large. How large?

The CDC estimates that, between October 1, 2019, and March 21, 2020:

  • Between 38 million and 54 million people suffered the flu
  • 18 million to 26 million people visited a doctor because of flu symptoms
  • 400,000 to 700,000 people were hospitalized with the flu
  • 24,000 to 62,000 people died from the flu

And as of March 31, 246,000 Americans had tested positive for influenza.

In other words, even though the confirmed cases of the flu are 246,000 (about of about 1.3 million tests administered), we know from historical analysis that number correlates to 38 million to 54 million actual cases of the flu.

No such historical data exist for COVID-19, so there are no estimates of the number of actual cases. We know only the number of confirmed cases.

When we take the number of deaths from the flu this year and divide it by the number of presumed cases in the United States (24,000 / ~45,000,000), we find that flu less than 0.1 percent of the people who presumably contract the disease.

When we take the number of flu-related deaths, however, and divide it by the number of confirmed cases of influenza, we get a 10 percent death rate—about the same death rate as for coronavirus in Italy.

The doctor’s conclusion, that coronavirus is deadlier than influenza, is not supported by his evidence, because he’s basing his evidence on two different kinds of data that cannot be responsibly compared to each other. This isn’t a medical error, it’s a data analysis error, and it’s an egregious one. He’d fail a basic course in data analysis for drawing a conclusion from a comparison of disparate data. In short, Dr. YouTube is using the same numerator but dissimilar denominators.

So was the doctor just wrong or was he trying to deceive?

The Doctor’s Deceptive Tell

I like learning about human behavior from people whose lives depend on understanding human behavior. And I’m not talking about psychologists. Their lives depend on government grants to the universities they work for. No, I’m talking about people who literally make life and death decisions in a split second based on a keen understanding of human behavior: cops.

Police officers, field FBI agents, CIA Clandestine Operations field officers, and hostage negotiators know how people behave. And how they misbehave. And the tells they display when they’re being deceptive.

One of the things I’ve learned from these streetwise behavioralists is how to spot a lie. And one of the best ways to spot a lie is how the person answers the question “why should I believe you?”

People telling the truth reflexively respond by saying so: “because I’m telling the truth,” or words to that effect. Liars won’t say that. They’ll give their bonafides or credentials. “Because I’m a scientist.” “Because I’m Hillary Clinton.” Or, maybe, just a simple, “because I’m an honest person.” All appeals to who they are, not what they’re saying. (Psychologists might tell you this is because they are subconsciously deflecting attention away from the false statement.)

What did this YouTube doctor say to the unasked question “why should we believe you?” He held up his diplomas and answered, “because I’m a doctor.”

Doctors study statistics, but they’re not data analysts. The YouTube doctor presented data analysis work and wants us to accept his conclusions unquestioningly because he’s a doctor.

Again, the YouTube doctor didn’t give a medical reason why coronavirus is, in his view, deadlier than the flu. He gave a statistical reason.

Because doctors do, indeed, study statistics and because this doctor displayed a lie tell in his video, I have to assume that he knows his evidence does not support his conclusion. He deflected our attention away from his false statistical statement to get you, the viewer, to focus on his medical credentials.

Dr. YouTube’s Conclusion Might Be Right

Look, it’s possible that the doctor’s conclusion is correct. We don’t know. We won’t know how deadly coronavirus is on a per-case basis until long after the pandemic passes, until we have tested enough people for the infection and its lingering antibodies to determine just how many people were infected. Coronavirus is far more contagious than the flu because no one had an immunity to it. It’s possible that this novel coronavirus is much deadlier than influenza.

It’s also possible that it’s not.

When someone tells you that COVID-19 is 100 times more deadly than the flu, he’s lying. For now, he’s lying. Available evidence responsibly presented is that 1 out of 10 people who have come down with the flu in the United States this flu season have died (24,000 out of 246,000 confirmed cases) and 2 out of every 100 people who have tested positive for COVID-19 in the United States have died (3,976 out of 187,347 as of 8:21 p.m., March 31, 2020.)

Dr. YouTube, instead, gave a false comparison of deaths per presumed cases on the one hand versus deaths per confirmed cases on the other when data for confirmed cases, the common denominator, was readily available for both.

When a guy on the street makes this mistake we can chalk it up to ignorance. When a doctor does this, we have to conclude it’s a case of deception.

Watch the video here and judge for yourself.