Statistical crimes are off the charts (no pun intended) in 2020. Why? It’s easy to accept a chart or graph on the internet at face value without diving into it any deeper. A little book I love is a short read written by Darrell Huff in 1954 titled “How to Lie With Statistics”. After reading it, you may never view a chart, graph, or headline involving statistics the same way again.
Most of us are already aware that statistics are commonly framed in a way to fit a narrative. That doesn’t mean we are always less prone to falling for the bait. I suspect this will be the first of many posts I make regarding statistical lies, so we’ll start with the most pressing issue at the present time, COVID-19.
I know what you’re thinking “You are a financial planner, not a virologist, why should we read anything you say about COVID?”. I admit that I know nothing about virology, epidemiology, or medicine. My interest is in how data is framed and presented. Finance is full of misleading statistics, but we’ll get into that in another post.
Let us dive into some COVID-19 statistics. What if someone wanted to frame the argument that reopening states too early was a bad idea, and keeping restrictions in place was the best thing to do? The map below would fit that argument quite well.
This map was shared tens of thousands of times on social media. Many of those posts had a negative tone towards southern states that have reopened more aggressively.
Is the map correct? When we isolate two states like Illinois and Texas, it would appear so. The info below is from worldometers.info.
The cases in Texas are definitely rising, while the cases in Illinois are dropping.
Let’s look a little deeper into this. Pay attention to the Y-Axis in the charts above. First notice that the scale for each y-axis is different. Illinois is hovering around maybe 500-800 new cases per day, while Texas is around 2,000. What is not mentioned in the charts is that Texas has 29 million residents while Illinois only has 12.67 million. If Texas is seeing 2,000 new cases per day, it would be equivalent to Illinois seeing about 880 on a population adjusted basis.
Illinois is experiencing one of the largest decreases in daily new cases, mostly because we already had one of the worst outbreaks. You’ll also notice New York is a darker green state above, and they had the worst outbreak of any state.
Imagine another example: City A has 100 murders per capita one year, and the next year they have 80, a nice 20% decrease in the murder rate. The neighboring town and arch rival, City B, has 10 murders per capita one year and 15 the next, a 50% increase in the murder rate.
What do you think City A’s newspaper says? “CITY B EXPERIENCES A 50% INCREASE IN MURDERS”. Of course if you’re considering moving to either city, you would feel safer in City B……unless there is more to the story for the underlying trends.
What if I wanted to show that Illinois is doing a terrible job at handling the pandemic and Texas is doing a fantastic job? I would share this info below from June 15th. (also from worldometers.info):
|Cases Per 1 million People||Deaths Per 1 million People|
Yikes, Illinois has over 7 times as many deaths per capita from COVID and more than 3 times as many cases.
But this isn’t the whole picture, the trend matters. As we saw above, Illinois is clearly on the downtrend and Texas is going the wrong direction.
Why is Texas seeing such a large uptick in cases while Illinois is not? There are a number of reasons but one is that Texas has gone from testing about 10,000 people per day in April to over 30,000 people per day in June. The more you test, the more cases you will find. Why is testing going up? One reason is probably because more people are feeling sick, which isn’t good. Another reason for more testing is finding asymptomatic carriers, or testing everyone that a positive case came in contact with.
The percentage of people tested in Texas that have tested positive is lower today than it was in April (about 7% today versus 10% in April), they are just testing many more people.
If you are ramping up testing, you will likely ramp up the number of positive cases since it appears there are many asymptomatic cases.
Illinois was testing a higher percentage of their population earlier, which led to more cases being discovered sooner. Illinois, like Texas, also had about a 7% positive test rate until the beginning of June, and since then it has declined to about 4%. Here are the charts for testing between the states (pulled from John Hopkins University website):
The question really is, will the spike in case count in states like Texas lead to a spike in deaths or hospitalizations? So far it has not led to higher death counts in Texas (charts below and once again pay attention to the Y-Axis), but that could change quickly of course as deaths do lag cases.
Hospitalizations in Texas are now the highest they’ve been since the pandemic started, but on a population adjusted basis Texas will have to experience a widespread extended outbreak to catch up with the numbers in Illinois. As we’ve seen, that could certainly happen.
What Do We Make Of All of This Data?
We won’t know until well into the future which states handled this the right way, and which did not. There will be plenty of Monday morning quarterbacks, and I suspect that those who end up looking bad from their decision making during this crisis will still find ways to frame statistics in their favor.
It’s good to know how things are presented to drive one to a certain belief. There is almost always more to the story than what is being told by the journos. If this type of data interests you, Charlie Bilello of Compound Advisors wrote a good piece on their blog which you can find here.