Tuesday, October 18, 2011

Data'll do, Pig. Data'll do.

Data will do just about anything you ask of them if you squeeze them hard enough. The trick is to know how much pressure to apply. Squeeze just right and you've got data-supported facts. Squeeze too hard and you've got data-extorted crap.

Good data are vital to good communications. But you must treat them with great care. If you force your data to perform stunts that they just aren't capable of doing, you're going to end up with muffin on your face. Some examples:

Muffin but the truth: The Justice Department's acting inspector general accused the agency of spending too much on conferences, citing charges for a $16 muffin. That fallacy flew around the world faster than a CERN neutrino. Turns out the IG's office misread a sloppy hotel invoice.

Honey, I lost the kids: Google "800,000 children disappear every year" and you'll get 2,600 hits, many on national media websites. So 2,000 kids a day -- poof -- just disappear and we're all OK with that? Now google, "800,000 children disappear for at least a time every year," which is a more realistic interpretation of the data, and you get a far more realistic, far less sensational, and far less reported (48 hits) story. As Slate points out, very few children suffer the fate that those slinging the startling 800,000 statistic want you to believe.

We're praying for you,
Dennis, and all the blondes
out there.
Blondes have more fun, fewer days: "A study by experts in Germany suggests people with blonde hair ... will become extinct by 2202," reported the BBC, among other leading media. But as Snopes proves, the story is not only false, it has also been around in one form or another since 1865.

Which brings us to the latest data abuse.

99% confused: As slogans go, "We are the 99%" is fan-freakin-tastic. Right up there with "Hey Jerry, what's the story?" It's powerful. It's pithy. It's got a great beat and it's easy to march to. It's also unsupportable. Worse, it may very well force the Occupy Wall Streeters to redistribute their wealth with the entire global family or face the inevitable cries of hypocrisy. 

Here's why: On the global scale, we Americans all were born on third base in the metaphoric game of life. Based on the Occupiers' chanted objective, when we 99-percenters finally do wrest the wealth from those in the sky box, we're going to have to share the wealth with those on first base, those on deck, and the hordes in the stands who have absolutely no hope of ever even playing in the game.

As NYTimes Economix columnist Catherine Rampell explains, the dirt-poorest among us are better off than almost 70% of the rest of the world. And a whole heck of a lot of us are better off than 90% of the rest of the world.

So if we're really serious about global economic justice, better get your ATM cards out.

Last point: If this response to Ms. Rampell's column is typical, you 99-percenters might want to get ready to do some damage control: "Concepts such as poverty only make sense relative to the particular society in which one lives; not having a car or cell phone in rural India is not the same as not having a car or cell phone in the U.S."

For a great book on number abuse and general innumeracy, check out Innumeracy by John Allen Paulos. It's brilliant, fun and a little scary. A must-read for you PR types.


  1. From the comments on the Times site:

    Take the assertion that "America’s poorest are, as a group, about as rich as India’s richest." Well, India has nearly 4 times the population of the US, so its top ventile has nearly 60 million people (more than all but the top 20 or so most populous countries in the world), compared to a mere 15 million in a US ventile. Brazil, on the other hand has about 2/3rds the population of the US, so the top Brazilian ventile has just 10 million people in it. If you defined the top 60 million Brazilians as "rich" and took their average income, you'd be a lot less impressed than after reading this blog entry.

    If you accept Rampell's interpretation, then you might as well compare the top ventile in Gabon (population 1.5 million, ventiles of 75,000, GDP per capita $10,000, PPP GDP per capita $14,400) and conclude that "the rich in Gabon are wealthier than the rich in Brazil." To be consistent, you'd also have to conclude that "the rich in Gabon are wealthier than the rich in the US."

    I guarantee you if you looked at the income of the top half or one million Indians, and compared it to the same for Brazil, it'd be quite similar. Obviously, the average Brazilian (PPP GDP per capita $10,000) is "richer" than the average Indian (PPP GDP per cap $2,800) but that's quite different from suggesting that the "richest" Brazilians are almost as wealthy as the "richest" Americans while the "richest" Indians are poorer than the "poorest" Americans.

    How do people look at a simple chart and come to "funny" conclusions - perhaps that old trope about damn lies and statistics? Maybe it's innumeracy or just hewing to some popular narrative, but either way the article is very misleading.