Bring It On!

Two and Two Ain’t Always Four

December 3rd, 2008 | by Omnipotent Poobah |

Bernacke

If Dick makes $57 per hour and Sally makes $70, who has the better PR agent?

Well, both silly.

Dick and Sally both make the same amount, regardless of what the numbers apparently say. That’s because both numbers are more “truthy” than absolute. Opposing PR firms - one for the UAW and the other representing anti-union activists - spun their press releases using the same pool of crunched numbers to produce entirely different results. Then, they delivered the numbers to desperate audiences prone to believe that one or the other is God’s-own-truth. As an added bonus, neither side is literally lying - even if they are being misleading.

We live in a data-driven society in which numbers are the Holy Grail. We use them to measure progress in schools, how competitive America workers are, and whether The Surge is working. In the process, we cook the raw data so it most closely resembles the ideological point we’re trying to make. We buy into the fiction that numbers actually tell you something. The problem is that numbers can tell you everything, anything, and nothing at the same time. And, it’s something that makes actual policy-making a damnably difficult thing to do.

On a recent Thom Hartmann radio show, Thom and his guest accidentally demonstrated the point. In one breath, they pummeled “the media” for unquestioningly accepting the “untrue” $70 figure while not giving any coverage to the $57 figure. Forgetting for the moment that Thom IS “the media” and that stories - including Thom’s - DID appear, will it make any difference in the final outcome?

Not in the slightest.

Numbers seem real to humans. They suggest orderliness while ignoring the chaos and capriciousness of humans. If numbers don’t jibe with our notions, we align them in less chaotic and more palatable ways. But when we do, we only present the illusion of orderliness. The $70 faction and the $57 faction will both continue to sit upon their number as if theirs the only true one. Meanwhile, the two sides will bicker over bail out details for The Big Three based on their truthy figures rather than the meat of the bail out.

If the amounts of money were only a few hundred bucks, the decision-making wouldn’t be so difficult. But when the numbers are big - astronomically big - our estimate of the amount actually needed falls somewhere between $0.00 and $3 billion plus. No one knows the true figure and won’t until after the exercise is over. At that point, the “winner” will blame the “loser” using updated - but still truthy - numbers.

There are three primary rules about collecting metrics. One, you must be able to collect data consistently. Two, data must be accurate. Three, data must mean something. This example fails all three tests.

However, Thom and his guest did have one thing right. Never accept any crunched number as absolute. There’s a 99.9% chance (how’s that for a metric) some enterprising PR agent has spun it like a top.

If you don’t, your rhetoric is as useful as a chat with Joe the Plumber - damn the numbers and full speed ahead!

Cross posted at The Omnipotent Poobah Speaks!

Share and Enjoy:
  • Digg
  • Sphinn
  • del.icio.us
  • Facebook
  • Mixx
  • Google
  • e-mail
  • YahooMyWeb
Sphere: Related Content

  1. 3 Responses to “Two and Two Ain’t Always Four”

  2. By Matthew O'Keefe on Dec 4, 2008 | Reply

    Amen Poobah! Numbers don’t lie, people reading them do though!

  3. By Paul Watson on Dec 4, 2008 | Reply

    I highly recommend Darrell Huff’s book, How to Lie with Statistics. It goes through the most common ways statistics are misunderstood or presented to give an impression different to the actual facts. Although after you’ve read it, you’ll never take a statistic at face value again. But skepticism is a good thing in my opinion.

  4. By Chris Radulich on Dec 5, 2008 | Reply

    If the unions were making all the decisions in the car companies, why was management paid so much money?

Post a Comment

Fish.Travel