Sunday, September 9, 2007

This Weekends Contemplation - You’d Think They Could Leave Statisticians Out of the Politics

As a statistician I often found myself in the middle of political “wars” concerning various decisions or directions that senior management wanted to pursue. The reason I was often in the middle was not because I wanted to be, but because I could access “objective” numbers that would substantiate one direction or another. My favorite process was for senior management to make a decision and ask my team to provide the numbers to substantiate their position. This ran contrary to the way I was trained. Being a classically trained statistician, I was taught to allow the data to do the talking. As my advisor once told me, “the data leaves a trail, your job as a statistician is to learn how to follow it.”

By the way, the article below from the WSJ uses the war in Iraq as an example of how data is used. The purpose of this post is not to discuss the war in Iraq and I have no “public” opinion on the issue of the war in Iraq.

Most of us go about our daily lives without having to handle sums in the billions and trillions. So when advocates, politicians or business leaders want to get a rise from us over, say, wasteful spending, they have to figure out how to get us to understand what their studies have uncovered -- and to see it their way.

These days, one of the numbers we hear most often in the news is that the federal government has spent more than $400 billion on operations in Iraq. That figure, calculated by the Congressional Budget Office, is frequently cited by opponents of the war when they urge troop withdrawal.
The challenge is to make that number provoke shock and awe.

One Web site divides the cost among U.S. adults, showing visitors how each adult could have bought two pairs of Manolo Blahnik patent-leather pumps, or 1,858 longneck bottles of Bud or Bud Light beer. A blogger conjured up the notion that $400 billion would buy enough golf balls to fill all the tractor-trailers it would take to reach across a 13-lane highway, bumper-to-bumper, from Los Angeles to New York.

Readers of the Reason Magazine blog offered their own analogies, such as the number of Reason subscriptions (nearly 17 years' worth for every person on the planet) that could be purchased for $2 trillion -- an economist's projection of the past and future cost of the Iraq war -- after a writer for the magazine said his "head is too small" to comprehend figures that big.

He's not alone. Human beings really do have a hard time understanding big numbers. The figures used in newspapers and political debates can seem like confusing abstractions, cognitive psychologists say. "The brain probably didn't evolve to process very big numbers," says Michael McCloskey, a Johns Hopkins cognitive scientist.

Mathematician John Allen Paulos, author of several books about numbers and how they are used, says his students at Temple University in Philadelphia often can't grasp the scale of large figures. He has told them a million seconds is equivalent to 11.5 days, and then asked how many days contain a billion seconds. "I often get answers like twice as much or 10 times as much." (It's 11,500 days, or about 31.5 years.)

Michael Ranney, a cognitive scientist at the University of California, Berkeley, says journalism students in his number-training class who should know the U.S. population is 300 million sometimes guess California has one billion people (instead of 36.5 million).

For those looking to use policy figures to stun rather than edify, one popular device is to describe a stack of bills stretching into space. Ronald Reagan used that image -- a tower of $1,000 bills 67 miles high -- in 1981 to depict the national debt, which was headed toward $1 trillion. In a 1998 Brookings Institution report, Stephen Schwartz wrote that the total cost of nuclear weapons to that point equaled a stack of dollar bills stretching to the moon and nearly back.

But "people don't have much of a grasp of the distance from here to the moon," Prof. McCloskey points out. He suggests making numbers small, such as figuring costs per person.

Mr. Schwartz, editor of the Nonproliferation Review, admits to hesitating before using the stacking image. He decided to bring several bricks of dollar bills to a news conference about his report and recalls those went over well.

Robert Siegler, a Carnegie Mellon cognitive psychologist, calls the stack-in-space image "silly" and adds, "It doesn't provide any more intuition than people already had."

These days, Stephen Crockett, an online co-host at Democratic Talk Radio, keeps the actual heights of such stacks vague when he asks listeners to picture one million stacks of one million $1 bills, representing projected war costs of $1 trillion. That's because he recalls visiting the Bureau of Engraving and Printing when he was in high school and seeing stacks of bills -- not singles -- worth $1 million. "They were not that high," he confesses. (Using the bureau's formula, a stack contains about 233 new bills per inch.)

Other war critics note how much the budget for Iraq could buy in, say, port security or teachers' salaries. This line on war costs suggests that federal spending is a zero-sum game, meaning that money spent on the war could easily be substituted elsewhere in the budget.
That's not entirely the case, but then the purpose of such arguments is often more to advocate than to explain.

UCLA economist Lee Ohanian says the Iraq war is a "drop in the bucket," about 5%, of the federal budget, especially compared with World War II, which consumed at least 70% of federal expenditures from 1942-1945. The WWII figure, however, includes all defense spending for the period. "No matter what the number is, you need to put it in perspective," says Prof. Ohanian.
Numbers can have a powerful impact. Prof. Ranney says the real statistics behind controversial topics often surprise people -- and sometimes provoke them.

"By giving people numbers, you can change what they prefer," he says. "What can be more empowering than that?"

No comments:

Post a Comment