Sen. Snowe and the Mad Biologist's Rule of Base Ten Numbers

My colleagues have all heard of the Mad Biologist's Rule of Base Ten Numbers: when you see too many numbers that end with zero, become skeptical. That's because only one in ten numbers should do should end in zero. So, if you read news reports that routinely say, "Today, American forces blew the crap out of [number that ends with zero] enemy forces" (and with a globe-spanning garrison empire, we do read a lot of those, don't we?), nobody has a good idea what actually happened.

Likewise, if you're reading a grant proposal and each treatment will be done ten times, there probably hasn't been a power calculation to determine how many times the experiment should be done. Why not perform nine replicates? Why not eleven? (An aside: If I ever read in a grant proposal, "We are perform eleven replicates because this grant goes to eleven", I would vote for funding in a heartbeat).

Which brings me to Republican Senator Olympia Snowe.

President Obama in his recent speech said that the cost of his health insurance plan over the next decade would be $900 trillion. Then the good Senator Snowe chimed in:

Another Republican negotiator voiced concerns to Fox. Sen. Olympia Snowe, R-ME, said there is still concern about the size of the package which is carrying a near $900 billion price tag. "Maybe we could shrink that to $800 billion or below," the moderate senator said, citing a skeptical public with bailout fatigue and concern for rising deficits.

This isn't a serious proposal. Why not shrink it to $827 billion? Steve Benen notes (italics mine):

Economists saw a trillion-dollar hole in the U.S. economy. Centrists, with an odd fondness for round numbers, kept wanting to shrink the size of the recovery response, just because. They wanted a smaller number, just so they could say it was smaller. They eyed $100 billion in cuts, because $100 billion had a nice ring to it. They were thrilled to fall under an $800 billion ceiling, not for any policy goal, but because it sounded "reasonable."

...It is, in other words, entirely arbitrary. Obama is eyeing $900 billion for health care reform. Snowe is now thinking about "$800 billion or below." Why? Because it just sounds better. Less is necessarily superior to more, the argument goes, for vague, personal reasons that have nothing to do with addressing the problem at hand.

I realize we're talking about a lot of money here, but the difference between a $900 billion reform package and an $800 billion package is $10 billion a year. Given the size of the U.S. economy, the federal government's budget, and the willingness of lawmakers to spend freely when it was debt-financed Bush-era initiatives on the line, an additional $10 billion a year to help Americans have quality, affordable health coverage is more than reasonable.

Of course, Obama didn't pick $900 billion out of a hat either....

More like this

In contrast, my training as a physicist and engineer tells me that when someone talks about a $913,287,563,192.14 price tag they're making it up.

Precision in excess of reasonable accuracy is a sure sign of bogosity.

By D. C. Sessions (not verified) on 15 Sep 2009 #permalink

The more things change, the more they stay the same.

... centrism is a pose rather than a philosophy. And to support that pose, the centrists are demanding $100 billion in cuts in the economic stimulus plan â not because they have any coherent argument saying that the plan is $100 billion too big, not because they can identify $100 billion of stuff that should not be done, but in order to be able to say that they forced Obama to move to the center.

Another way to catch bogus numbers is that most numbers start with 1.

Oh yes. Real-world data is usually arranged logarithmically.

So if you examine a data set of - say, money, or mass, or duration, or anything other kind of data where negative numbers make no sense, and the first digits have an equal chance of being 1-9, then the data is almost certainly faked.