David Kravets of Wired magazine has an interesting post on a number that has made its way into a US Chamber of Commerce letter to the president, encouraging him to sign into law a bill focused on stronger enforcement of intellectual property law, including the creation of a cabinet-level "copyright czar" to oversee these efforts. The number in question declares that 750,000 American jobs have been lost due to counterfeiting and piracy of intellectual property, a number which, as Kravets points out, is pretty impressive, and would make up 8% of the 9.4 million Americans who are currently unemployed.
The only problem is, no one seems to know where the number came from. The US Chamber of Commerce says they got the number from the US Department of Commerce (yes, that's two different agencies), but the US Department of Commerce says they got the number from the US Chamber of Commerce. Not only that, but the number is also quoted by the US Customs and Border Protection, which says the number comes from the International Anti-Counterfeiting Coalition, but the IACC says the number comes from, you guessed it, the CBP.
This kind of circular quoting of statistics is not nearly as uncommon as you might think. Daniel Gardner, in his book "The Science of Fear," tells the story of "50,000 pedophiles prowling the internet at any given time," a number widely quoted by a number of news reports and government officials, including former US Attorney General Alberto Gonzales. This number turned out to be essentially a number that an FBI agent once heard and didn't know where it came from, but thought was a "fairly reasonable figure." (Gardner offers this as an example of the use of what psychologists call the anchoring and adjustment heuristic.)
This is not to downplay the value of experts estimating the size of a particular phenomenon. Experts often have insight into their fields, or expertise in certain modeling or statistical analysis techniques, that others don't have; that's what makes them experts. But when people (or government agencies) are presenting information in an attempt to persuade us of the correctness of their position (or in an attempt to convince the president to sign a bill into law) it's appropriate to ask how exactly these estimates were arrived at, not to mention who exactly arrived at the estimate, as that knowledge may affect how much weight we assign to it. If an estimate turns out to be just a "fairly reasonable figure" that someone essentially pulled out of the air, that's something that's good to know.