Sunday, May 21, 2006

Lies, Damn Lies and Statistics

Gerald W. Bracey dug out the source behind three numbers in today's WP: 600,000, 350,000 and 70,000. These are, allegedly, the number of engineers produced in 2004 in China, India and the United States, respectively.

His conclusion:

"Statistics that end up as conventional wisdom even when they're wrong usually become popular by being presented as fact in a highly visible and respected source -- such as a cover story in Fortune or a National Academies report."

Read more here.

But who is to blame?

The news reporters who do not bother to check the accuracy of their numbers. Or the consumers of those numbers, like you and me?

If the responsibility is on us, the consumers, to verfiy the authenticity of the numbers, how much time can we spare each day for doing that? Time constraints imply that we are extremely easy to be misled by such statistics. Our decisions which are based upon at least partially on those statistics will be similarly affected as well.

Will competition among different papers ensure that reporters do their checking but citing statistics? One can only hope so, but I doubt it.

Also puzzling is why reporters are not as diligent in checking their numbers as one would expect.

To build up credibility and reputation (or simply to protect these attributes in the case of renown reporters), one would expect the reporters have the right incentives to do their checking. Is it because a number that is intended to stir up emotions of the readers matters much more in boosting sales than an accurate one? Afterall, the readers have no time to figure out the number's authenticity anyway.

No comments: