Saturday, September 25, 2004

Imaginary Jobs Reports: Following Up

Haven't updated this blog in a while due to heightened work responsibilities and travel, but I wanted to revisit a prior subject because of some discussions that resulted mainly taking place in other blogs.

Quite a few people criticized my earlier postings discussing how most of the jobs thought to be created in the last year and reported in the Bureau of Labor Statistics' establishment survey were imaginary. In a representative criticism that actually is somewhat valid as far as it went, Jason Williscroft said:

"No place do you demonstrate that the BLS numbers actually are wrong, nor do you show that the fundamental assumptions underlying their counting methodology are actually incorrect. The best you seem to be able to do is to cherry-pick data that seem to support your assertion, while dismissing any that don't. These guys perform such counts for a living; I would suggest that, if you are going to attack their results, the onus is on you to demonstrate how they actually are wrong, not how they might be wrong in some alternate universe." (quote found here.)

This of course is a double negative. Proving something is NOT so is virtually impossible. If I asserted that, in a good friend's frequent example, that Dover, Delaware is REALLY intended to be the capitol of the United States, others would have a hard time proving such, because of the strange scarcity of documents directly asserting that Dover is NOT the US Capitol. All one can do in such a challenge is show indirect evidence that the capitol was really some other city. Such evidence wouldn't make a skilled internet flamethrower pause for even a second, and huge quantities of bandwidth would continue to be consumed debating the proposition.

In this case the BLS is making the assertion, not I. They are claiming that over a million jobs have been created that have not EVER been detected by their actual sample surveys month after month. They make the claim based on

No comments: