Thursday, June 15, 2006

Meaningless Statistics

A couple of days ago Brad DeLong and Susan Rasky posted 12 tips for journalists reporting on economics and 12 tips for economists serving as journalists sources. The tips are very useful. The tips for journalists are actually good rules for undergraduates to follow when writing papers. So I especially encourage my undergraduate readers to check them out.

Anyhow, rule number three is:

3.No naked numbers. Don’t report numbers by themselves. Numbers have meaning only in context. And context is almost always impossible without explicit comparisons to other numbers. How does this number compare to other cities, other states, other countries, other eras? How does this number compare to total spending, spending on necessities, spending on luxuries, spending on other kinds of goods?


I point this out because Tyler Cowen is pimping reporting of a McKinsey study in the FT stating that Sweden's "real" unemployment rate is 15% (the additional 10 percent come from including those who want to or could work). Here's the key fear-mongering by the FT reporter, "The numbers cast a pall over Sweden's international reputation as a thriving welfare state with low unemployment..." Of course no where does the reporter discuss what the "real" unemployment rates in other countries would be if similar calculations were performed. Colman posts a diary at DailyKos arguing that a similar calculation for the US would produce a real unemployment rate of 13.3%. I haven't looked at the data myself, so I have no idea if his number is correct. Colman's instinct, however, is correct. Without a comparison, Sweden's "real" unemployment rate of 15% is completely meaningless.


Comments: Post a Comment

Subscribe to Post Comments [Atom]





<< Home

This page is powered by Blogger. Isn't yours?

Subscribe to Posts [Atom]