Why you should be suspicious of "people on the street were more ignorant of X than Y" stories

[mmcirvin]

I missed this in March: Mark Liberman of Language Log takes apart the "more people could name the Simpsons than First Amendment freedoms" story.

The people who could correctly name N First Amendment freedoms (I assume they didn't split freedom of religion into the establishment and free-exercise clauses) decreased monotonically with increasing N, whereas the Simpsons distribution was bimodal, with more people ignorant of even one Simpson than of one First Amendment freedom, but also a fairly large population who could name all of them. From these numbers it is possible to extract just about any message you want, especially if people repeating the story are inclined to garble the result in the desired direction.

Comments

[bottlroktt.livejournal.com]

uh oh, jay leno's in trouble!