Wednesday, September 29, 2010

On average www.nytimes.com headline guy does very well but he or she struggled yesterday

Boy did he or she ever. The headline has changed twice. I don't know the order, but the current headline is (no screen shot) "Basic Religion Test Stumps Many Americans" which seems fine. But there was a third headline

"On Basic Religion Test, Many Doth Not Pass"

Which appears to be an incorrect use of "doth" (I doth not have a clue, but Belle Waring asked “Did they learn about religion from a bunch of old Thor comics?”

No answer to that question, however, John Holbo shows how headline writing is to be done

The Gray Lady Doth Protest Too Much, Methinks

Forsooth, methinks that John Hoblo art a genius.

Also his update art not bad

"UPDATE: Yon Gray Lady Hath Fixethed It."

Old post on a headline which doth by no means lead to such brilliant criticism seguent.

The headline at www.nytimes.com for a much noted story is odd

"Atheists Outdo Some Believers in Survey on Religion"

The story reports that the average score of atheists and agnostics on a quiz about religion was higher than the average score of each of many religious groups.

I suspect that an earlier candidate headline (perhaps just in headline guy's head) was "Atheists Outdo All Believers in Survey on Religion"

That hypothetical headline would be incorrect. That means either that no single believer scored better than the lowest scoring atheist or, at least, that no single believer beat the atheist average.

But the current headline is a very weak claim. "Some believers" might be 2 believers. I think that a better effort would be "Atheists Outdo Believers in Survey on Religion." There is no simple English qualifier which indicates that one is speaking about averages, since the English language was old when the concept of an average was invented.

I think the correct headline would be "On Average Atheists Outdo Believiers on Religion." Headlines must be short, but there is no way to accurately describe averages without explicitly invoking the concept.

This is part N in a series on how we have trouble with basic statistical concepts.

No comments: