More NY Times funny business?

Today's Times has an article under this headline: World's View of U.S. Sours After Iraq War, Poll Finds. Yet how reliable is this poll? At first glance, it looks reasonably thorough:

In a survey conducted since the end of hostilities in Iraq, the Pew Research Center for the People and the Press has released new data as part of an ambitious project to assess changes in global attitudes. The nonpartisan center surveyed 38,000 people in 44 nations during the summer and fall of 2002, and followed up with interviews of 16,000 people in 20 nations and in the areas administered by the Palestinian Authority after the end of the war.

This is the article's third paragraph, following up on the opening summary of the poll's most newsworthy findings (although the UN's dismal approval ratings somehow got buried down in paragraph 13). Yet in the very last of the article's 19 paragraphs, we read this:

The survey, called "Views of a Changing World," was released today by Madeleine K. Albright, the former secretary of state who is chairwoman of the global project. The latest survey was conducted with about 500 to 1,000 adults in 20 countries and the area administered by the Palestinian Authority. The polling was done by telephone and face-to-face interviews in late April and May. In most countries, the survey involved national population samples, but others were done only in urban areas. The margin of sampling error ranges from plus or minus three percentage points to four percentage points.

What do they mean, "500 to 1,000"? Don't they know how many were polled? And even if it was 1,000, that comes out to an average of under 50 persons per country. I'm no statistician, but I'll bet the proverbial dollars to doughnuts that you can't get accuracy of under 4% from a national survey of 50 individuals, even in a small country (though I also suspect "margin of sample error" means something rather narrower, despite how the expression is used without explanation in the article)

UPDATE: Mea culpa -- as a good scholar I should have first gone back to the original source, which is available here. And sure enough, it was a total of 16,000 individuals polled -- but I stand by my criticism of the Times, which should have stated, "The latest survey was conducted with about 500 to 1,000 adults in each of 20 countries. . . . "
Sloppy writing is sloppy reporting.

Posted by David on June 4, 2003 3:53 PM