Saturday, July 15, 2006

How NOT to report a poll

In recent years, the Associated Press seems to have decided that maintaining a reputation for fairness and credibility is far less important than crusading for liberal causes and promoting the Democratic Party line.

Case in point: Donna Cassata's Friday poll-mongering story, which claims that Democrats are favored by a 3-to-1 majority over Republicans. What utter tripe!

The most important words in the story are found in the fourth paragraph: "AP-Ipsos poll of 1,000 adults ..."

Adults! Random freaking adults!

As I explained in March: Random adults don't vote. All political polls of "random adults" will skew leftward for the very reason that Democrats do much better among people of limited education, people who don't know anything about the issues, people with limited interest in politics, the kind of people who NEVER VOTE.

This dirty little secret explains why Democrats have to run those "knock and drag" voter-turnout efforts, with activists going door-to-door on Election Day in low-income districts, offering bus rides to the polling places.

If you have a job and an education, chances are you've never experienced this treatment. Why? Because Democrats know that somebody like you is (a) civic-minded enough to get to the polling place on your own, and (b) disproportionately likely to vote Republican.

So you, the suburban homeowner, have never seen a knock-and-drag operation, and you don't understand why this type of activity is absolutely essential to the existence of the Democratic Party. This AP poll is a reminder: Democrats have an overwhelming majority of support among ignorant people, the kind who can't even name their U.S. Senators.

The less you know about the issues, the more likely you are to support Democrats but ... the less likely you are to vote at all. And these "random adults" who aren't likely voters tip the AP poll heavily to the left.

But Donna Cassata of the AP doesn't explain this to the reader. She doesn't distinguish "adults" from "likely voters," nor does she explain that this telephone poll probably tells us absolutely nothing about what's going to happen on Nov. 7.

Because any poll includes a number of non-voters -- even when they're trying to screen for likely voters -- all polls tend to overpredict the Democratic vote, because Democrats do better among non-voters. In fact, the person who tells a poll-taker that he will vote, but doesn't actually vote, is just about always a Democrat.

The Associated Press's poll-based reporting has become almost as bad as the New York Times, which notoriously misrepresents poll results in its coverage.

Sweetness & Light does a good job of dismantling this AP poll:
The AP article goes on to reiterate in several variations how impossible it will be for Republicans to hold onto their majority in Congress.
But, as regular readers of this site will
recall, the French company Ipsos always over-samples Democrats. And this case is no different
...
The larger picture, in terms of press bias, is this: The U.S. electorate has been trending steadily toward the Republicans for about 40 years. Yet every election gain for the GOP is greeted by the press as something shocking. The press was stunned by the three consecutive landslides (1980, '84, '88) during the Reagan-Bush era. The media gasped at the 1994 "Republican Revolution."

What have we seen in the past six years? Bush barely squeaked by in 2000, but scored a surprising off-year victory in 2002, then in 2004, became the first U.S. president to win an absolute majority of the popular vote in 16 years. Still, the press clearly wishes to believe that Bush and the GOP are widely hated; the press slants the news to put Republicans in the worst possible light; and yet, somehow, year after year, the election results don't match the media slant. Hmmmm.

Is it possible that the GOP could lose on Nov. 7? Sure, anything's possible. But it's not bloody likely. And this kind of slanted AP reporting, while perhaps helping the DNC's fundraising, will only increase Democratic voters' sense of betrayal and disappointment when (as we might reasonably expect) the Dems once again go down to defeat.

-- McCAIN