Captain's Quarters Blog


« Boston Labor Dispute Worsens | Main | McCain Steps Up Again »

July 17, 2004
CBS/NYT Poll: No Bounce For Edwards, Better Reporting, Same Old Methodology

The New York Times improved its polling reporting this week by having Richard Stevenson and Janet Elder write the article instead of Adam Nagourney, and the result is a more balanced look at a still-imbalanced poll. The overall result shows Kerr/Edwards leading nationwide by five points, an outlier of recent polling showing the race still a dead heat after the selection of John Edwards gave no momentum to the Democrats:

But naming Mr. Edwards did not immediately win over any substantial number of voters for the Democratic ticket, and the campaign between Mr. Kerry and President Bush remains statistically deadlocked as Mr. Kerry heads toward the Democratic convention and his best opportunity to make a strong impression on the country, the poll found. ...

While Mr. Kerry's selection of Mr. Edwards did not substantially alter the race, it corresponded with a deepening of the support for Mr. Kerry, though that support remained less intense than that for Mr. Bush. Forty-one percent said they strongly favored Mr. Kerry, up 10 percentage points from last month. Also, the percentage of people who said they were backing Mr. Kerry because they disliked the other candidates fell to 27 percent from 37 percent.

Among Mr. Bush's backers, 60 percent said they strongly favored him, up from 56 percent last month. Eight percent said they supported him because they disliked the other candidates, down from 11 percent last month.

The Times posts a PDF file of the poll analysis, but unfortunately, it's the poll analysis from the end of June, and not this poll. It shows the totals from each question, but one question in particular appears to call the results into question. When asked, participants responded that 35% were registered Democrats, 29% Republican, and 29% independent. As I've pointed out before, that doesn't represent the electorate properly, and if CBS/NYT weights it out as they've done in the past, they'll likely have made the situation worse, not better:

The CBS poll, on the other hand, uses only 1113 registered voters, broken down in an unusual manner: 346 Republicans, 390 Democrats, and 377 independents. Since when are there that many more Democrats than Republicans? The poll then shows its "weighting" (although it doesn't explain what it means), and the numbers get even worse: 330-R, 401-D, 381-I. According to the University of Pennsylvania in 2003, Republicans accounted for 32.5% of the registered electorate, while Democrats accounted for 33.7%. In a sample of 1113 voters, you would then expect to see 361-R, 375-D, 376-I. The result of CBS's sample is to throw off representation for Republicans by 8.6%, while bolstering Democrats by 7% and independents by 1.3%, using CBS' weighting.

In this case, the numbers appear to be only slightly closer to reality, and since neither outlet has released their methodology this time -- a highly suspicious omission -- it looks as though we have another skewed poll from the Tiffany Network and the Paper of Record. Undersampling GOP voters by 4% and oversampling Dems by a shade over 1% just about equals the gap between Kerry and Bush in their poll.

If CBS and the Times get around to releasing the methodology and raw data from this poll, I'll update you. The results are interesting as a trend, but I wouldn't put a lot of stock in the specific data as a prediction of voting behavior.

UPDATE: Longtime reader/commenter Gerry from Daly Thoughts has the link to the proper PDF for this study. As in the June PDF, no methodology of weighting is provided, but as Gerry notes in the comments, the gap in Dem/Rep representation is wider than first thought:

Despite a 2-point increase in the percentage of Democrats in the sample, the results were the same as the late-June result: Kerry 49, Bush 44 among registered voters. This time, CBS/NY Times included 29% Republicans, 37% Democrats, and 30% independents (4% did not know or refused to answer). In June they included 35% Democrats.

So we have 4% underrepresentation of the GOP and 4% overrepresentation of the Democrats, and still the Kerry/Edwards ticket only leads by five points. This can't build much confidence in the Democratic ticket.

Sphere It Digg! View blog reactions
Posted by Ed Morrissey at July 17, 2004 8:45 AM

Trackback Pings

TrackBack URL for this entry is



Design & Skinning by:
m2 web studios





blog advertising



button1.jpg

Proud Ex-Pat Member of the Bear Flag League!