« Economist/YouGov Poll: Kerry Up by 2 | Main | Harris Poll: Kerry Ahead by 7 in Swing States, Down 2 Nation-wide »

PRC Poll: Kerry Up 6 in Battleground, Tied Nation-wide

John Kerry and George Bush are tied at 45 percent nation-wide RV's, with 1 percent for Nader, according to a Pew Research Center Poll conducted 10/15-19. The Poll also found that Kerry leads in "battleground states" 49-43 percent and Bush's approval rating is 44 percent.


Can someone please explain to me why these polls all have Kerry in the lead yet when I watch Hardball tonight they are saying Kerry is down in BGround states?
Please don't say Hardball is a conservatively skewed show, I think Mathews really does a good job of keeping an even hand.
But out of curiosity why is it that many shows list Bush as having these leads yet when I go to Liberal leaning websites they all list Kerry. Which one do I believe?

I noticed that Bush had been enjoying a significant lead among white Catholics in the previous three Pew Research Center Polls. In the most recent poll the results are almost reversed with Kerry up 7% among white Catholics. Is this just an anomaly? If not, what accounts for the improvement? Was it the last debate?

Andy Knox, have you been watching the same Chris Matthews on Hardball as I have?? The bias for Bush is unbelievable!! I gave up on Matthews when he constantly was crowing over how Cheney really got Edwards good, when he said he never met Edwards until the debate. Even after it was shown that Cheney had met Edwards several times, Matthews was still saying Edwards really got kicked by Cheneys' statement.

The education breakdowns on this poll are extraordinary. The college-educated and some college categories have been quite stable while the high school or less category has fluctuated wildly. The entire change from the last poll is due to high school or less going from 47-37 Bush to 46-41 Kerry. The pattern for income is similar, although less extreme, with the greatest fluctuations in the less than 20,000 category. I suspect that this labile behavior is due to small sample size because of the difficulty of reaching low-income respondents which increases both the sampling error and the estimation error of the weighting coefficients.

Note that since the last poll Kerry improved by 2% among whites and by 1% among non-whites, yet he imrpoved 3% overall. Thus much of Kerry's improvement is caused by the latest sample containing more non-whites.

It would appear, based on these observations, that in the polls Sept. 22-26 and Oct. 1-3 polls that showed Bush ahead, low-income pro-Kerry voters were underrepresented among the low-education respondents. The Sept. 17-21 and Oct. 15-19 polls that show a strong Kerry lead among voters with less than $20,000 income probably are much closer to the reality throughout the period.

This is striking evidence of the problems caused by the lack (for good practical reasons) of income-weighting of respondents in a year when voting behavior is very income-dependent, and education is not usable as a proxy for income because voters of the same income and different education levels vote very differently. See my posting on the new Wisconsin poll.

Only because the Pew poll gives more detail than other polls can we criticize it in such detail. I suspect that looking at other polls would turn up similar anomalies.

Matthews is a total chameleon. Actually, he is worse than
O'Reilly because he pretends to throw a few more bones to the dems. I can't believe this guy was employed by Thomas P. O'Neill. Look at his lineup! Andrea Mitchell? I haven't heard her say a positive thing about Kerry in months. Red Sox win and now Kerry wins!!

Andy--- There are LOTS of polls, and surely some put Kerry down in the battlegrounds. For example, if the spreads of the latest Zogby internet poll were the spreads on election day, Kerry would be in trouble. Matthews is not a pollster, and seems to pick his polling data in a haphazard manner. Moreover, he uses broad summaries of the numbers to support whatever story he thinks fits as the narrative of the race.

My take is that Kerry is up in the battlegrounds, and the last round of Zogby internet polls were the only datapoints I have seen that contradict this view.

This is a promising result, although Pew Center polls have yo-yoed around in ways hard to explain. Still, the trend seems at the moment to mostly be pointing in our direction, although it's so close it's hard to ascertain what changes are real and which are statistical "noise".

Ben Ross -- thanks for the analysis. I couldn't figure out why, after having what seemed to me consistent and believable results over the last few months, Pew suddenly diverged from other polls and its own prior results, showing a 7-point lead for Bush. That had me worried more than Gallup and the others, which tend to be all over the place anyway. Your focus on the volatility among the low-income and lower-educated is interesting and probably explanatory.

By the way, can anyone figure out what Pew means by a "battleground" state? They say they're using a new group of them, but I couldn't find what states they're talking about.

North Carolina Paper endorses Kerry/Edwards.

"Top of the Ticket: Kerry and Edwards"
The Pilot
Pinehurst, NC
October 21, 2004

President George W. Bush says that he’s a conservative. And that the opposite of conservative is “liberal” — which is what his Democratic opponent, Sen. John Kerry, is. He thinks that’s all voters in conservative places like Moore County need to know.

But conservative is not just the opposite of “liberal.” It is also the opposite of “reckless.” And Bush has behaved more recklessly than any president in living memory by, among other things, rushing America into an unnecessary war on a faulty pretext.

This reckless war, manned by a backdoor draft of reservists, has kill-ed more than 1,000 Americans and left thousands maimed. It has stretch-ed our military to the breaking point and diverted resources and attention away from the real war on terror and the real nuclear threats in Iran and North Korea. It has isolated the United States and made us weaker, not stronger. What’s conservative about that?

Possibly the most interesting part of this poll is the poll on Party ID (i.e. shoring up your base). Bush's early advantage in the polls was due largely to his partisan advantage among republicans versus weaker support among Democrats for Kerry. Notice the erosion of support for Bush and break toward to Kerry, especially among Democrats. Clearly due to Kerry standing up for himself and hammering Bush in the debates:

Numbers read Bush/Kerry +/- Bush
poll dates: 9/17-21 9/22-26 10/1-3 10/15-19

Republican 91- 4 90- 3 90-3 89-7 -1
Democrat 8-85 10-81 9-85 7-88 -2
Indep. 40-41 46-38 42-39 43-43 +1

Notice that Democratic support of Kerry has equaled Republican partisanship for Bush (89% Rep. vs. 88% Dem.). If we assume an overall Democratic Party ID advantage on election day similar to 2000, Bush is in bad shape. Notice also independent undecideds breaking +4 (39 to 43) for Kerry to +1 Bush (42-43) since just prior to the debates.

In past elections, the Republicans had higher partisan loyalty than Democrats, but party ID advantage sometimes gave Democrats the victory -- if their base turned out (as in 1992 and 1996) but victory to the Republicans if their base turned out in greater numbers (2002).

If partisan loyalty is really equal in this race as this poll indicates, then that is very bad news for Bush indeed, UNLESS you assume that party ID has shifted by about 4 points to the Republicans since 2000 because of 9/11 (as Gallup states to justify their poll weighting). We'll have to see after the election who's right.

I don't know how accurate this poll is overall, but this clearly shows a trend to Kerry and also indicates a very tight race - actually a dead heat. Not at all what is being promoted on CNN & Fox News.