Monday, August 2 2004
There seems to be some confusion over which poll numbers we use in our RCP Polling Averages. Today's Gallup results are a perfect example. As soon as the results were posted yesterday, we started to receive dissatisfied emails wanting to know why we were posting only the Likely Voters results as opposed to the Registered Voters numbers.

There is a lot of emotion among partisans on both sides in the outcome of this election and naturally each side wants to see the most favorable poll results pushed for their respective candidate. Kerry supporters obviously would prefer the press focus on the Registered Voters result which has Kerry tied 47% - 47% (or even better the National Adult sample that has Kerry ahead 48% - 45%) instead of the Likely Voters result we posted that shows Kerry trailing Bush 50% - 46% in the three way race.

Two weeks ago, we had Republican partisans emailing us that we had the wrong FOX News poll numbers posted when, in fact, we had the correct Likely Voters results. These people were incorrectly looking at the Registered Voters result which (in a rarity) had President Bush doing better among Registered Voters than among Likely Voters.

The first thing we try to do is to be consistent in regards to polling data. We will always use Likely Voters numbers over Registered Voters, and Registered Voters numbers over National Adults. The reason is clear and simple. Likely Voters results are a better predictor of how an election actually turns out. In other words, they are more accurate.

People on both sides of the aisle usually recognize this fact without much argument, though you will occasionally hear complaining from Democrats that this usually tilts the polling results toward the GOP. But Democratic complaints really shouldn't be with pollsters who release Likely Voters results, but rather with the Registered Voters and Adults who respond in the surveys but don't bother to vote. Again, when a polling firm releases LV numbers and RV numbers we will always post the Likely Voters result.

Another source of confusion is whether we should highlight the three-way numbers with Nader or just the Bush-Kerry head-to-head results. Unlike the issue over LVs vs. RVs, this is more debatable and reasonable people can disagree over which series will be more accurate. We post both sets of results here at RealClearPolitics, but when just quoting one number from a poll we will use the three-way with Nader.

Again, Democrats sometimes will balk at this because usually it works to Bush's advantage. Now if Nader doesn't end up on the ballot in Florida, Ohio, Pennsylvania, Wisconsin, Missouri and Michigan then perhaps at some point we'll switch to using the head-to-head number as the one to quote. But given the results of the 2000 election it seems foolish to ignore the presence of Nader in the race.

To summarize: when we get a poll like the just released CNN/USA Today/Gallup with six different horse race numbers in the same poll. (3-way LV, 3-way RV, 3-way NA, head-to-head LV, head-to-head RV, head-to-head NA), we will always post the Likely Voters result in our RCP Poll Averages and we will highlight the three way with Nader as long as Nader remains in the race and on the ballot in critical states.

BUSH 50%, KERRY 46%, NADER 2%: It's hard to know what to make of the Gallup poll showing a negative 'bounce' for the Kerry/Edwards ticket. My first reaction is that everyone should take a deep breath before drawing huge conclusions from this one poll.

However, the results from the other post-convention poll from Newsweek doesn't provide much solace for the Kerry campaign when you read this: "Kerry’s four-point “bounce” is the smallest in the history of the NEWSWEEK poll."

Rasmussen's daily tracking poll shows a similar 2-4 point bounce and Zogby's poll taken July 26-29 during the Democratic Convention gives the Kerry/Edwards ticket a three point bump from their previous poll.

All told, the early evidence seems to indicate that Kerry will not receive much of a bounce, if any, from his convention in Boston. That in and of itself isn't a killer, because it is easily possible that given the polarization of the electorate neither candidate will receive much of a bounce from their convention. But the negative bounce in the Gallup poll, will be a serious concern for the Democrats if the other major polls confirm the same trend.

But even if the other polls are more similar to Newsweek and Rasmussen showing only a small bounce for the Kerry/Edwards ticket, that isn't necessarily good news for the Democrats. For while it is possible that people's opinions are so set there will not be any big bounces from either convention, it is also entirely possible Kerry gets a meager bump while Bush still gets the normal big boost from his convention.

So while these early polls aren't the end of the world for Kerry - or even necessarily conclusive evidence as to what kind of bounce Kerry may ultimately have gotten from his convention - there is no way the Kerry campaign can spin them as good news for their candidate. J. McIntyre 8:56 am Link | Email | Send to a Friend


© 2000-2004   All Rights Reserved

RCP Polling Information
3 Way Race: Bush-Kerry-Nader
Head-to-Head Race: Bush-Kerry
State Polls: Battleground | Other

RCP Blogroll
Andrew Sullivan Milt Rosenberg
Armavirumque Morning Grind
Atrios No Left Turns
Best of the Web The Note
Bill Hobbs Oxblog
The Corner Pejmanesque
Daily KOS Polipundit
Dan Drezner Political Animal
Donald Luskin Political Wire
Donald Sensing PowerLine
Drudge Rich Galen
Easterblogg Robert Tagorda
First Read Roger L. Simon
Hit and Run Ryan Lizza
Hugh Hewitt Scrappleface
Instapundit SD Politics
James Lileks TPM
Jeff Jarvis Tapped
John Ellis TNR
Kausfiles Tim Blair
Kevin McCullough Virginia Postrel
Matt Rosenberg Volokh
Matthew Yglesias Wonkette

Archives - 2004
7/19-7/25 | 7/12-18 | 7/5-11 | 6/28-7/4 | 6/21-6/27 | 6/14-20 | 6/7-13 | 5/31-6/6 | 5/24-30 | 5/17-23 | 5/10-16 | 5/3-5/9 | 4/26-5/2 | 4/19-25 | 4/12-18 | 4/5-11 | 3/29-4/4 | 3/22-28 | 3/15-21 | 3/8-14 | 3/1-7 | 2/23-27 | 2/16-22 | 2/9-15 | 2/2-2/8 | 1/26-2/1 | 1/19-25 | 1/12-18 | 1/5-11 | 12/29/03-1/4/04

Archives - 2003
12/22-28 | 12/15-21 | 12/8-14 | 12/1-7 | 11/24-11/30 | 11/17-11/23 | 11/10-11/16 | 11/3-11/9 | 10/27-11/2 | 10/20-26 | 10/13-19 | 10/6-10/12 | 9/29-10/5 | 9/22-28 | 9/15-9/21 | 9/8-9/14 | 9/1-9/7 | 8/25-8/31 | 8/17-8/24 | 8/11-8/16 | 8/4-8/10 | 7/28-8/3 | 7/21-7/27 | 7/14-7/20 | 7/7-7/13 | 6/30-7/6 | 6/23-6/29 | 6/16-6/22 | 6/9-6/15 | 6/2-6/8 | 5/26-6/1 | 5/19-5/25 | 5/12-5/18 | 5/5-5/11 | 4/28-5/4 | 4/21-4/27 | 4/14-4/20 | 4/7-4/13 | 3/31-4/6 | 3/24 - 3/30 | 3/10 - 3/17 | 3/3-3/9 | 2/24-3/2 | 2/17-2/23 |
2/10-2/16 | 2/3- 2/9 | 1/27 - 2/2 | 1/20 -1/26 | 1/13-1/19 | 1/6-1/12 | 12/31/02-1/5/03

Archives - 2002
12/23-12/29 | 12/16-12/22 | 12/9-12/15 | 12/2-12/8 | 11/25-12/1 | 11/18-11/24 | 11/11-11/17 | 11/4-11/10 | 10/28-11/3 | 10/21-10/27 | 10/14 -10/20 | 10/7-10/13 | 9/30-10/6 | 9/23 -9/29 | 9/16-9/22