Federal Review Composite Poll: Bush 290, Kerry 248

[phillybri]

http://www.federalreview.com/compositepoll.htm

Stable Race, Bush Still Leads Bush 49.00 | Kerry 47.59
May 11, 2004

Despite dire warnings from the pundits about Bush's electability after the USA Today Gallup Poll showed Bush's approval ratings relatively unchanged since February, the race remains tight, with Bush maintaining a slight lead over Kerry in this week's Composite Poll. Bush leads the Composite Popular Vote 49.0% to 47.6%, and he leads the Electoral Vote Prediction 290-248. When eliminating the Battleground states (which are in the "Lean" category where the current projected margin for any candidate is less than 6%), Bush has the advantage in states with 182 electoral votes, Kerry has the advantage in states with 108 electoral votes and the Battleground states are worth 248 electoral votes. Obviously, despite what Zogby says, this race is far from over. It is anybody's to win.

Although the reported numbers in the Electoral College are relatively unchanged from last week, there are significant changes in the raw numbers and the corresponding map colors. This results from a methodology change (I promise, it will probably be my last significant change). I have weighted recent state polling more heavily than in the past, assigning greatest weight to the most recent polls and less weight to older polls and to the Composite Poll baseline, which is based on 2000 results. Thus, a recent California poll showing Bush down by only 1% in California, throws that state into the "Lean" category for Kerry. But since Gore won California by 12% in 2000, this analysis takes that into some account, and we currently project Kerry to win California by 5.6% (in an April poll, Kerry held a 10 point lead in California). As future polls are included, we'll get a better idea of whether there really will be a race in California and whether the campaigns better conserve some cash for those expensive media markets. Do I believe the California poll that showed Kerry with a 1 point lead? I don't know, it does seem like an aberration, but the same poll showed Barbara Boxer with an 8 point lead over her Republican opponent, a result consistent with her 10 point win in 1998. At the very least, I think we can say that Kerry is garnering about 7% less support than Boxer is. But, in short, I need another poll!

[lancdem]

1. I dispute some of their projections

Pennsylvania, New Mexico, Florida and Arkansas are not leaning Bush. The last Pa. poll I've seen has them tied. All the states I cited are tied.

[Name removed]

3. Deleted message

Message removed by moderator. Click here to review the message board rules.

[LynneSin]

13. My knowledge of Pennsylania says it will go Dem, just like the last 3

Seriously, I'm not sure how anyone can POSSIBLY think Bush could win Pennsylvania, especially with Ohio (a state that has been voting repuke for years) is in the blue.

Simple, first we've won a statewide race with Ed Rendell as our governor. Anyone who wants to win PA will take a lead from that race and will pour 75% of their effort into getting out the vote in Philadelphia and the suburbs.

2nd, last time Arlen Spector ran for senate the same year their was a president race, the state split between the party. Pennsylvania tends to be a state where folks are NOT committed to pulling the lever for just one party. I've worked many campaigns in the Philly suburbs and it amazed me how most counties in that area would pick a repuke for senate and/or house and still vote dem for president.

3rd - the Unions, most notably the Steel Laborers, are very VERY upset with Bush. The thing is, central PA union members tend to vote republican and many of them are angry at Bush for revoking the Steel Tariffs (which were paying for the Health Care Tax Credit - a system that would allow retires not yet SSN age to get their health care at 35% cost. You don't think that's a big deal? Try telling that to the thousands of Bethlehem Steel retirees that lost their Health Care and Pensions last year) (BTW - good news - those Central PA steelworkers are also NOT keen on repukes trying to unseat democrat Tim Holden by throwing in a guy who hasn't worked more than 90 days in his life because his daddy is famous - Scott Paterno)

4th - Pennsylvania hasn't voted for a republican president since Reagan. Why would they suddenly change.

I still say that PA will be a tight race - but without a doubt it will be a blue state. Too bad we probably won't be able to get Joe Hoeffel in as senator

[zeemike]

2. I have often wondered

Polls are taken of "likely voters" How do they determine what a likely voter is?

[Name removed]

4. Deleted message

Message removed by moderator. Click here to review the message board rules.

[Thrill]

5. I don't get how polling

600 people is suppose to take the whole pulse of a state. And what is a likely voter?

[Name removed]

6. Deleted message

Message removed by moderator. Click here to review the message board rules.

[sangh0]

14. And reading helps, Brent

The post you responded to has a link to a page that explains how they figure out who a "likely voter is". You might get an answer to your question (ie. "What's a likely voter?") is you clicked on the link and read it

[Radical Activist]

8. A likely voters is...

someone who is likely to vote. So they only call those for the poll who vote regularly in elections, not people who only vote every four years. Other polls just call any registered voter. That means any poll of "likely" voters in a Presidential election year is biased towards Republicans because Republicans vote more often in non-Presidential elections. It really doesn't make sense to only poll likely voters in an election year that will probably have a high turnout of people who don't vote very often. It also means that if Democrats do a good job of motivating new voters and people who don't vote very often then Kerry will be way ahead of what those polls claim.

[phillybri]

7. Kick

:kick:

[snippy]

9. The first sentence of your post contains a false statement.

I do not know whether you are mistaken, have been misinformed, or are being mendacious, but the following statement is false:

the USA Today Gallup Poll showed Bush's approval ratings relatively unchanged since February

The accurate and truthful statement is:

"Bush's job approval rating has dipped to 46% -- the low point of his administration"

http://www.gallup.com/content/?ci=11668

I do not know whether there are any accurate statements in the remainder of your post as I stopped reading after the first sentence. However, I previously have reviewed the description of your methodology and wonder what your source is for the statistical validity of averaging different polls, conducted at different times for different clients and for different purposes, and of different sized samples of different population groups.

[Name removed]

10. Deleted message

Message removed by moderator. Click here to review the message board rules.

[snippy]

11. You did it again.

False statement:

"Bush's highest approval rating since February according to Gallup was 51"

The accurate and truthful statement would be that Bush's highest approval rating since February according to Gallup was 53.

http://www.gallup.com/content/?ci=11668

http://www.pollingreport.com/BushJob.htm

There is nothing wrong with honesty or accuracy.

You did not provide a source for the statistical validity of averaging different polls. When I took statistics I would have received a failing grade for doing so.

[Name removed]

12. Deleted message

Message removed by moderator. Click here to review the message board rules.

[sangh0]

15. Don't avoid the issue

What source do you want for statistical validity of averaging different polls?

You were asked to cite the sources *YOU* used. Don't you know the sources you used?

You don't like it. Fine. Don't look at what I've done. You don't find it useful. Good. Go bother someone else.

It's not a matter of like or don't like. You were asked to cite YOUR source, so it could be checked, an advisable activity seeing as how you've made at least two mistakes. For someone who is not an authority on the subject, and who has made some errors, your credibility is not so high that you can get sanctimonious about being questioned.

[Name removed]

17. Deleted message

Message removed by moderator. Click here to review the message board rules.

[snippy]

18. "There isn't one." Exactly. Thank you.

Averaging polls is meaningless, regardless of how bombastic the language used to make it appear as though such averaging is based on sound statistical analysis or laws of probability.

[MallRat]

16. California is LEANING Kerry? My ass.

WI should be leaning Kerry.
PA and FL should be too close to call.

Whose numbers is this guy using?

-MR