Don't forget to hit refresh every time you visit to get the latest info!!!
FINAL POLLS:
___________________ Mean observed bias from major pollsters, when compared to other pollsters (number of observations): For Kerry: Even: For Bush: Averaging was done over hand-picked time periods of 2 1/2 weeks or less (in most cases) where there were at least 3 polls available. Rasmussen results were used if they overlapped the entire time period. Time periods listed below do not include Rasmussen dates outside the time period. If Rasmussen was used for two averages, the errors were combined into one observation. Strategic Vision was removed from the poll averages due to the inherent bias in a party-commissioned poll. Polls not on the list above were not included in the averages. States and dates used are listed below, and only polls after 04/30/2004 were used. Results were rounded to the nearest tenth percent. Bush wins, Kerry concedes. My comparison of my projections and the actual results is at the right. Below are data on state poll biases. State Poll Bias: Here are the observations for those two winning pollsters. Mason-Dixon: Predicted/Actual/(error) . Rasmussen: Predicted/Actual/(error)
Bias of other polls: Early (and very unreliable) exit polling results: Kerry up big in PA, up in FL, OH, NM. Also, a rumor that the race in VA may be very close. If you look at my map below, you'll see I picked a close one there. I'm also pulling in all the final poll data, so I can do analysis of the poll bias pretty quickly. One more thing: Vote if you haven't already!! Map updated, go hit Race2004.net if you want the polls, too many have come out. Here's my final prediction: Kerry 284, Bush 254. On the right, you'll see my state by state prediction of the margins. Since Florida and Ohio are so close, this could go either way. But Florida has Nader on the ballot, and Ohio does not, plus challenges are being curtailed/banned in Ohio. Don't forget to vote tomorrow!! Polls and map updated. Polls updated, lead switches to Kerry. Kerry's making great inroads in Ohio, Wisconsin, Minnesota, Colorado, and Iowa. However, Bush is striking back in states Kerry must win to get a victory: Michigan and Pennsylvania. Both states have now moved to tossup, and Bush has taken the lead in New Mexico as well. I have no idea how all these last minute movements will shake out. I will make a prediction on Monday, we'll see how good I am. And post-election, I'll give you some final detailed poll-bias numbers. Polls updated, still a Bush lead. However, note that Kerry's lead in OH and Bush's leads in FL and WI are less than 1.5%, basically a coin toss. Polls updated, Bush now in a slight lead. However, Bush's lead in many of the narrow states is very thin, especially considering the factor that undecideds usually go against the incumbent. In states like Iowa, Wisconsin, Florida, Ohio, New Mexico, Bush has rarely polled above 48%, which could mean he is in serious trouble in those states. New polls, moving Ohio to Kerry and Iowa to Bush, which puts Kerry in the lead. Ohio is still razor thin, but the Iowa lead is a bit stronger for Bush. Kerry can afford to lose Iowa, but only if he gets WI and OH, or if he gets FL. Also, the latest polls have the Colorado initiative failing. New polls, but no change in the total. I plan to do a post-election analysis in November on the poll biases, since I will have the election itself as a measure of how accurate the final polls were. A lot of new polls from Zogby and Survey USA. Florida and Ohio are both Bush leads by less than a point. In a surprising turn, the Bush leads in Virginia and North Carolina are now both under 4 points, moving them to Tossup Bush. Maybe John Edwards is having an effect after all! Some new polls, the race is getting a lot closer in North Carolina, but the polls still say Bush wins for the moment. However, Kerry would probably win Ohio (and maybe Florida) if the election were held today. But the effect of the Sinclair Broadcasting stunt is still unknown. Will it sway voters, or will the cynical undecideds be turned off by a large corporation trying to change their votes? Also, what effect will the protests by viewers and by Sinclair staff have? There has already been a major firing there, the Washington bureau chief is gone. My prediction from the beginning still holds: Bush needs both OH and FL. Kerry will likely get 246 EVs any way you cut it: HI/CA/OR/WA/NM (82), MN/IL/MI (48), DC/MD/DE/PA/NJ (52), NY/CT/RI/MA/VT/NH (61), and 3 of 4 EVs in ME. If Bush loses Florida, Kerry's at 273 and the game is over. If Bush loses Ohio, Kerry's at 266, meaning Bush can't afford to lose any of WI, IA, NV, or the CO initiative. Since it's a virtual tie in WI and IA before breaking the undecideds, and since the CO initiative is likely to pass, Ohio means Bush loses as well. A new poll has Bush leading in Oregon. However, given its smaller sample size, and the fact that all other polls have Kerry leading there, I would still assume that state will be a Kerry win. In Ohio, the margin is razor thin right now. Only a few new polls, and no change to the map. Before the election, I will do a state by state prediction. If I had to do it right now, I'd predict that Ohio would go to Kerry. Likely-voter polls usually exclude new registrants, and the registration effort in Ohio has been huge. But that is not enough, Kerry likely needs Wisconsin and Iowa as well. Bush will probably spend a lot of money in those states. New polls show Wisconsin and Iowa tilting ever so slightly to Kerry now, and Ohio is almost dead even, although I'm still giving it to Bush for the moment. A few more polls have been released, most just confirm that states like MD will go to Kerry and AR will go to Bush. Still waiting for new polls in OH, FL, WI, and IA, these could tell us more clearly how the race will end up. From the weekend news: Bush and Kerry basically tied in the debate, which can be seen as a victory for both: for Bush, it means he didn't get beaten twice, for Kerry it means he went into Bush's strongest format and didn't lose.
Also in the news: "Bulge-gate" won't go away, there's more speculation about Bush being prompted in the first debate. The speculation is especially sharp, since Bush has already been caught once being audio-prompted for a speech. And don't forget to check out the newest video from JibJab. The political jokes are a month old (it took me a bit to figure out they were lampooning recently outed NJ governor McGreevey), but it was pretty funny.
A new controversy in the debate rages: what did the candidates bring to the debate? Kerry pulled a pen from his pocket. At first, GOP bloggers said he pulled a card, but later views of the Fox footage clearly shows it was merely a pen. Now the focus has turned to Bush, and the mysterious "bulge" on the back of his coat. Does Bush have a receiver, and is he being prompted? Some evidence points to this being a possibility, which would amount to serious cheating. Checkout this link, read it, and tell me what you think. It's from a left-leaning source, but it is worth a read. Did Bush get prompted in the debate? Oh, what a tangled web we weave!! The Colorado EV-splitting initiative is likely to pass, but now the most recent polls show Kerry tied or slightly ahead! It's not hard to envision a scenario where Kerry wins Colorado, barely wins the EC, but the initiative passes. Then the GOP would be fighting in the courts FOR the intiative, and the Democrats would be fighting against the bill they just passed!! Also, don't forget the debate tonight. Lots of new polls out, they're mostly mild good news for Kerry. Yes, he's getting a debate bounce. No, it's not enough yet. In Kerry's key states of Ohio and Wisconsin, he's pulled back even. In states he should win, Michigan and Pennsylvania, he's holding on to his slim lead. Good news for Bush in Florida, where it looks like he may have an keep the advantage. If Bush can gain a few few points in Florida, it forces Kerry to win Ohio, which is a tough sell for a Democrat. The debate is tonight, and the format favors Bush a great deal. My best guess is that Bush will win this one by a little, or tie. However, that's all up to Bush himself. We know how Kerry will perform: solid, on point, and attacking the Bush record. The question is will Bush be able to connect with the audience? If he looks as unprepared as he did in the first debate, it could be a disaster. But I think that the ability to interact with people works to his favor, he is a natural at using a little humor and relating to an individual. Unless Bush flops, or he shines and Kerry stumbles, expect very little in the way of poll movement after this debate. In interesting news, Survey USA has polled Maine's two congressional districts, and it looks like they'll split, with the rural northern one going to Bush, and the more urban southern district going to Kerry. Overall, Kerry is still leading, so he should get 3 EVs, while Bush gets 1 EV. I'm updating my prediction accordingly. I watched the Cheney-Edwards debate, and my impression was that although Cheney should technically be given the win, Edwards still looked pretty good, he hit Cheney really hard with his comments about Meals on Wheels and Martin Luther King Day. Overall, the polls seem to give Cheney a slight win. But what's interesting about the poll numbers is that Edwards seems to have won among independents and undecideds. Either way, I have heard both conservatives and liberals say "our guy kicked ass". I guess they're both right, in that this was NOT the civilized discussion we saw in the 2000 veep debate. There was blood everywhere. All the polls are updated, but no states move as of yet. I will be interested to see any newer Ohio and Wisconsin polls. I have updated the states, leaving in a slight bias towards Kerry in borderline cases to reflect the huge swing in the national polls due to the debate. The first post-debate national polls are showing a huge swing toward Kerry. Newsweek has him up by 2, and even the Gallup poll has it as even. For Gallup, that's an 8-point debate swing. The net result of the newest polls is that Kerry solidifies his support base, but doesn't gain any states. However, only one or two post-debate polls have come out. Look for possible changes in Wisconsin, Iowa, Florida, Nevada, Ohio, or maybe even Colorado. No matter which way you spin it, Kerry won the debate. Even the GOP talking heads are admitting it, most are just trying to minimize the damage. It appeared that only Kerry realized that you were still on camera when the other guy was talking. Bush's smirks are reminiscent of Gore's sighs in the 2000 debate: they mean nothing, but yet they mean everything to the undecided voter. One liberal wag has dubbed it Bush's "constipated monkey" look. Polls are updated, but I won't change the map, the changes were small, no states switched sides, and there are no post-debate polls out. Gallup has released a poll that shows that Kerry clearly won the debate, especially among independent voters. Since Gallup usually skews GOP by 1% or so (see my analysis at the left), this is horrible news for the president. In side news, if you think hanging chads were controversial, just wait for Ohio in 2004, especially if the election is close. Apparently, there is a little-known provision in Ohio state law that allows for challenges to individual voters. You can't challenge everyone, but those who are challenged must answer a few questions under oath. It seems that the challenges will probably not scare away voters or disqualify legitimate voters, but they could be used to snarl the voting lines, thus discouraging voters from going to the polls. The GOP is planning to use them in Democratic neighborhoods. (Additionally, there is a case running through Ohio courts about whether or not you can vote even though you are at the wrong polling place.) So forget "chads", and get ready for "challenges"! Pacific West AZ: Bush +11 (B+7) CA: Kerry +11 (K>+10) HI: Kerry +9 (K+3) NM: Bush +1 (K+1) NV: Bush +3 (B+4) OR: Kerry +5 (K+6) WA: Kerry +6 (K+7) Open CO: Bush +7 (B+3) ID: Bush >+10
KS: Bush >+10
MT: Bush >+10
NE: Bush >+10
ND: Bush >+10
NE: Bush >+10
OK: Bush >+10
SD: Bush >+10
UT: Bush >+10
TX: Bush >+10
River AR: Bush +9 (B+3) IA: Bush +1 (K+1) IL: Kerry +11 (K>+10) LA: Bush +15 (B>+10) MN: Kerry +3 (K+4) MO: Bush +7 (B+3) WI: Kerry +1 (K+1) Heartland IN: Bush >+10 KY: Bush >+10 MI: Kerry +3 (K+5) OH: Bush +2??? (K<+1) TN: Bush +15 (B+8) South AL: Bush >+10 GA: Bush >+10 FL: Bush +5 (B<+1) MS: Bush >+10 NC: Bush +7 (B+6) SC: Bush +17 (B>+10) VA: Bush +8 (B+3) Mid-Atlantic DC: Kerry >+10 DE: Kerry +7 (K>+10) MD: Kerry +13 (K>+10) NJ: Kerry +7 (K+7) PA: Kerry +2 (K+3) WV: Bush +13 (K+4) Northeast CT: Kerry +10 (K>+10) MA: Kerry >+10 ME: Kerry +8 (K+9) NH: Kerry +1 (K+2) NY: Kerry >+10 RI: Kerry >+10 VT: Kerry >+10 It's a lot easier to analyze this from the Kerry point of view. First, let's see what Kerry will most likely win, barring major disruptions in either campaign. In the West, he'll win CA(55), WA(11), OR(7), NM(5), and HI(4). That's 72. He'll get IL(21) and MI(17) for sure, that's another 38, bringing the total to 110. In the Mid-Atlantic and Northeast, Kerry will win DC(3), MD(10), DE(3), NJ(15), NY(31), CT(7), RI(4), MA(12), VT(3), and the southern congressional district in ME (1). That's 89 more, bringing him up to 199 which we're pretty certain of. Now we come to the "likely" states for Kerry. First on the list are MN(10) and PA(21). Even before the debate, the biased Strategic Vision poll had Kerry leading in MN. And every post-debate poll in PA has Kerry leading by 2 to 7 points. Given that the challenger usually picks up more of the undecideds, and that new registrations have been much higher for Dems, we'll count these for Kerry. 199+31=230. Now let's add in a few "probable" states. Kerry is leading in some post-debate polls in WI(10), IA(7), NH(4), and he's also leading the "overall" vote for the state of ME(2). That's 23 more, 230+23=253. Kerry must pour some money into these states, he can't afford to lose any of them. We can not overemphasize how much it hurt when Kerry mispronounced the name of Lambeau field! So Kerry is now 17 short. The odds of him getting two of MO(11), TN(11), and AR(6) are pretty slim right now, only the Democrat-leaning Zogby has him up in those states, and only by a fraction of a percent. He's behind in NV(5), the chances of getting votes from the Colorado initiative are questionable, so that leaves (you guessed it) Ohio and Florida. Looking at the latest polls, OH(20) is much more likely than FL(27), but both are extremely close. If Kerry is running out of cash and time, I'd say Ohio is his best bet. The economy there is faltering, factories are closing, and they've had an enormous surge in Democratic voter registrations. If he can swing a few soccer moms, he can cinch it up. Bush's strategy is then counter to Kerry's. First, he has to try and sew up FL and OH. If he loses both, he's done. If he loses one, then he desperately needs to win in Wisconsin and maybe Iowa. So Bush should lock up MO & TN & AR, put Kerry on the defense in MN, WI, and IA, and try to win OH and FL. Of course, a rising tide lifts all boats, so we'll have to wait and see how the debates go.
Electoral and Poll websites
Electoral-Vote: Good map (Links are liberal.)
Am. Rsrch. Grp.: 1.4% (31)
Zogby: 1.4% (62)
Research 2000: 1.1% (9)
Survey USA: 0.3% Bush (22)
Rasmussen: 0.3% Bush (51)
LA Times: 0.5% Bush (7)
Quinnipac 1.0% (10)
CNN/USAToday/Gallup 1.3% (16)
Strat. Vis. (GOP): 2.7% (29)
Mason-Dixon: 3.3% (18)
Fox: 4.6% (4)
.
States and time periods covered:
AR 8/16-8/22
CO 9/10-9/17
FL 5/17-5/23
FL 6/12-6/27
FL 7/6-7/11
FL 7/13-7/23
FL 8/2-8/10
FL 8/17-8/22
FL 8/24-9/3
FL 9/12-9/21
IA 5/18-5/26
IA 7/26-7/30
IA 8/15-8/26
IA 9/12-9/21
MI 6/15-6/23
MI 7/6-7/10
MI 7/13-7/23
MI 7/26-8/4
MI 8/14-8/24
MI 8/28-9/3
MI 9/9-9/21
MN 5/18-5/24
MN 7/12-7/23
MN 8/15-8/21
MN 8/28-9/3
MN 9/10-9/17
MO 6/5-6/20
MO 7/6-7/10
MO 7/19-7/23
MO 8/15-8/24
MO 9/7-9/19
NH 9/12-9/17
NJ 8/6-8/23
NJ 9/11-9/19
NM 8/15-8/26
NM 9/12-9/17
NV 7/19-7/23
NV 8/14-8/21
NV 9/11-9/21
OH 5/10-5/25
OH 6/1-6/8
OH 6/15-6/23
OH 7/17-7/23
OH 8/9-8/24
OH 9/4-9/13
OH 9/13-9/20
OR 5/3-5/23
OR 7/26-7/30
OR 9/10-9/21
PA 5/18-5/25
PA 6/1-6/9
PA 6/15-6/23
PA 7/6-7/11
PA 7/18-7/23
PA 7/26-8/4
PA 8/11-8/21
PA 8/23-9/3
PA 9/4-9/14
PA 9/13-9/19
TN 6/15-6/21
TN 9/11-9/18
WA 6/1-6/6
WA 8/15-8/24
WA 9/5-9/18
WI 8/14-8/26
WI 9/9-9/17
WV 9/12-9/20
This did not play out as I expected. I did expect that Zogby leaned left and Strategic Vision leaned right, and they did. However, the center did not turn out to be the average of the polls, it was favoring the Democrats by about 2%. By far the most accurate polls with the least amount of error (at least for the swing states) were Mason-Dixon and Rasmussen. Mason-Dixon did have a few Bush outliers (including the wrong winner in MN), and Rasmussen had a Kerry outlier, but most of their predictions were very close. On average, Gallup did okay, until you look at the variance. What that shows is that Gallup is way off from the true number, but it could go either direction.
MO 10/27-10/29: B+5/B+7.3/(2.3 to K)
WI 10/27-10/29: K+2/K+0.5/(1.5 to K)
AR 10/27-10/29: B+8/B+9.3/(1.3 to K)
FL 10/27-10/29: B+4/B+5.1/(1.1 to K)
OR 10/27-10/29: K+6/K+5.3/(0.7 to K)
OH 10/27-10/29: B+2/B+2.5/(0.5 to K)
CO 10/27-10/29: B+7/B+7.0/(0.0 to -)
PA 10/27-10/29: K+2/K+2.2/(0.2 to B)
NH 10/27-10/29: K+1/K+1.4/(0.4 to B)
MI 10/27-10/29: K+2/K+3.1/(1.1 to B)
NM 10/27-10/29: B+4/B+1.7/(2.3 to B)
NV 10/27-10/29: B+6/B+2.7/(3.3 to B)
IA 10/27-10/29: B+5/B+1.1/(3.9 to B)
MN 10/27-10/29: B+1/K+3.6/(4.6 to B)
NJ 10/17-10/30: K+12/K+6.2/(5.8 to K)
FL 10/25-10/31: B+3/B+5.1/(2.1 to K)
OR 10/29-10/29: K+7/K+5.3/(1.7 to K)
WI 10/28-10/28: K+2/K+0.5/(1.5 to K)
IA 10/28-10/28: -+0/B+1.1/(1.1 to K)
MI 10/25-10/31: K+4/K+3.1/(0.9 to K)
NV 10/23-10/23: B+2/B+2.7/(0.7 to K)
PA 10/25-10/31: K+2/K+2.2/(0.2 to B)
OH 10/25-10/31: B+4/B+2.5/(1.5 to B)
NM 10/26-10/26: B+4/B+1.7/(2.3 to B)
MN 10/25-10/31: K+1/K+3.6/(2.6 to B)
(These first 4 had huge variances from the actual, there may be some inherent error in their methodology)
Harris: biased towards Kerry by average of 5.1% (3 observations)
LA Times: biased towards Kerry by average of 1.1% (3)
Fox/Opinion Dynamics: biased towards Kerry by average of 0.8% (4)
CNN/USAToday/Gallup: biased towards Kerry by average of 0.6% (6)
.
(The rest had low error levels-except Zogby-but an inherent bias)
Zogby: biased towards Kerry by 3.1% (14, moderate error)
ARG: biased towards Kerry by 1.5% (8)
Survey USA: biased towards Kerry by 1.5% (11)
Rasmussen: biased towards Kerry by 0.7% (11)
Mason-Dixon: biased towards Bush by 0.6% (14)
Quinnipac: biased towards Bush by 2.1% (3)
Strategic Vision: biased towards Kerry by 2.2% (8)
.
(States used for this analysis were AR, CO, FL, IA, MI, MN, MO, NH, NM, NJ, NV, OH, OR, PA, WI.)
Update (10/6/04):
AK: Bush >+10
NORTHERN MAINE: Kerry +4
SOUTHERN MAINE avg: Kerry >+10
Analysis
RealClearPolitics: All polls updated. (Editorials are conservative.)
Historical U.S. Electoral College results
D.C. Political Report: Good insider info on political races
ARG's Electoral Vote Calculator
Dales' Electoral College Breakdown (very conservative, good info on weakness of Rasmussen)
270 To Win: Great historical info, plus a clickable map to figure out strategy!
LOWV interactive map: strategies plus prior election results!
Triapas: Cool site, you can choose which polls to include/exclude!
Ed Fitzgerald's Unfutz: a compilation of EC numbers from a bunch of websites like mine.
Race2004.net: The BEST site to find new polls, hands down.
Great collection of electoral websites.