Poll watch: As LV and RV models split, Obama leads VA, McCain stops bleeding in yet another FL poll

October 22, 2008

We are starting to see polling taken after the week-end (and thus after the Powell endorsement and McCain’s socialism charge), and there is little sign that McCain is closing the gap. He does gain a bit in two of the tracking polls, but he loses ground in four others, as Zogby, Research 2000 and Washington Post/ABC now all show Obama leading by double-digits. In all 10 of the national polls released today (including the AP survey, about which I will talk in a minute), McCain is stuck in the low 40s, between 40% and 45%.

One possible worry for Obama is that the size of his national lead is due to his gains in states that will not influence the electoral college: We have been seeing Obama open dramatic leads in safe blue states like California and Washington and cut margins significantly in places like Texas and Kentucky. The trends in places like Ohio and Florida are at a much smaller scale (surely because the volume of campaigning and advertisement makes these states less susceptible to follow national trends). So could the size of Obama’s lead in non-battleground states be obscuring a tighter race in the electoral college?

There isn’t much evidence of that in polls from battleground states, where Obama continues to get strong numbers - though he hasn’t put it away the way the way he appears to have secured a popular vote lead. But he dominates in Virginia, where CNN/Time finds him leading by double-digits yet again. Mason-Dixon does find the Old Dominion within the margin of error, but its previous survey had been the only one with McCain ahead since September. Furthermore, Obama leads outside of the margin of error in three out of the five CNN/Time polls (VA, Nevada and Ohio) and is leading within the MoE in two polls of North Carolina.

The good news for McCain comes from Florida: his lead in Mason Dixon is well within the margin of error, but it is the fourth survey in a row to find McCain gaining in the Sunshine State, a significant break from Obama’s fifteen consecutive - many of which were outside of the margin of error.

The second good news for McCain comes from the much-discussed AP poll that has a 1% lead. But three remarks apply here. First, McCain is stuck in the same range as every other poll (the low 40s), and Obama is much lower than his national average. As long as McCain cannot break 45% (or 46% in his best Rasmussen days), he doesn’t have much hope of besting Obama nationally. Second, Marc Ambinder remarks that evangelicals make up about twice as much of the sample as they usually do. Third, this gets us to the important slip between registered voters and likely voters.

Obama leads by 5% among registered voters in AP’s poll, a differential that also exists in Gallup’s tracking (+9% among registered voters, +5% or +8% among likely voters). And it is most dramatic in CNN/Time’s state polls. In all five, Obama performs better among RVs than among LVs (especially in Nevada, where he is ahead by 13% among RVs). What this means is very simple: Obama will benefit from higher turnout, and the size of his lead is partly dependent on how tight a likely voter screen pollsters apply.

There are clear indications that turnout will be larger than usual, particularly among Democrats, meaning that Obama’s lead could range somewhere between the LV screen and the RV results. Early voting numbers are going through the roof among Democrats and African-Americans in North Carolina, Georgia and Nevada; furthermore, Gallup’s tracking poll acknowledges that the traditional LV model might not apply - which is why they have an expanded model which closely mirrors the RV results.

That said, it is impossible to predict how large turnout will be and whether Obama’s organization will fully function. And that’s why we have elections. On to the full roundup of the day’s polls:

  • Obama remains in command of the tracking polls, though they are not moving as uniformly in his direction yesterday. Obama gains 2% in Zogby (up 52% to 42%), 2% in Research 2000 (up 51% to 41%), 2% in Rasmussen (up 51% to 45%) and 2% in Washington Post/ABC (54% to 43%). Hotline finds a stable margin (47% to 42%). McCain gains 2% in Gallup’s expanded likely voter model (52% to 44%, with a 9% lead for Obama among RVs and a 5% lead in the traditional LV model), 2% in IBD/TIPP (46% to 42%). To recap, Obama’s leads are: 4%, 5%, 6%, 8%, 10%, 10%, 11%.
  • Obama leads 49% to 40% in a national Fox News poll conducted Monday and Tuesday. He led by 7% two weeks ago. Who knew a few months ago that Obama would achieve the support of 88% of Democrats (versus 83% of Republicans for McCain)? Interestingly, 66% of Democrats and 47% of independents think that spreading the wealth is a good idea.
  • Obama leads 44% to 43% in a national AP/GfQ poll conducted Thursday through Monday. He led by 7% three weeks ago. Obama leads by 10% among all adults and by 5% among registered voters, however.
  • Obama leads 50% to 42% in a national poll conducted by Ipsos/McClatchy conducted Thursday through Monday.
  • Obama leads 54% to 44% in a CNN/Time poll of Virginia. Among registered voters, Obama leads 54% to 42%. When other candidates are included, he leads 51% to 44%.
  • Obama leads 47% to 45% in a Mason Dixon poll of Virginia. McCain led by 3% two weeks ago.
  • Obama leads 51% to 47% in a CNN/Time poll of Ohio, just within the margin of error. Among registered voters, Obama leads 51% to 45%. When other candidates are included, he leads 49% to 44%, with 2% for Barr (50% to 43% among registered voters).
  • Obama leads 48% to 46% in a WSOC-TV poll of North Carolina.
  • Obama leads 50% to 46% in a CNN/Time poll of North Carolina, just outside of the margin of error. Among registered voters, Obama leads 51% to 46%. When other candidates are included, he leads 51% to 46%, with 2% for Barr.
  • Obama leads 51% to 46% in a CNN/Time poll of Nevada. Among registered voters, Obama leads 54% to 41%. When other candidates are included, he leads 49% to 43%, with 3% for Nader and 2% for Barr.
  • McCain leads 46% to 45% in a Mason Dixon poll of Florida. Obama led by 2% two weeks ago.
  • Obama leads 52% to 41% in a Research 2000 poll of Wisconsin, conducted Monday and Tuesday.
  • Obama leads 51% to 38% in a Wisconsin Public Radio poll of Wisconsin. However, the poll was was conducted form the 9th to the 17th, so it is not at all an indicator of what is going on currently on the ground.
  • Obama leads 55% to 36% in an Elway poll of Washington.
  • McCain leads 53% to 42% in an Ivan Moore poll of Alaska. McCain led by 17% two weeks ago.
  • McCain leads 53% to 44% in a CNN/Time poll of West Virginia, just outside of the margin of error. Among registered voters, Obama leads 51% to 44%.
  • McCain leads 42% to 41% in a one-week old poll of West Virginia conducted by Democratic-firm Rainmaker.
  • McCain leads 54% to 42% in a Rasmussen poll of Tennessee. He led by 19% last month.

Meanwhile, in down-the-ballot polls:

  • Mark Begich is ahead 47% to 46% in an Ivan Moore poll of the Alaska Senate race. Begich led by 4% two weeks ago.
  • Kay Hagan leads 44% to 43% in a WSOC-TV poll of North Carolina’s Senate race.
  • McConnell leads 50% to 43% in a Rasmussen poll of Kentucky’s Senate race. He led by 9% three weeks ago.
  • Mary Landrieu leads 54% to 34% in an internal poll of the Louisiana Senate race.
  • Chris Gregoire leads 51% to 39% in an Elway poll of Washington’s gubernatorial race.
  • Perdue and McCrory are tied at 44% in a WSOC-TV poll of North Carolina’s gubernatorial race.
  • In AK-AL, Ethan Berkowitz leads 51% to 43% against Don Young in an Ivan Moore poll. He led by 9% two weeks ago.
  • In FL-18, Rep. Ros-Lehtinen leads 48% to 41% in an internal poll for Democratic candidate Annette Taddeo.

No big surprises in this batch of congressional polls. If anything, the news is good for the GOP as Sens. Stevens and Dole stay within the margin of error in their respective cases (as we await the verdict of the Stevens trial) and as Mitch McConnell remains ahead outside of the margin of error in Rasmussen’s survey. But the Louisiana numbers are naturally excellent news for Democrats; while Landrieu’s own survey might be overstating her lead, it does confirm the conventional wisdom that the incumbent is ahead.


Eleventh presidential ratings: Obama consolidates electoral college lead

October 20, 2008

A week after Obama surged to a dominant position, the ratings remain relatively stable, with only one state shifting in or out of a candidate’s column. There is movement under the surface, however, as McCain’s base continues to erode while Obama solidifies his hold on a number of states; a total of 26 electoral votes move from the lean Obama to the likely Obama column, giving the Democratic nominee a base of 260 electoral votes.

In my September 20th ratings - posted exactly a month ago - 18 states were listed in a competitive category (lean or toss-up). Of these, not a single one is today in a more favorable category for McCain but fourteen have shifted towards Obama. In fact, 8 of these states are no longer competitive at all - and they now all belong to the Democratic nominee. They have been replaced by four new red states that were solidly anchored in McCain’s column a month ago and are now considered competitive.

What better way to illustrate how much the electoral map has shifted towards Obama over the past month, and how most of these changes will not be erased no matter how much McCain closes the gap in the final 16 days. Unless some major event turns the campaign on its head, Michigan or Iowa, for instance, are now out of contention.

This also illustrates how narrow McCain’s electoral strategy has become: He needs to sweep nearly all of the 14 states currently rated as competitive, including all three red states that are in the Obama column. That is no small feat, and it is revealing of just how much Obama is command. That said, there is a reason these states are still listed as competitive: they could go either way, and a slight wind pushing McCain over the final two weeks could help him accomplish that.

Without further delay, here are the eleventh electoral college ratings (states whose ratings have been changed towards Obama are colored blue, those whose ratings have been changed towards McCain are colored red):

  • Safe McCain: Alabama, Alaska, Idaho, Kansas, Kentucky, Louisiana, Mississippi, Nebraska (at large + 3rd congressional district), Oklahoma, South Carolina, Utah, Tennessee, Texas, Wyoming (116 EVs)
  • Likely McCain: Arizona, Arkansas, Nebraska’s 1st district, South Dakota (20 EVs)
  • Lean McCain: Georgia, Montana, Nebraska’s 2nd district, West Virginia (24 EVs)
  • Toss-up: Indiana, Missouri, Nevada, North Carolina, North Dakota, Ohio (65 EV)
  • Lean Obama: Colorado, Florida, New Hampshire, Virginia (53 EVs)
  • Likely Obama: Iowa, Oregon, Maine (at-large + 1st district + 2nd district), Michigan, Minnesota, New Jersey, New Mexico, Pennsylvania, Washington, Wisconsin (107 EVs)
  • Safe Obama: California, Connecticut, DC, Delaware, Hawaii, Illinois, Maryland, Massachusetts, New York, Rhode Island, Vermont (153 EVs)

This gives us the following map and totals:

  • Safe + Likely Obama: 260 electoral votes
  • Safe + Likely + Lean Obama: 313
  • Toss-up: 65
  • Safe + Likely + Lean McCain: 160
  • Safe + Likely McCain: 136

I will naturally not attempt to provide an explanation for every single one of these ratings and will concentrate instead on those that have shifted over the past two weeks:

Alaska, likely McCain to safe McCain: Like in other red states Obama had been eying, McCain jumped to a commanding lead in Alaska in the aftermath of the GOP convention and of the Palin pick. Unlike in some of these other red states (say, North Dakota and Montana), McCain’s surge has not faded over the past month. The Sarah Palin effect is strong, and it appears to have put Alaska’s once-promising 3 electoral out of Obama’s reach for good. In fact, the GOP’s recovery is so pronounced that it could very well save Sen. Stevens and Rep. Young.

Arkansas, safe McCain to likely McCain: Arkansas is very rarely polled, but perhaps there would be some interesting results to be found. The state remains heavily Democratic, though it is made up of conservative Democrats who vote GOP in federal races. Obama was not expected to do well among conservative Democrats and blue-collar voters, but the startling finding that he is competitive in West Virginia means that he is making inroads in the type of constituency that could help close the gap in Arkansas.

Maine’s 2nd district, lean Obama to likely Obama: Despite a week of GOP advertisement and a visit by Sarah Palin, the GOP does not appear satisfied with the odds of snatching away one of Maine’s four electoral votes, as we learned this week that the RNC is moving out just as quickly as it moved in to help protect red states. The McCain campaign is staying on the state’s airwaves but a recent Research 2000 poll showing Obama with large leads in both districts and statewide suggest that the RNC’s pull-out was a wise decision.

Minnesota, lean Obama to likely Obama:  On paper, Minnesota should not have been have been as vulnerable as neighboring Wisconsin or Michigan, but the polls here tightened more than in other blue states throughout August and September. But a sign of Democratic confidence came from the two campaigns’ expenditures: Minnesota is the only state in which Obama let McCain outspend him by significant amounts, signaling that he believed Minnesota remained solidly anchored in his camp. Now, Obama is matching McCain’s spending (another sign of Democratic confidence given that Obama is outspending his opponent by massive amounts in every other battleground state but Iowa), and polls are reflecting the state’s return to its Democratic roots. Obama leads by double-digits in CNN/Time, Research 2000, Star Tribune, Quinnipiac… Even SUSA now has Obama leading outside of the margin of error. Do I need to say anything else?

New Mexico, lean Obama to likely Obama: New Mexico was the second red state to move to the Obama column - and it did so early. In fact, Obama started enjoying double-digit leads in New Mexico polls well before he did in blue states like Minnesota or Michigan. One significant factor has been Obama’s strength among Hispanics; when it was still believed (back in primary season) that Obama might have problems among that group, it looked like the Southwest could be promising territory for McCain. But it will be hard for the Republican to stay competitive in the state unless he can perform at Bush’s level among Latinos - and every indicator suggests that he is underperforming.

North Dakota, likely McCain to toss-up: Three successive polls released over the past week have found an Obama lead or an exact tie in a state that Democrats abandoned in mid-September, after McCain’s post-convention surged appeared to put North Dakota and the rest of the Mountain West out of contention. With 15 days to go until Election Day, there is increasing speculation that Obama is looking to put resources in the state in a last-minute bid to recapture its electoral votes - and polls indicate that would be a wise decision. One interesting fact about this state is that it does not have any voter registration: any one who has lived in a precinct for the past 30 days can show up and cast a ballot.

South Dakota, safe McCain to likely McCain: The latest polls from the state find a large lead for the Republican nominee, but we have had no result since mid-September. Since then, Obama has made gains in the Mountain West, and it is unlikely that he has been able to tie the race in Montana and North Dakota without also making some inroads in South Dakota.

Wisconsin, lean Obama to likely Obama:  Among the tightest states of the 2000 and 2004 contests, Wisconsin does not look like it will be decided in the early hours of the morning this year. In fact, the Badger State never emerged as a true battleground this year; only during a brief patch in mid-September did Obama’s lead descend in the mid single-digits - certainly nothing to be panicked about. Since then, Obama has recaptured a double-digit lead, and while Quinnipiac’s 17% margin might be overstating his advantage, but the Univ. of Wisconsin, SUSA, or Research 2000 aren’t that far off. And we got confirmation of the fact that Wisconsin is no longer in the top-tier of competitive races when the RNC’s independent expenditure arm pulled out of Wisconsin this week; it had been airing ads in the state since its very first wave of expenditures back in June.

History of Campaign Diaries’ electoral ratings:

  • October 20th: + 153 Obama (313 for Obama [153 safe, 107 likely, 53 lean] and 160 for McCain [116 safe, 20 likely, 24 lean])
  • October 12th: + 150 Obama (313 for Obama [153 safe, 81 likely, 79 lean] and 163 for McCain [122 safe, 17 likely, 24 lean])
  • September 27th: + 55 Obama (239 for Obama [154 safe, 43 likely, 42 lean] and 174 for McCain [122 safe, 38 likely, 14 lean])
  • September 20th: +6 Obama (222 for Obama [154 safe, 19 likely, 49 lean] and 216 for McCain [119 safe, 41 likely, 56 lean])
  • August 31st: + 16 Obama (243 for Obama [154 safe, 29 likely, 60 lean] and 227 for McCain [93 safe, 75 likely, 59 lean])
  • August 20th: + 14 Obama (238 for Obama [151 safe, 32 likely, 55 lean] and 224 for McCain [90 safe, 75 likely, 59 lean])
  • July 30th: + 38 Obama (238 for Obama [151 safe, 42 likely, 45 lean] and 200 for McCain [90 safe, 75 likely, 35 lean])
  • July 16th: +28 Obama (255 for Obama [150 safe, 43 likely, 62 lean] and 227 for McCain [90 safe, 78 likely, 59 lean])
  • July 2rd: +11 Obama (238 for Obama [143 safe, 50 likely, 45 lean] and 227 for McCain [93 safe, 78 likely, 56 lean])
  • June 18th: +22 Obama (238 for Obama [86 safe, 97 likely, 55 lean] and 216 for McCain [87 safe, 87 likely, 42 lean])
  • June 4th: +20 McCain (207 for Obama [76 base, 107 likely, 24 lean] and 227 for McCain [97 safe, 77 likely, 53 lean])

Poll watch: Trackings tighten but Obama remains in command; Hagan, Collins and M. Udall lead

October 18, 2008

The latest tracking polls suggest the presidential race might be tightening. Though the trend lines are not uniform, Obama’s biggest lead is 7% whereas last Saturday three trackings had him at 9%, 10% and 12%. In fact, compared to last week four of five trackings have McCain in a significantly stronger position, while the margin is the same in the fifth. That said, it also looks like all the trackings are converging to a similar 6-7% range, with most of the six polls released today going in opposite directions but towards that margin:

  • Research 2000 had shown a double-digit lead for Obama (between 10% and 13%) since September 13th; today’s poll showed a lead of “only” 7%, 50% to 43%. Given the stability of the past three weeks, this was certainly a noticeable drop, especially because it is due to two consecutive nights of tighter polling (+6% on Wednesday night and +7% on Thursday night). On the other hand, IBD/TIPP had been showing a tighter race than the other trackings, but Obama has now opened up a 7% lead.
  • A third tracking poll (Diego Hotline) also shows a 7% lead, down from 10% yesterday but similar to the last pre-debate poll. Two other trackings have slightly smaller margins, but no uniform trends: Obama’s lead in Zogby goes from 5% to 4% while his edge in Rasmussen goes from 4% to 5%.
  • That leaves us with the silliness that has become Gallup’s tracking poll. Gallup’s reputation should make its tracking poll the most important of the bunch, but the firm’s decision to release three different daily measures is making this impossible to follow: Obama’s lead today is 2% (likely voters, traditional model), 4% (likely voters, expanded model and down from 6% yesterday) and 8% (registered voters, up from 7% yesterday). Gallup thus ensures it cannot be wrong by offering a variety of margins and a trend towards both men - but how are we to know just what model Gallup thinks reflects the situation on the ground, what model Gallup’s interviews are suggesting should be the closest to the truth?

Taken together, these polls should give the McCain campaign some reason to hope, as Obama’s lead appears to now be hovering in the mid-to-high single digits, a far more manageable gap than the double-digit advantage Obama was posting last week. It should also serve as a big relief for Republicans, as the post-debate snap polls favored Obama and suggested that the Democrat’s lead would grow rather than shrink.

That said, a 5% to 7% lead in the second half of October is far more significant than a similar lead in the summer. Bush only led by more than 5% in three polls in the entire month of October 2004. Furthermore, it will take a few more days to see whether the national tightening (if there is indeed a tightening) will be spread homogeneously across the country or if it will primarily affect certain regions or demographics. Yesterday, we saw a few post-debate state polls that had Obama in a commanding position (CO, NV and MO). Today’s state polls were taken before the debate, but they have good news for Obama:

  • Despite most recent polls showing Obama in the lead in Virginia, McCain leads in Real Virginia. It is unclear how many electoral votes the McCain campaign expects to gain from Real Virginia, nor by how much the GOP’s polls show McCain leading Real Virginia.
  • Obama leads 46% to 44% in a Research 2000 poll of North Carolina. A month ago, McCain led by 17%, but that survey looked like an outlier. The poll was taken on the 14th and 15th, so right before the debate.
  • Obama leads 53% to 38% in a Research 2000 poll of Maine. He leads 58% to 35% in the first district and 52% to 41% in the second district, so there is little danger of McCain snatching away an electoral vote.
  • Obama leads 47% to 43% in a poll of Florida conducted by a Democratic firm, Hamilton Campaigns, before the debate. One big problem in the poll’s internals is that Obama leads by 18% among Hispanics. Republicans are very strong among Florida Hispanics because of the high number of Cubans, and most other polls show McCain with a narrow edge among that constituency. Obama leads in the crucial Tampa region.

Meanwhile, in down-the-ballot polls:

  • Susan Collins leads 53% to 40% in a Research 2000 poll of Maine Senate race. 32% of Democrats cross-over to vote for Collins. She led by 19% in mid-September.
  • Mark Begich leads 48% to 46% in a Research 2000 poll of Alaska’s Senate race. A month ago, he led by 6%.
  • Ethan Berkowitz leads 50% to 44% in a Research 2000 poll of AK-AL. A month ago, he led by 14%.
  • In MN-06, an internal poll for the Bachmann campaign has her leading 44% to 33%. The poll was taken on the 12th and 13th. A DCCC poll taken last week had Bachmann leading by 4%.
  • In VA-02, Rep. Drake leads 47% to 42% in an internal poll for her Democratic challenger.
  • In CA-50, an internal poll for the Bilbray campaign has him leading 48% to 35%; an internal poll for the Leibham campaign has Bilbray leading 44% to 42%.

Senate: No surprises whatsoever among these four surveys: The Alaska race is entirely dependent on the outcome of Stevens’ trial (the jury should start to deliberate on Tuesday) while Hagan retains a slight edge. As for Maine, it is not looking good for Allen - just as we have known for months. It might make sense for the DSCC to invest in the race given how cheap the state’s media markets are, but the money could be put to better use with a bigger DSCC investment in Kentucky or Georgia.

House: Bachmann’s race is one many people are now interested in, and it confirms what we already know: Bachmann is slightly favored, but she is vulnerable. Depending on how much the DCCC invests, this race could certainly emerge as a hotspot, though Tinklenberg’s newly-raised half-a-million will certainly come in handy. In AK-AL, Berkowitz retains a lead but Young remains within striking distance, something other polls have also found. It is very much possible that a not guilty verdict for Stevens could also prove a boost for Young.


Rating changes, Senate edition: Make it eleven

October 12, 2008

Democrats have been trying to expand the Senate map since the start of the cycle, and as of mid-September it looked like the GOP had managed to stop the bleeding by solidifying its position in a number of races that Democrats were eying. But Senate Republicans are now back in their downward spiral, the prospect of several incumbents badly damaged by the economic crisis. For the first time in the long history of my Senate ratings, there are a stunning 11 Republican-held seats that are rated as highly vulnerable (meaning that they are rated lean retention or above), two more than in my previous rankings.

Five of those seats are rated lean retention, but they are not the five you might expect, as Alaska and North Carolina are trading seats in these rating update. While her race remains hotly contested, Sen. Dole has fallen slightly behind in recent weeks (and Republican operatives are even more worried than poll numbers seem to justify); on the other hand, Sen. Stevens is looking surprisingly strong, and it looks likely that he would survive this race if he is acquitted in the coming weeks.

Complicating calculations is the fact that highly competitive seats tend to break heavily towards one party, as the national political breeze of the final days tips the balance to one side. That allowed the GOP to nearly sweep the table in 2002 and 2004, while in 2006 Democrats won 6 out of the 7 seats that could have been described as highly contested in the final week of the campaign (MO, MT, NJ, MD, VA, RI and TN). That makes it unlikely that Democratic pick-ups are somewhere at the middle of their range, say 6 or 7. Far more likely are that Republicans slightly improve their situation and save most of the seats that are now rated as toss-ups or lean Republican (holding Democrats at 4 or 5 pick-ups) or that Democrats push their advantage and sweep most of the seats now in contention (getting them to 9 or 10 pick-ups). At the present moment, this ought to worry Republicans, for Democrats have all the momentum and the blue wave is threatening to transform itself into yet another tsunami.

The full ratings are available here. Here are today’s four rating changes:

Alaska, lean Democratic to toss-up: I acknowledge that this rating change is contrary to the conventional wisdom about this race, but there is no reason to categorize this seat as lean take-over when the result now looks to be almost entirely dependent on the decision of 12 Washington D.C. residents who are serving as the jury of Stevens’ trial. Mark Begich had pulled far ahead in July, but Stevens closed the gap over the past two months and seems to be convincing a sizable number of Alaskans that whether or not he is corrupt is irrelevant to the fact that he brings a lot of money to Alaska. The race is now a toss-up in the polls - the latest Ivan Moore survey has Begich leading by 4%, the latest Rasmussen has Stevens leading by 1%. If Stevens is found guilty, it should be enough to propel Begich over the top; but if Stevens is acquitted, it should him give him the last-minute boost he needs to come out on top. And as I explained yesterday, I am getting increasingly uncomfortable with the former scenario given how openly the prosecution has trampled defense rights in this trial.

Georgia, likely Republican to lean Republican: Who knew that Saxby Chambliss was this vulnerable? Late late spring, there were many other candidates to join the group of highly competitive Senate contests: Maine, Texas, Kansas, Nebraska, Oklahoma, Idaho… Democrats were too busy figuring out what was happening in a divisive primary in Georgia to think ahead to the general election. But some polls showed potential immediately after Jim Martin’s nomination - and that as before the financial crisis undercut Chambliss’s defenses along with those of Republican candidates nationwide. Chambliss retains an edge, but polls have shown that the race has dramatically  closed to the low single digits. Yet, the DSCC is not running ads on his behalf. That is certainly understandable (the DSCC is running ads in 9 states…), but unless the Democratic wind is really strong it could mean that Martin is unable to take full advantage of his opening. Martin’s success might very well depend on Obama’s coattails and on whether the Illinois Senator increases the share of the black vote in the state. Early voting indicators are very strong for Democrats, but Martin would have been in a far stronger position had Obama not pulled out of the state in mid-September.

Kentucky, likely Republican to lean Republican: Kentucky’s inclusion in the list of competitive seats is less surprising than Georgia’s. We have known that Mitch McConnell is highly vulnerable  since polls released in the fall of 2007 showed him barely beating a number of Democrats. Yet, many top-tier Democrats passed on the race and Bruce Lunsford simply did not seem a strong enough candidate to make it this tight (nor would he be a reliable enough Democrat for progressives to get excited over). But the race has been highly engaged for months, with both candidates running vicious ads (Lunsford has been particularly smart by contrasting McCain’s reformist image to McConnell’s insider status). And in the context of the financial crisis, it might no longer matter how good or bad a candidate is as long as he has a “D” next to his name; as the Senate Minority Leader, McConnell is that much more likely to suffer from voter anger about Washington and about Republicans. Over the past few weeks, polls have shown the race has closed, with Mason Dixon going as far as showing a tie; most other pollsters show McConnell up in single digits. The DSCC has just gone up on air against McConnell, further proof that Democrats are determined to bring down the Senate’s Republican leader.

North Carolina, toss-up to lean Democratic: This race is the biggest surprise of this year’s Senate cycle. While we certainly knew that Senator Dole was vulnerable, Democrats were certainly not expecting for the incumbent’s standing to collapse so easily. Dole led by double-digits through July, but a hard-hitting campaign by the DSCC took care of Dole’s numbers, as the race quickly became a toss-up by the end of August. Since then, Hagan has inched ahead in a number of polls, with her lead extending as much as a 9% in the latest PPP survey! To make matters worse for Dole, Hagan will benefit from Obama’s turnout machine while McCain’s ground game is minimal in the Tar Heel state. Somewhat surprisingly, Republican operatives are very pessimistic about this race, dropping quotes in a multitude of press outlets about how bad Dole’s situation is. Though there is no question that Dole’s campaign has been strikingly weak and its attakcs on Hagan have failed to catch on, polls do not justify such a high level of despair. The race remains highly competitive, and it could easily find its way back to the toss-up column in upcoming weeks. For now, however, Hagan is riding the DSCC advertisements and Obama’s momentum in the state, and that is proving a lethal combination.


Eighth presidential ratings: Blue and red states tighten, leading to narrowest projection yet

September 20, 2008

Since my last ratings, the presidential race has given heartburn to partisans of both parties. Thanks to the enthusiasm generated by the Palin pick and the momentum of his convention, McCain seized the lead in national polls for the first time of the general election, and with it came strong polling numbers at the state-level. Over the past week, however, Obama has erased McCain’s gains and taken a lead of his own, as the economy-fueled momentum has swung back his way. Now that these dynamics appear to have stabilized, it is time to take a look at the electoral college once more.

While the national numbers are now back to where they were mid-August, before the parties’ convention, there has been movement below the surface. In fact, a number of states appear to have tightened, whether they started blue or red. McCain has been gaining in some Obama-leaning states (Pennsylvania, Wisconsin, Minnesota, Washington and even New New Jersey all look more competitive than they did a month ago) while Obama has been gaining in some McCain-leaning states (he has closed the gap in Indiana, for instance, and looks more competitive in Florida than he did in August).

Because most of McCain’s vulnerable states were long known to be competitive, it is Obama who loses the most electoral votes in this rating round. As Pennsylvania moves to the toss-up column, Obama’s advantage over McCain is the smallest it has been since June 4th and the Illinois Senator now has to think about playing defense - something that was not on his campaign’s mind throughout the summer. In fact, only 6 electoral votes now separate Obama and McCain - the smallest advantage either candidate has ever enjoyed in the history of Campaign Diaries‘ presidential ratings.

There are now 18 states that are rated lean or toss-up, a testament to how unusually large the playing field remains less than seven weeks from the election and how unpredictable the final tally is likely to be. A small national uptick for Obama or for McCain could be enough to deliver most of the toss-up states to one candidate, so the fact that these projections remain so tight in no way guarantees that there won’t be an electoral college blowout.

Without further delay, here are the eighth electoral college ratings (states whose ratings have been changed are in bold). Remember that states that are in the “lean” category are considered to be very competitive and certain to be hotly contested, but it is possible to say that one candidate has a slight edge at this time.

  • Safe McCain: Alabama, Arkansas, Idaho, Kansas, Kentucky, Louisiana, Mississippi, Nebraska (at large + 3rd congressional district), Oklahoma, South Carolina, Utah, Tennessee, Texas, Wyoming (119 EVs)
  • Likely McCain: Alaska, Arizona, Georgia, Nebraska (1st and 2nd congressional districts), North Dakota, South Dakota, West Virginia (41 EVs)
  • Lean McCain: Florida, Missouri, Montana, North Carolina (56 EVs)
  • Toss-up: Colorado, Indiana, Michigan, Nevada, New Hampshire, Ohio, Pennsylvania, Virginia (100 EV)
  • Lean Obama: Iowa, Minnesota, New Mexico, Oregon, Washington, Wisconsin (49 EVs)
  • Likely Obama: Maine (at-large, 2nd district), New Jersey (19 EVs)
  • Safe Obama: California, Connecticut, DC, Delaware, Hawaii, Illinois, Maine’s 1st district, Maryland, Massachusetts, New York, Rhode Island, Vermont (154 EVs)

This gives us the following map and totals:

  • Safe + Likely Obama: 173 electoral votes
  • Safe + Likely + Lean Obama: 222
  • Toss-up: 100
  • Safe + Likely + Lean McCain: 216
  • Safe + Likely McCain: 160

I will naturally not attempt to provide an explanation for every single one of these ratings and will concentrate instead on those that have shifted over the past two weeks:

Alaska, lean McCain to likely McCain: Sarah Palin’s pick caused me to move Alaska out of the toss-up column in my seventh ratings, but I had left it in the lean column until we got more information. The past three weeks have left little very doubt that Palin’s presence on the GOP ticket has been enough for McCain to address whatever weakness he had in this traditionally red state. Polls have shown McCain-Palin opening up a huge lead in a state Obama had poured money in throughout the summer. This also means that we won’t have to wait until Alaska’s polls close on Election Day - and that’s a good thing for those of us on the East Coast.

Indiana, lean McCain to toss-up: McCain is making a huge gamble by staying out of the Hoosier state. Sure, with all things equal it would be unlikely for Obama to win Indiana while losing Ohio, but all things are certainly not equal when one campaign is pouring resources and organizing and the other is refusing to admit that the state is competitive. Now, polls are showing McCain’s lead is at best within the margin of error; in fact, a survey released by Selzer & Co (a very reliable polling survey) found the Illinois Senator leading this week. The Obama campaign has been building up an organization in the state ever since they turned their attention to the crucial May 6th primary, and that has fueled a massive registration drive: 500,000 new voters have registered since January 1st alone (that represents about 20% of the 2004 electorate). This dramatic expansion of the electorate partly explains how a state Bush won by 21% can now be competitive, and the McCain campaign might come to regret not having taken the warning signs seriously.

Pennsylvania, lean Obama to toss-up: Democrats were hoping that the Keystone state would be more reliably blue than it was in the past two cycles. Spring and early summer polls showed Obama up by comfortable margins, and in few states did Democrats post as significant registration gains as they did in Pennsylvania. But Republicans clearly think otherwise, perhaps fueled by the intuition that Pennsylvania is one of the states in which Obama’s race could hurt him the most. McCain spent $1.6 million advertising in Pennsylvania last week alone, making it (by far) the biggest investment by either campaign in any state. And McCain’s investment appears to be paying off, as recent polls have found that Obama’s lead has vanished. The Big Ten poll and Rasmussen found a tie, while Quinnipiac’s latest numbers show Obama falling from 12% to 3% in three months. Securing Pennsylvania would have allowed Obama to divert his time and resources to picking up Ohio and Florida, but Democrats can no longer afford to feel confident here.

Texas, likely McCain to safe McCain: Bush was no longer going to be on the ballot, the share of Hispanics continues to increase in the state, and a series of spring and summer polls showed McCain’s lead within single-digits. All of this opened the door for Obama to make a move in the country’s biggest Republican state, and one McCain could not afford to spend time or resources contesting. But it’s now mid-September, polls have worsened for Obama and the Democrat’s campaign is now telling its Texan volunteers to travel to New Mexico to help Obama’s campaign in that crucial state. Texas will surely become a bigger focus of presidential campaigns in upcoming cycles - but not in 2008.

Washington, likely Obama to lean Obama: A few months ago, Washington looked to be solidly anchored in the Obama column, but a wave of recent polls have shown McCain gaining. In fact, the race has tightened in a number of states in which Obama once looked relatively secure - those include Wisconsin, Minnesota and Pennsylvania. That said, while McCain is investing millions in those three states, he has stayed away from the Pacific Northwest, testifying to the fact that the GOP still doesn’t believe it can pick up Washington and making it a less competitive “lean Obama” state than the other ones in that category.

West Virginia, safe McCain to likely McCain: In the Democratic primaries, Obama’s most problematic region was the Appalachian - and ever since he got crushed in the Kentucky and West Virginia primaries, no one has given him much of a chance in those two states. But Democrats have been eying West Virginia as of late, with some speculation that the Obama campaign is thinking about moving in the state. It’s getting late, of course, and Obama is an unlikely candidate to reverse the Republican gains in this state that used to be reliably Democratic. But the past two polls show McCain’s lead well within single-digits, so perhaps Obama can find an opening in the closing weeks.

History of Campaign Diaries’ electoral ratings:

  • September 20th: +6 Obama (222 for Obama [154 safe, 19 likely, 49 lean] and 216 for McCain [119 safe, 41 likely, 56 lean])
  • August 31st: + 16 Obama (243 for Obama [154 safe, 29 likely, 60 lean] and 227 for McCain [93 safe, 75 likely, 59 lean])
  • August 20th: + 14 Obama (238 for Obama [151 safe, 32 likely, 55 lean] and 224 for McCain [90 safe, 75 likely, 59 lean])
  • July 30th: + 38 Obama (238 for Obama [151 safe, 42 likely, 45 lean] and 200 for McCain [90 safe, 75 likely, 35 lean])
  • July 16th: +28 Obama (255 for Obama [150 safe, 43 likely, 62 lean] and 227 for McCain [90 safe, 78 likely, 59 lean])
  • July 2rd: +11 Obama (238 for Obama [143 safe, 50 likely, 45 lean] and 227 for McCain [93 safe, 78 likely, 56 lean])
  • June 18th: +22 Obama (238 for Obama [86 safe, 97 likely, 55 lean] and 216 for McCain [87 safe, 87 likely, 42 lean])
  • June 4th: +20 McCain (207 for Obama [76 base, 107 likely, 24 lean] and 227 for McCain [97 safe, 77 likely, 53 lean])

Poll watch: Obama leads in IA, PA, MI while IN remains very tight; Dems lead in AK-AL and CO-04

September 19, 2008

Another day of strong polling results for Obama - this time at the state level. SUSA confirms that the Illinois Senator can feel more confident about Iowa than about many Kerry states, Marist finds larger leads than we have seen lately for Obama in the crucial states of Michigan and Pennsylvania (two states that are quasi-must wins for Obama) and two surveys from Indiana find the race within the margin of error. Who knew the Hoosier State would be polled so much?

What is fascinating about the Marist polls is that the surveys were taken over the week-end (thus before the financial crisis exploded) in Ohio and Pennsylvania, and at the beginning of this week in Michigan. The share of voters who say that they are most concerned about the economy is far greater in the Michigan poll (51%), which explains why Obama has such a large lead and confirms that the dominance of economic issues this week is helping fuel Obama’s comeback. Here’s the full roundup of today’s polls:

  • First, the trackings, where the movement is less uniform than it was yesterday: Obama gains one in Research 2000 (leads 49% to 42%) and in Gallup (leads 49% to 44%). Rasmussen doesn’t move (tied at 48%) and McCain gains 3% in Diego Hotline (but still trails 45% to 44%).
  • Obama leads 47% to 45% in a Marist poll of Ohio. The two are tied among registered voters. Those who say that the economy is the most important issue for them vote Obama by 14%. Obama gets 90% of Democrats. This poll was taken Thursday through Sunday.
  • Obama leads 49% to 44% in a Marist poll of Pennsylvania. The margin is 3% among registered voters. Obama gets 87% of Democrats and leads among independents. This poll was taken Thursday through Sunday.
  • Obama leads 52% to 43% in a Marist poll of Michigan. The margin is the same among registered voters. Obama gets 92% of Democrats, leads by 14% among those who say the economy is the most pressing issue. This poll was taken Tuesday and Wednesday, after the Wall Street collapse.
  • McCain leads 49% to 47% in a Rasmussen poll of Indiana. He led by 6% in August.
  • McCain leads 47% to 44% in an ARG poll of Indiana.
  • Obama leads 54% to 43% in a SUSA poll of Iowa. He gets 89% of Democrats and leads by 11% among independents. Among voters who are sure of their vote, he leads by 15%.
  • McCain leads 53% to 42% in an ARG poll of North Dakota.
  • Obama leads 50% to 44% in an ARG poll of Washington.
  • McCain leads Obama 64% to 31% in a SUSA poll of Alabama.
  • McCain leads 61% to 34% in an ARG poll of Oklahoma.

There is good news for McCain as well in this batch of surveys, most notably his strong margin in North Dakota (a state Obama has been contesting). A Rasmussen poll last week had found McCain jumping to a strong lead there after struggling through the summer. Republicans will also be satisfied to see that Obama is struggling in yet another poll from Washington - confirming that the Northwestern state is far less safe than people thought a few weeks ago.

Meanwhile, in down-ballot:

  • Betsy Markey leads Rep. Marilyn Musgrave 47% to 38% in a Grove Insight poll for Emily’s List of CO-04.
  • Andy Harris and Frank Kratovil are tied at 36% in a DCCC poll of MD-01.
  • Mark Begich leads Ted Stevens 50% to 44% in a Research 2000 poll of Alaska’s Senate race.
  • Susan Collins leads 55% to 42% in a Rasmussen poll of Maine’s Senate race.
  • Mitch Daniels leads Long Thompson 56% to 40% in a Rasmussen poll of Indiana’s gubernatorial race.
  • Daniels leads Long Thompson 46% to 42% in a Selzer poll of that same race.
  • Dino Rossi inches ahead 48% to 47% against Gregoire in a Strategic Vision poll of Washington’s gubernatorial race.
  • Lautenberg leads 49% to 42% in a Rasmussen poll of New Jersey’s Senate race.
  • Chambliss leads 52% to 33% in an internal poll conducted for his campaign in Georgia’s Senate race.
  • Inhofe leads 55% to 39% in a Rasmussen poll of Oklahoma’s Senate race.

The House races bring some excellent news for Democrats. Musgrave and Young are among the most vulnerable Republican incumbents, and those are not isolated polls. The CO-04 survey, for instance, confirms what SUSA found a few weeks ago. Democrats have been trying to kick Musgrave out for a few cycles, and it looks like this could be their year. As for MD-01, it has a very high percentage of undecideds, and in a heavily conservative district they are more likely to vote Republican. But it remains remarkable that Democrats are competitive in a district the GOP should be safe in.

As for the Senate races, Democrats will be satisfied that Begich is holding on to a lead, though the race is undoutedly much tighter than they would like it to be. There isn’t much else for the DSCC to get excited about here. Tom Allen, Bruce Lunsford, Jim Inhofe and Jim Martin are making little to no inroads in their respective Senate races, making it increasingly unlikely that Democrats will be able to contest more than the 9 races they have already put in play.


Morning polls: ARG releases wave of state surveys, PPP polls Virginia

September 17, 2008

American Research Group just released an unusually large collection of state polls. Though some of the most competitive battlegrounds (FL, PA, MI, NH, VA) are missing, this certainly gives us a good idea of the field of play heading in the final run. Overall, more swing states favor McCain (he narrowly leads in Colorado and Nevada, more comfortably in Ohio and by double-digits in North Carolina), but most numbers are well within the margin of error and Obama gets some good results as well (he leads in New Mexico and is very competitive in both Montana and West Virginia).

First, some other presidential polls that have been released since last night - including a new poll from Virginia:

  • Obama leads 47% to 45% in a national poll released this morning by Reuters/Zogby. This is a 7% shift in his favor since the August poll. The poll was taken from Thursday through Saturday. Both candidates get 89% of their party’s vote.
  • There is a tie at 45% in another national poll, released by AP Ipsos. The poll was taken Thursday through Monday and is a one point gain for Obama since last week’s survey that found McCain up 1.
  • Obama leads 48% to 46% in a PPP poll from Virginia. This is the 4th PPP poll in a row to find Obama leading by 2%. Obama gets 91% of Democrats but trails among independents by 17%.
  • Obama leads McCain 52% to 36% in a Field poll of California. Sarah Palin’s favorability rating is by far the worst of the four candidates.
  • Obama leads 55% to 42% in a Rasmussen poll of New York. McCain had 32% in August and 28% in July.

No surprises, nor anything particularly stunning in those surveys, though they confirm that the race has moved back to a dead heat nationally. Democrats will also be reassured by PPP’s Virginia poll, as McCain seems to have gained ground in other swing states (PA, OH, MN) but not Virginia. Now, on to ARG’s polls, starting with those from competitive states. All polls have a margin of error of 4%, and they have not all been taken at the same time:

  • McCain leads 50% to 44% in Ohio. The poll was taken the 10th to the 13th. Obama only gets 79% of the Democratic vote. (The partisan breakdown is much more Republican than most polls that have been released of late; SUSA’s poll last week had a 9% edge for Democrats but this one is equal.)
  • McCain leads 46% to 44% in Colorado. The poll was also taken the 10th to the 13th. There are more Republicans than Democrats, and Obama leads by 14% among Democrats.
  • McCain leads 49% to 46% in Nevada. Here again, more Republicans are polled than Democrats but Obama leads among independents. The poll was taken over the week-end.
  • Obama leads 51% to 44% in New Mexico. Democrats make up 51% of the sample (40% in 2004) and Obama leads among independents.
  • McCain leads 49% to 47% in Montana. Ron Paul was not included, and neither were Barr and Nader in a state in which third party candidates could make a difference. The poll was conducted early, the 7th to the 9th.
  • McCain leads 49% to 45% in West Virginia.
  • McCain leads 52% to 41% in North Carolina, a disappointing result for Obama who only gets 25% among white voters. The poll was conducted over this week-end.
  • McCain leads 50% to 45% in Missouri. The poll was conducted Thursday through Monday.
  • Obama leads 51% to 41% in Maine.
  • McCain leads 58% to 36% in Alabama, 55% to 39% in Alaska, 56% to 39% in Arizona, 68% to 25% in Idaho, 63% to 31% in Kansas, 57% to 37% in Kentucky, 50% to 43% in Louisiana, 57% to 36% in Texas, 65% to 29% in Utah and 66% to 28% in Wyoming.
  • Obama leads 82% to 13% in DC, 51% to 40% in Delaware, 63% to 32% in Hawaii, 51% to 45% in Illinois, 55% to 38% in New York, 59% to 33% in Rhode Island.

It is remarkable how few surprises there are in these polls, with most results - including those in Ohio, Colorado, Nevada and New Mexico - tracking the average of recent polls from these states. Colorado and Nevada confirm that they are among the ultimate toss-ups of this year’s contest. The least expected results are surely those from West Virginia (this is the second poll in a row to find a competitive race), Illinois (does anything think Obama has something to fear there) and North Carolina, where pollsters seem unable to find a coherent model and where numbers are all over the place - from a 2% race to a 20% race.

Obama will also be reassured by the Montana poll, as the only recent survey we had seen (from Rasmussen) had McCain surging to a lead in the aftermath of the convention. The question facing his campaign now is whether to invest in West Virginia, a state that had long been ruled out for Obama because of his problems in Appalachia. There have been rumblings of that as of late, but no sign for now that Democrats will move in there. [Update, and partial correction: As Ben points out in the comments section, Obama ads are running in many of the state's markets because of overlap with advertising in neighboring states. The same is true for McCain in New Jersey.]


More polls: Dead heats in key states, McCain leads in OH, Dem opening in KY-03 and CA-04

September 8, 2008

The two-week state polling embargo seems to finally be broken! After the day’s first polling wave showed McCain enjoying a bounce nationally but the race remaining stable in the key battlegrounds of Michigan, Virginia and Colorado, more surveys released by Rasmussen confirm that neither candidate is catching a clear break in swing states. All the Rasmussen polls were conducted exclusively on Sunday (one-day polling is generally frowned upon, especially on a week-end) and carry a relatively large margin of error of 4,5%:

  • In Colorado (polling history), Obama leads 49% to 46%. He leads by 10% among independents. Last month, McCain was up by 1%.
  • In Ohio (polling history), McCain has a solid 51% to 44% lead. Last month, McCain led by 5%. Obama has two problems: He only has 78% of the Democratic vote and he trails by 26% among independents! (Note that this is the third Rasmussen poll in a row to find McCain with a big lead in Ohio, something no other polling outlet has found.)
  • In Florida (polling history), it’s a tie at 48%. Obama has a big lead among independents, but he is here again weak among his base (79%). Last month, McCain led by 2%.
  • In Pennsylvania (polling history), Obama has a 2% lead, 47% to 45%. He led by 3% last month. Here again, Obama must solidify his Democratic base: he is only at 74%.
  • In Virginia (polling history), McCain is up 49% to 47%. That’s only a 1% improvement over August.
  • Finally, a last state poll came from SUSA in Washington. SUSA finds Obama’s lead collapsing to only 4% - down from 8% in August and 16% in July.
  • Finally, the Alaska poll commissioned by the NRSC that I blogged about earlier also contained presidential numbers, and confirms that McCain is now ahead, 55% to 34%. This also means that the NRSC’s New Hampshire poll is the only one for which presidential numbers were not released (contrary to Colorado and AK), further suggesting that the NH numbers might not have been good for McCain.

To recap: There is almost no movement in any of these states. Colorado moves by 4% in Obama’s direction, and Florida by 2%. Ohio moves by 2% in McCain’s direction, while Virginia and Pennsylvania move by 1%. All these margins (but Ohio’s) and trend lines are well within the margin of error. If anything, the most worrisome result for Obama comes from Washington, and that’s not a state in which McCain has any investment for now.

In other words, the electoral college situation seems largely unmoved after the two conventions, and there is much less movement in the key battlegrounds than in national polls. This is certainly not inexplicable: While veepstakes and convention coverage happens in a vacuum in states like Maryland or Texas (thus amplifying its impact), it is largely drowned by an army of volunteers and millions worth of ads in states like Ohio and Virginia.

Furthermore, keep in mind that if Obama keeps Michigan and Pennsylvania (and he has led in every survey from both states all summers) and lives out his advantage in Iowa and New Mexico (which are currently both leaning Democratic), any of the other four states polled by Rasmussen would be enough to put Obama over the top. That is not to say that Obama is in any way assured of victory, just to point out that the election is still being waged on red territory. That’s in some sense good for both candidates: The states McCain needs to win have leaned Republican in past elections, and Obama can concentrate on offense.

Meanwhile, we got two House polls:

  • In CA-04, an internal poll for the Charlie Brown campaign has him with a narrow lead against Republican McClintock, 43% to 41%. That’s obviously well within the margin of error.
  • In KY-03, a SUSA poll finds Rep. Yarmuth keeping a lead against former Rep. Anne Northup, 53% to 45%. Yarmuth led by 10% two months ago.

Neither district is considered to be in the top-tier of competitive House races, as both are currently rated lean retention. It would be nice to see independent polling in CA-04 to see how much of a shot Brown really has in this conservative district. In KY-03, Yarmuth might be the incumbent but Northup cannot be regarded as a complete challenger considering she was a longtime representative of the district and she mounted an unsuccesful gubernatorial run in 2007.


Poll watch: First signs of a McCain bounce, a third tracking poll (!), Indiana is competitive

September 5, 2008

The two week stretch of fast-paced game-changing events is finally over, and we now have to wait to see the state of the race as of the end of the veepstakes/convention cycle. Will there be much difference from where the race was as of August 22nd? Will Obama’s gain among Clinton supporters and McCain’s among evangelicals cancel out? Did undecided voters, independents move at all? Did Palin’s pick open the door for McCain among women?

In the wake up yesterday’s CBS poll showing McCain gaining 8% in three days, today’s trackings show an uptick for McCain (see below), but we will of course have to wait a few more days to see the convention’s full impact and look at the internals of major national polls to see (1) whether McCain improved his image and (2) whether Obama’s improvements from last week survived the RNC convention. Also, pollsters will now go back in the field to bring us more state-level presidential and Senate surveys. We have not been getting a lot of those lately.

Let’s look at the day’s presidential numbers:

  • The two tracking polls are starting to detect signs of a McCain bounce. In Rasmussen, Obama’s 5% lead shrank to 2%, 48% to 46%. In Gallup, Obama’s lead also shrinks by 3%, and he is now up 48% to 44%. A third of this tracking was conducted after the Palin speech, none was conducted after McCain’s. (You can also check this interesting attempt at guessing the individual night numbers.)
  • And we now have a third tracking poll! Diego-Hotline just released its first wave, taken over the same period and finding Obama leading 46% to 40%. (The previous Hotline survey, released a few days ago had Obama leading by 9%. but it was not part of the tracking series so I am not sure whether the methodology is the same and trendlines can be used.) Obama’s favorability rating is at 58%, McCain’s at 52%, Palin’s 43% and Biden’s 42%. Other interesting findings: Energy is the only issue voters are unsure who would handle best (Obama 43, McCain 41). More voters think Palin is prepared (46) than unprepared (45) and only 41% of women think she is “prepared.” This tracking will not
  • In Indiana, a Howey-Gauge poll conducted after the Democratic convention and the Palin pick has a tight race, with McCain leading within the margin of error, 45% to 43%.
  • In Alaska, the new Ivan Moore poll confirms that McCain has secured the three electoral votes thanks to Palin. He led by 3% two weeks ago; now, the GOP ticket dominates, 54% to 35%.

We will be closely monitoring how numbers in red state evolve post-Palin, as it is in states with a strong Republican base that McCain’s gains among conservatives could be the most useful. That said, Obama continues to air ads in Obama and McCain has still not moved in - and given that he is now restricted to public financing he is unlikely to invest there unless he really has to. This is the third poll in the past two weeks that show a tight race in Indiana.

Down-the-ballot:

  • In the Indiana gubernatorial race, that Howey-Gauge survey has Governor Daniels leading Long-Thompson 53% to 35%.
  • I already posted Ivan Moore’s Senate numbers from Alaska, which had Begichup 49% to 46%. The survey also tested possible match-ups in AK-AL. Berkowitz led Rep. Young 54% to 37%, but Parnell led Berkowitz 49% to 38% (the margin was only 4% three weeks ago).

This is the clearest exposition yet of the stakes of Alaska’s Republican primary. What is extraordinary is that the primary already took place… ten days ago! The margin was less than 200 votes, with thousands of absentee ballots left to be counted. The Anchorage Daily News (a very useful paper these days) says that the results might be released tonight, so we will soon know whether AK-AL should be considered likely take-over or lean retention.

If Parnell prevails, it would be a big disappointment for Democrats - as is Indiana’s gubernatorial race. This is a seat Dems have eyed for two years, and spring polls suggested the general election would be etremely tight. But Daniels’s approval ratings has been recovering and he has opened a solid lead in most polls released over the summer. At least, Democrats can be relieved that Daniels’s victory will not also cost them a Senate seat, since Barack Obama did not tap Evan Bayh.


Polls: McCain gains 8% in CBS poll, Obama takes first lead in red ND but loses opening in AK

September 4, 2008

37 million viewers tuned in to watch Sarah Palin last night, almost as much as the 38 million that watched presidential nominee Barack Obama last week even though his speech was transmitted on four more networks. Palin has become a huge story, and whatever voters thought about her speech will surely have repercussions.

CBS just released a new national poll, and it finds quite a sharp change in just a few days: Obama was leading by 8% in a poll taken over the week-end, but the race is now a tie in a poll taken from Monday to Wednesday, suggesting that whatever bounce Obama got faded quickly - and before Palin even took the stage.67% of Clinton supporters now back Obama (58% last week) but McCain has a similar uptick among evangelicals (57% to 66%). Obama’s lead among female voters is smaller, 43% to 38%. Note that the previous poll was taken after the Palin pick, so this is a post-post VP and pre-convention bounce for the Republican.

However, the Gallup and Rasmussen trackings are showing little movement in their latest polls taken over the same period. Gallup has Obama leading 49% to 42% (he led by 6% yesterday), and Rasmussen has Obama leading 50% to 45% (same as yesterday). At this point last week, Obama had already posted gains in the tracking, but keep in mind that the Denver convention had a Monday night, but the GOP did not.

And another national poll, taken by Democracy Corps (a Dem. pollster) yet again in the Monday-Wednesday period, has Obama leading 49% to 44%. As usual, the pollster put a heavy focus on the 18 battleground states Obama has invested in (12 of which were carried by Bush in 2004), and there Obama is ahead 49% to 43%. That’s an impressive lead, it can partly be explained by the fact that Obama has big leads in some of the Kerry battleground states (MN, WI) but it clearly goes to show that the election is being waged in Republican terrain and Obama is still doing well.

For now, the Palin effect seems to be increased polarization. McCain is gaining among Republicans and the base is very excited - Rush Limbaugh has dropped all his animosity to the ticket and is now calling McCain “McBrilliant.” But the Obama campaign is saying that they raised $8 million offline since yesterday’s primetime ended, an outstanding sum that I believe exceeds the amount McCain raised last week in the 18 hours after the Palin announcement, a clear suggestion that the intensity of Democratic partisanship also shot up last night. Also, keep in mind that the McCain campaign can no longer raise money, as it now has to limit itself to $84 million (note that Republicans can still donate to the RNC, which raised $1 million since yesterday).

Let’s get to the day’s state and district-level polling:

  • In North Dakota, Obama has taken his first lead ever in a DFM Research poll conducted just before Obama’s acceptance speech. He is up 43% to 40%, within the margin of error.
  • In Alaska, the last poll had Obama leading by 5% but the first post-Palin survey has McCain jumping to a huge lead, 57% to 33% with Palin’s approval rating at 86%. However, this is a poll that is being circulated by the McCain campaign (though the campaign claims it did not commission it); it is extremely rare for a presidential campaign to leak an internal poll, so this just goes to confirm how eager they are to highlight Palin’s appeal among voters who know her.
  • Democrats got some bad news in MO-09, a district they are contesting. SUSA found Republican Blaine Luetkemeyer leading Judy Baker 50% to 38%. Obama was crushed 60% to 36%, numbers that are much worse than Kerry’s in 2004.
  • In KY-02, however, a Democratic internal poll suggests an opportunity for Dems to take a very conservative district (even though the district’s partisan ID is 58% Democratic, it voted for Bush with 65%). The poll shows Democratic state Senator Boswell leading fellow state Senator Brett Guthrie 41% to 33%.

Alaska, Montana and North Dakota are the three extremely Republican 3-electoral vote states that the Obama campaign has been airing ads in over the past 2 months. In the first two, at least one poll had shown Obama taking a stunning lead some time in the summer (see Montana’s polling history and Alaska’s polling history). Now, Obama seems to be making progress in North Dakota. The prior two polls (taken by two different institutes) showed McCain only by 1% and 3%, but seeing a Democratic presidential candidate leading by any margin here is truly remarkable.

3 electoral votes might seem negligible compared to the high stakes of Ohio and Michigan, but keep in mind that if the election is close we are talking about very small margins in the electoral college. McCain can certainly not afford to give up on either North Dakota or Montana, and even if there is no other benefit to the Palin pick, it is significant that she anchored Alaska’s 3 votes in the Republican column. As I said, the last poll had Obama leading by 5%. While we will be looking for independent polling that does not come from the McCain campaign to confirm this survey, I had already moved Alaska out of the toss-up column in my latest “veepstakes special edition” presidential ratings.

MO-09 and KY-02 are comparable in the sense that they are both deeply conservative open seats in which Democrats usually stand no chance. Both have been polled over the summer, and it is interesting to compare the results. An independent (SUSA) poll from KY-02 in early July found Boswell up by 3%, suggesting that the internal poll is not completly out of line. An internal poll for Baker in MO-09, however, found her leading by 2% just two weeks ago. Losing these districts would be devastating for the GOP and would mean that Democrats have a shot at a big wave like 2006; we will know more about whether (or rather where) Democrats are the most competitive by looking at the party’s expenditures over the next 2 months.


Seventh electoral ratings: Veepstakes have little immediate impact

August 31, 2008

This is not Wednesday - the usual day of my presidential ratings - but considering that the week ahead is supposed to that of the Republican convention, we will have enough to talk about then. Consider this a special veepstakes edition. McCain and Obama had an opportunity to reshape the electoral map with their vice-presidential pick. For Democrats, choosing Kaine, Bayh, Richardson could have boosted his chances in key battleground states; for Republicans, Romney would have helped in Michigan and Pawlenty could have sufficed to transform Minnesota into a toss-up.

Instead, both candidates chose a pick from states with 3 electoral vote - one of which was considered a swing state (Alaska), the other was not (Delaware). Thus, the veepstakes’ only immediate impact is to move Alaska out of the toss-up column. Democrats would surely point out that Biden solidifies Obama’s claim on Pennsylvania (the Keystone State was already rated lean Obama in my previous ratings). The Obama campaign is playing up Biden’s roots in Scranton, PA (even airing an ad only in northeastern Pennsylvania) and is hoping Biden can help among blue-collar voters who voted for Clinton on April 22nd.

Beyond this home-state phenomenon, it is very much possible Palin and Biden’s impact will be felt in some states with more strength than others. If Biden helps Obama among blue-collar whites, that could be significant not only in PA but in places like Ohio and Virginia; if he helps him with older voters by adding gravitas to the ticket, look to Florida as a place numbers could move. Meanwhile, if Palin helps McCain among undecided women, that could be particularly important in the suburbs of Pennsylvania and Michigan; if she boosts his conservative credentials, Republican turnout in conservative regions of the Midwest and the parts of the South that are contested could increase; and if she makes him look more libertarian, she could prove a boost in Western states like Montana, North Dakota and even Colorado.

For now, most states are remaining in their place, and the latest polling supports this stability. The latest from Pennsylvania continue to show Obama with a consistent edge while Ohio and Virginia surveys are among the tightest in the country. As for Florida, McCain was ahead in seven of the eight polls released in August (Obama was leading by 1% in the eight). Stability also reigns in Colorado, where polls have stabilized and find both candidates in the lead after months of Obama holding a clear lead. Results in Michigan, New Hampshire and Minnesota, meanwhile, are finding Obama ahead but by widely differing margins.

I have moved New Mexico and Indiana’s ratings this month, however. Both states were very rarely polled, but we three and two polls from these states over the past two weeks. Combined with what we know about the two campaign’s ad investments and ground game, that is enough for a change.

Without further delay, here are the sixth electoral college ratings (states whose ratings have been changed are in bold). Remember that states that are in the “lean” category are still considered to be very competitive and certain to be hotly contested, but it is possible to say that one candidate has a slight edge at this time.

  • Safe McCain: Alabama, Arkansas, Idaho, Kansas, Kentucky, Louisiana, Mississippi, Nebraska (at large + 3rd congressional district), Oklahoma, South Carolina, Utah, Tennessee, West Virginia, Wyoming (90 EVs)
  • Likely McCain: Arizona, Georgia, Nebraska (1st and 2nd congressional districts), South Dakota, Texas (64 EVs)
  • Lean McCain: Alaska, Florida, Indiana, Montana, Missouri, North Carolina, North Dakota (73 EVs)
  • Toss-up: Colorado, Michigan, Nevada, New Hampshire, Ohio, Virginia (68 EV)
  • Lean Obama: Iowa, Minnesota, New Mexico, Oregon, Pennsylvania, Wisconsin (60 EVs)
  • Likely Obama: Maine (at-large, 2nd district), New Jersey, Washington (29 EVs)
  • Safe Obama: California, Connecticut, DC, Delaware, Hawaii, Illinois, Maine’s 1st district, Maryland, Massachusetts, New York, Rhode Island, Vermont (154 EVs)

This gives us the following map and totals:

  • Safe + Likely Obama: 183 electoral votes
  • Safe + Likely + Lean Obama: 243
  • Toss-up: 68
  • Safe + Likely + Lean McCain: 227
  • Safe + Likely McCain: 154

I will naturally not attempt to provide an explanation for every single one of these ratings and will concentrate instead on those that have shifted over the past two weeks:

Alaska, toss-up to lean McCain: To everyone’s surprise, the Obama campaign had included Alaska in its list of 18 targeted states and was airing ads in this traditionally Republican state. That was paying off in the polls, as Obama had taken the lead for the first time in a survey released mid-August; Obama trailed in low-to-mid single digits in most other surveys. But Sarah Palin joining the Republican ticket makes the state even more of an uphill battle than