Saturday, October 25, 2008
Obama's Largest Lead of the Race -- Getting Closer to Over
PROJECTED POPULAR VOTE: OBAMA/BIDEN 52.6%, MCCAIN/PALIN 45.4%
PROJECTED ELECTROAL VOTE: OBAMA/BIDEN 375, MCCAIN/PALIN 163
TIME UNTIL FIRST POLLS OPEN: 9 DAYS, 14 HOURS, 49 MINUTES
This race has never looked worse for John McCain or better for Barack Obama. While Obama's edge in popular vote projection, at 7.2% is slightly below his October 11th peak of 7.6% and his second October 18th peak of 7.5%, it is his third highest water mark of the race.
On the electoral map, Obama has opened up his largest lead yet. Here are the changes:
Indiana -- flips to Obama -- the most recent two polls show Obama up 4% (Survey USA) and 10% (Big 10 Battleground) respectively. I'm much more inclined to believe the Survey USA poll, I don't believe Obama could possibly be up 10% here and we have seen some funky recent results from Big 10.
Indiana -- remains a key battleground after the swap
North Carolina -- promoted from serious to key battleground -- a Rassmussen poll shows Mccain up by 2%, the first poll showing him ahead (although still within the margin of error) since October 7th. We leave it with Obama for now, but it may be getting closer.
Missouri, Nevada, Florida and Montana all remain as serious battlegrounds.
Georgia -- promoted from fringe to substantial battleground -- Insider Advantage poll shows Obama up 1%, the first poll to show him with a lead here. Rasmussen poll from the day before still shows Mccain with a 5% edge.
Ohio -- demoted from Substantial to Fringe -- two polls show huge leads for Obama: Big 10 at 12% and Insider Advantage at 10% -- other polls show more modest 3-4% lead
Colorado -- demoted from Substantial to Fringe -- Rocky Mountain News shows Obama with a 12% lead, McCain has not led in a poll here since September 14th.
West Virginia -- demoted from Serious to Fringe -- apparently the one poll a few weeks back that showed Obama up 8% was just an outlier, last 3 polls have Mccain +9%, +9%, +6% (CNN/Time, Rasmussen and Mason/Dixon respectively.)
New Hampshire -- added as fringe battleground -- welcome back, NH! After weeks of being out-of-sight and seemingly in the bag for Obama, the last 4 polls all show Mccain has cut it to single digits -- he still hasn't shown a lead in a poll since September 23rd, but his relentless campaigning here (against the advice of most of us) seems to be starting to pay some dividends
Virginia and North Dakota also remain as fringe battlegrounds
Need Some Polls, Please
To have an accurate projection, there are a few states where there hasn't been much polling that more data is needed.
North Dakota has only had two polls in the past two months, both showed a neck-and-neck race, but more information is needed to establish a clear leader.
Arizona hasn't had a poll in quite some time. Everyone assumes it is in the bag for Mccain, but all the old polls showing him up double digits are from a time when he was even-to-leading in the national polls.
Also would like some more recent polls in Iowa where Mccain has been pushing hard, not sure if he is making progress or not.
All The Things I've Been Wrong About
Since talk is cheap and apologies are precious, I thought I would take some time to reflect on all the mistakes I've made in my predictions this year:
1. Early on in the Democratic nominating process, I felt sure that Hillary Clinton would easily defeat Barack Obama (I'm sure she did too!)
2. I expected the Republicans to nominate a traditional conservative like Fred Thompson (before he fizzled out) rather than John Mccain
3. I said at the time that I thought Sarah Palin would be a boost to the Republican ticket -- clearly she has not been (although I may have been partly right because she did seem to provide an initial bump)
4. I thought the conventions were tied, but clearly Mccain came out ahead
5. When Mccain was a 2:1 dog on Intrade, I said I expected the odds to get closer by election day. They now stand at 7.4:1.
6. I thought Mccain won the third debate -- clearly he hasn't moved the polls
Things I've Been Right About (or at least stand by)
1. Joe Biden was a bad pick for Obama -- he has proved to be a liability at times and hardly ever a help
2. I thought Obama won debates #1 and #2 and I still believe he clearly did
3. I predicted that PUMA's (Party Unity My Ass Democrats) would be a non-issue and they have been
4. I thought that the controversy around the Democratic primaries in Michigan and Florida would be a non-issue by election time and I was right
5. I thought Colorado held the key to the election early on and it still may (although that will get lost if the race is a landslide)
On balance, I've probably been wrong more than I've been right this year. One more reason that my Presidential models are based on statistics rather than my own gut feelings. Presidential campaigns are unpredictable as we have seen throughout the year.
The only person who had this right so far was my wife, who is not a political junkie but after seeing Barack Obama's speech at the 2004 convention stated very directly "there's the next President." It seemed absurd at the time, a black guy with an Arab-sounding name who hadn't even won a senate seat yet. But here we are...
So it with a great grain of salt that you should view my analysis to come.
Early voting statistics are starting to emerge, so I thought I would give a rundown. It is mostly (but not entirely) favorable to Democrats:
Florida -- about 1.5 million ballots have been cast so far, representing about 20% of the likely electorate with an almost even split between Democrats and Republicans. Advantage McCain.
Georgia -- almost 1.0 million ballots have been cast so far, representing about 25% of the likely electorate with a huge advantage in both Democrats and African-Americans. Advantage Obama.
Iowa -- about 0.3 million ballots have been cast so far, representing about 15% of the likely electorate with a moderate advantage in Democratic turnout. Advantage neither as the trend has been Democratic but the numbers have been disappointing to the Obama camp.
Nevada -- almost 0.2 million ballots have been cast so far, representing about 25% of the likely electorate with a moderate Democratic edge in turnout. Advantage Obama.
New Mexico -- about 0.06 million ballots have been cast so far. Advantage neither -- this is only 7 to 8% of the electorate.
North Carolina -- about 1.0 million ballots have been cast so far, representing about 25% of the likely electorate with a heavy Democratic and African-American edge. Advantage Obama.
So, based on early voting turnout, Obama may run ahead of polling expectations in Georgia, Nevada and North Carolina, but may run behind expectations in Florida.
Is It Over?
Rather than definitively answer the question (I just proved earlier how bad I've been at predicting this year), I thought I'd present both sets of arguments.
Why It's Over:
1. Obama has his largest lead of the race with less than 10 days to go.
2. No candidate has come back from this far back with so little time in modern history.
3. Obama has all the advantages going in -- more money, a better ground operation and a by-and-large friendly media
4. The McCain campaign has looked completely disorganized and off message. McCain himself seems unable to mount a coherent attack.
5. The early voting numbers we just discussed are starting to "lock-in" the polls where they stand today
Why It Isn't Over:
1. A lot can happen in 10 news cycles.
2. McCain has a record of "from the dead" comebacks
3. The much heralded youth vote has a history of not showing up on election day
4. Obama consistently underperformed in some key states in the primaries: Pennsylvania, Ohio, New Hampshire. Call it the Bradley Effect or something else, but there is a pattern here that may re-emerge.
5. We don't know what information the McCain campaign has been waiting for the last week to release -- there could still be a bombshell. Remember Al Gore closing 5 points in the final few days of 2000? And that was just a DWI.
6. We have proven over and over again that polls don't vote, people do.
7. Will people really trust Obama on National Security when they walk into the privacy of the polling booth?
8. There is the possibility of mass chaos at the polls this year
Please Vote (!!)
Point number 6 on reasons why it isn't over is absolutely important. Whether you back McCain or Obama, get out and cast your vote, because we just don't know. Vote early if your state permits it, vote on election day if it doesn't. It's the most important thing you can do to protect our democracy.
State Polls vs. National Polls
With so much attention focused in recent days on the wide spread of the national polls, I thought I would attempt to validate the national numbers by comparing the state poll numbers, weighted based on the 2004 vote to the national numbers.
The state totals shown above would yield a 6.4% edge for Obama nationally. The national projection is 7.1%. Looks pretty right to me.
As a matter of disclosure, my final prediction map will be published on the evening of Monday, November 3rd. Some sites continue to update their projections on election day. I think this is silly -- you aren't predicting anymore when the event is already happening.
For the November 3rd prediction, I will be doing two things to ensure the accuracy of the results:
1. In states where I don't have recent poll data, older poll data will be "aged" based on the trend of the national polls since the state polls were taken
2. Once the data are aged, I will compare the national tracking polls to the composite of the state polls. The average of the two will form our final national vote projection. State projections will then be adjusted slightly to bring the totals in line with the national projection. States with the least polling data will get the biggest adjustment, states with the most polling data will get almost no adjustment.
Later in the week (probably that next weekend as I have to travel again on business starting the day after the election), I will compare my final projections against the Monday, November 3rd projections of other major respected sites such as electoral-vote.com, realclearpolitics.com, electionprojection.com and others. I am committed to accuracy and want to see how I do against these other sites.
Obama is back from visiting his sick grandmother in Hawaii and back out on the trail in Nevada today. He hits New Mexico for a rally this evening and then moves to Colorado all day tomorrow.
Biden is in Virginia today.
McCain is in New Mexico all day today. Tomorrow he hits Iowa briefly in the morning and then on to Ohio for the rest of the day.
Palin is in Iowa today and then Indiana this evening.
What became of the proudly announced Pennsylvania strategy for McCain? Is he now effectively conceding the state? If so, he needs to win all the rest of them: everything he is leading (including the battlegrounds of North Dakota, West Virginia, Georgia and Montana) PLUS North Carolina, Indiana, Florida, Missouri, Nevada, New Hampshire, Virginia AND Ohio. Tough climb.
My advice to McCain: focus on the big states. The top prizes are Florida, Ohio, North Carolina and Virginia. Stop spending time in New Mexico, which you will probably lose anyway. Go big and try to pick off one blue state (I guess Pennsylvania.)
My advice to Obama: Don't get greedy. Focus on what you need. Push hard in North Carolina and Virginia (you can easily hit both in single-day visits) and in Ohio and Pennsylvania (again, you can hit both in a day.) Shoring up Colorado is good, but ignore those 4 and 5 electoral vote battleground states. If you win EITHER Ohio OR North Carolina, it is over. Likewise, if you win Virginia and Colorado, it is over. Don't sweat Nevada, New Mexico, New Hampshire, etc. And don't bother with Indiana and West Virginia, they will probably go red in the final analysis.
That's all for this time...