Sunday, August 31, 2008

The Interestingness of Changelessness

The updated Gallup and Rasmussen tracking polls for August 30th showed no change in Obama's lead over McCain (at eight and four percentage points respectively). Because these polls were taken the day after Barack Obama's acceptance speech, one may conclude that Obama received no bounce from his well-received speech. Since the morning of August 30th was when Sarah Palin was named as McCain's running mate, one may also conclude the polling results suggest that McCain successfully stole Obama's thunder and counteracted the positive impact of Obama's speech. Further movement in these tracking polls towards McCain over the next several days would suggest the positive impact of Palin on McCain's fortunes.

My own model showed no change from the previous day. Obama's lead remains at 50.9% to 49.1% and his Electoral College vote margin remains at 304-234.

Saturday, August 30, 2008

A Bit More Bounce for Obama

Barack Obama continued to gain momentum according to the Gallup and Rasmussen tracking polls released on Friday and based on polling performed through the Thursday the 28th. The Gallup polls gives Obama an eight percentage point lead while the Rasmussen poll gives him a four percentage point lead. Imputting the result of these polls into my model produces a 1.8 percentage point lead for Obama, a one-day 1.4 percentage point swing:

This shift at the national level was enough to move Ohio, with 20 electoral votes, and Colorado, with nine votes, into the Democratic column. As a result, Obama now has a 34 vote cushion in the Electoral College:

The electoral map now looks like this:

John McCain Will Win

John McCain's choice of Sarah Palin was absolutely brilliant and, more likely than not, will deliver him the U.S. presidency in two months (and change) time. Here's why:

1. The boldness of the selection completely took the spotlight away from Barack Obama's acceptance speech and blunted its eventual impact.
2. By selecting a woman, McCain highlighted the fact that Obama passed over Hillary Clinton for an old white guy without seriously considering her. Had Obama selected Clinton, McCain would have gained little from selecting someone like Palin.
3. Because Palin is an evangelical Christian and stridently anti-abortion, McCain's pick will energize social conservatives who have thus far been sitting on the sidelines.
4. Palin successfully fought deeply entrenched political corruption in her own party. That suggests she's a tough woman and will be able to handle the pressure that will be thrown at her. She is being underestimated by the Democrats.
5. McCain's selection highlights the fact that he is the bold candidate in this election. Obama has in fact been revealed, by his selection of Biden, to be cautious and conservative.

The reaction of Democrats to McCain's pick is predictable; afterall, that's what they get paid to do and say. But if she turns out to be what McCain obviously thinks she is capable of, then that talk will be quieted.

Personally, McCain's pick of Palin probably seals the deal for me. I'm probably going to cast my vote for Obama-Biden in my home state of Pennsylvania. But I think the Democrats are, by foolishly selecting Barack Obama over Hillary Clinton and then by Obama foolishly deciding to pass over Clinton for Joe Biden, now at serious risk of losing an election that should have been their's almost by default.

Friday, August 29, 2008

A Barack Bounce

Finally, the polls have indicated the beginning of a bounce towards Barack Obama. The Wednesday tracking polling done by Gallup indicated a substantial one-day six percentage point lurch towards Obama. However, in sharp contrast, the Rasmussen tracking poll actually shifted one percentage point towards McCain on the same day. This just shows you how volatile polling can be. The best way to deal with this variation across polls is to average the results. When we do this and crunch the numbers through the model, we come up with a 0.5 percentage point shift towards Obama during the previous 24 hours:

The shift back over to Obama in the popular vote has also put him back on top in the Electoral College:


Obama's Electoral College vote lead is, of course, fragile since it is based on extremely slender margins in New Mexico and New Hampshire. The updated electoral map is shown below:


McCain's lead in Ohio appears to be extremely slender and his lead in Colorado is only slightly less fragile. My expection is that Obama's lead will jump in response to his well-received acceptance speech on Thursday evening. We should see that effect in the polls in two days time.

Thursday, August 28, 2008

A Day Two Bounce

Barack Obama experienced a bit of a bounce in the polls during the Day Two polling as both the Rasmussen and Gallup tracking polls moved one percentage point in Obama's direction. As I discussed in yesterday's post (as noted by Nate Silver of the blog 538.com), today's tracking polls likely reflected the impact of Day One of the convention. Five state polls also released yesterday also suggested a movement in Obama's direction, although the results in several states, including Now Mexico and Nevada, suggested a leap towards Obama that seems more than a bit questionable.

The results moved my own model a tick in Obama's direction, however, McCain remains on top according to the model by a 50.3%-49.7% margin. The Electoral College margin remains unchanged at 272-266 in McCain's favor. As has been the case over the past week or so, New Mexico and New Hampshire remain the most pivotable of the swing states in this election. Given that Obama was officially nominated today and gives his acceptance speech in the evening in the states, we should see some definite movement towards Obama in these polls over the next 48 hours.

The window of opportunity for Obama is narrow because John McCain will likely be naming his running mate on Friday. After that, the Republicans begin their convention in St. Paul, Minnesota early next week meaning that all of the attention will be on the Republicans shortly after the crowd files out of Mile High Stadium Thursday evening. Barack Obama will need to wow America this evening and convince them that he is a risk worth taking on.

Wednesday, August 27, 2008

The Biden Bounce . . .

appears to be somewhat underwhelming if two tracking polls from respected polling organizations (Gallup and Rasmussen) released over the past 24 hours are to be believed. In fact, feeding the results from these polls into my model suggests that McCain now has a modest popular vote lead over Obama of about 0.6 percentage points:

National Popular Vote Breakdown
As Nate Silver noted in the election blog 538, these two tracking polls finished their polling prior to the Monday evening speeches at the Democratic convention, so this movement should have nothing to do with the convention itself yet. But, as he notes, these polls should be showing the first reaction to the naming of Biden as Obama's running mate. It's not pretty so far, but this may be early and transitory.

McCain's lead translates into a narrow lead in the Electoral College according to my model:

Electoral College Vote Breakdown
The resulting current state-level map is shown below:

Current Electoral Map
New Mexico and New Hampshire have gone red according to the model. My model also shows that Missouri is getting uncomfortably close and that Obama's lead in Iowa may be under 2 percentage points.

Tuesday, August 26, 2008

Demo Convention Day One . . .

. . . . and no significant change yet in the numbers. The Rasmussen moving average tracking poll has shown a slight shift towards Obama the last two days (one percentage point a day) while the Gallup tracking poll has shown no movement whatsoever. The conclusion is that there is no clear budging from the 50%-50% point we have been stuck in for most of the past week. My model shows Obama with a 13,000 vote lead nationally out of an estimated 125 million voters right now. The coverage from the convention seems to be painting a positive picture of the goings on and the Clinton's appear to be cooperative at Obama's big party. Now we'll see if the numbers begin to move, as they really should.

Sitting here waiting for things to get unstuck . . .

Monday, August 25, 2008

The Demo Convention Begins . . .

with the election race as tight as it could possibly be. A number of polls released over the weekend taken in Rocky Mountain states painted a clear picture of an election looking more and more like a rerun of the last two presidential races. A Utah poll showed McCain with a lopsided 62% to 23% lead while a poll taken in neighboring Wyoming showed McCain running and hiding by a 62% to 25% margin. Assuming that the undecideds in these two states break for the two candidates in the same percentage shares, both states are presently within a mere percentage point or two of the percentage share breakdowns in 2000 and 2004. Now that Barack Obama has named another liberal northern senator to be his running mate, it is hard to imagine that the Democratic ticket will have any appeal whatsoever for states such as these.

Recent polling from several southern states shows that the race in much of the south has the same deep red texture that has been the norm for most elections since 1972. So much for Obama turning the deep south purple. Indeed, a poll released late last week taken in Mississippi shows McCain up by an 82% to 13% margin among Whites while Obama has a 97% to 2% advantage among Blacks. Such racially based margins would count as tribal warfare in more than a few unstable parts of the world. While there is an expectation among the Obama camp that a large Black turnout in the south could deliver states like Virginia and North Carolina into the Blue column, such racially based voting is as likely to cause poorer Whites in these states to turn out in larger numbers as well.

As for the race itself, my model is showing the race to be 50.0%-50.0%. The model is showing that Obama remains ahead with a paper thin 272 to 263 electoral vote margin but this paper thin margin overall is supported by paper thin margins in New Mexico and New Hampshire. My model is showing Obama with less than a 1,000 vote margin in either state which truly rates as an absolute tossup.

Sunday, August 24, 2008

Biden's Impact

While the world waited impatiently for the actual text message of Obama's VP selection to be sent out, word of Joe Biden's selection leaked out several hours in advance. The official announcement was in the end a non-event. So much for trying to appear hip and modern.

Obama's selection of Biden represented as safe a call as Obama could possibly have made. Biden is an eastern Pennsylvania/Delaware kind of guy, middle class, and ethnic Catholic (Irish descent) to boot. For goodness sakes, even today he commutes from Washington to Wilmington everyday by train. For this middle class ethnic Catholic with roots in Michigan and with time spent in eastern Pennsylvania and a firm believer in mass transit, Joe Biden is my kind of guy. Having Joe Biden on the ticket makes me a lot more comfortable somehow about pulling the lever for Obama.

And that's the point. Obama has picked Biden to solidify his base with liberals, particularly in the northeast. So much for adventure and so much for the 50-state strategy. With news that Obama has pulled ads out of a number of red states previously trumpeted to be in play, Obama has admitted even before the convention that the shape of this election is more likely to look a lot like the past two elections rather than a landscape of blue and purple. With this pick Obama has written off large swathes of the south and mountain west and instead will be focusing on the handful of states that are actually in play including Virginia, Ohio, New Hampshire, New Mexico, Missouri, Colorado and Nevada.

Biden's selection will enable McCain in turn to solidify his base. Expect a safe conservative pick such as Mitt Romney. So much for Change.

Saturday, August 23, 2008

No Room For Error - The Presidential Race on August 23rd

Another batch of polls over the past 24 hours show a renewed tightening of the presidential sweepstakes. Basically, it can hardly get tighter than this:

National Two Party Vote Shares
I estimate that Barack Obama's lead in the popular vote, were the election to be held today, at only about 250,000 votes out of a likely 125,000,000 to be cast for the two main presidential candidates. The tightening national vote appears to have moved Ohio into McCain's column:

Ohio Vote Shares
The flipping of Ohio from Barach Obama to John McCain has reduced Obama's Electoral College vote margin to only 12 electoral votes:

Electoral College Votes
The electoral map now looks like this:

It's too early to tell how the (apparent as of this moment) selection of Joe Biden to be Obama's Vice Presidential nominee will impact the presidential vote but I would anticipate that we are going to have a good deal of movement between now and the end of the Democratic convention late next week. That movement should be back towards Obama as he dominates the national stage during the Democratic convention and more of the electorate begins to pay closer attention to the election campaign.

Friday, August 22, 2008

Obama Gets a Breather

John McCain's recent "surge" that put has him within whispering distance of Barack Obama has apparently crested over the past 24 hours as Obama's popular vote lead has increased (according to my model) from 0.8 percentage points to 1.0 percentage points:

National TwoParty Popular Vote Shares
It will be interesting to see how McCain's "House-gate" statement plays out with the general public. Obama is attempting to take advantage of what he perceives to be a significant gaffle clearly hoping to add to his expected convention bounce next week. We should get Obama's vice-presidential nominee within the next 12 to 24 hours, and the polling should begin to reflect the effect of that choice (if any) over the next several days.

As has been the case for some time, the popular vote movement has done nothing to change the Electoral College vote standings that have been in place for some time. Obama continues to lead in the Electoral College by 295 to 243 votes.

Moving on to actual states where the election will really take place, a poll released yesterday by Rasmussen gave Obama a four percentage point lead in the swing state of New Mexico. My own model suggests that Obama's current lead in the state is 1.0 percentage points:

New Mexico Vote Shares and Popular Vote Margin

Thursday, August 21, 2008

It's Tied (More or Less)

Another day, another set of polls and yet more momentum towards John McCain. My models suggest that McCain slashed the gap between he and Barack Obama by 0.8 percentage points:

National Two Party Popular Vote Shares
The Democratic convention in Denver won't come a day too soon. While the Electoral College vote breakdown remains unchanged (at 295 for Obama and 243 for McCain) a sliver more movement towards McCain will put Ohio in his camp:

Ohio - McCain vs. Obama
The model sugggests that Obama's lead in Ohio is currently less than 25,000 votes. New Hampshire and New Mexico are also razor close with about 1 percentage point separating Obama from McCain. Both of these states would need to shift to McCain before he would take over the lead in the Electoral College.

While those states with the close contests usually get all the attention, I thought it would be interesting to show how the election is shaping up in those states somewhat less competitive. The most lopsided Republican state appears, once again, to be Utah where McCain's lead is nearly 2 to 1 over Obama:

Utah Two Party Vote Shares
If McCain names Mitt Romney as his running mate, he can count on piling perhaps another 5 percentage points or so on to his total in this competitive and vitally important state. Any bets?

Wednesday, August 20, 2008

Too Close to Call

A slew of national and state polls released over the past 24 hours paint a picture of a presidential election race that has narrowed to the point where it is too close to call. My election forecasting model suggests that at the national level McCain has pulled to within 1.6 percentage points:

National Two Party Popular Vote Shares
In spite of the narrowing spread at the national level, the Electoral College vote remains unchanged:
Electoral College Vote Totals
As is often the case, Ohio appears to be the state on the knife edge. I have Obama up in Ohio by only 1 percentage point:
Ohio Two Party Vote Shares
Should McCain pull ahead in Ohio, its 20 electoral votes would give him 263 electoral votes, only 7 short of the 270 needed for election.

A poll taken in Indiana gave McCain a 6 percentage point lead in the state. This is a state into which that, according to press reports, the Obama campaign has poured a significant amount of advertising and organizational resources. The conventional wisdom has been that Indiana has the potential of falling into the Democratic column this year for the first time since 1964 (when Barach Obama was 2 years old). Indiana may yet vote Democratic but only if there is a landslide at the national level, and that doesn't seem to be in the cards at this point. My own model shows McCain with a larger lead than this:

Indiana Two Party Vote Shares
Obama's campaign has been pouring campaign resources into southern states such as North Carolina and Georgia with relatively little to show for it as well. Obama's 50-state campaign is beginning to look like Richard Nixon's 50-state campaign in 1960. Nixon ended up devoting far too many resources in too many states that he was destined to either win or lose easily; in the end, he lost the election by 100,000 popular votes and by slender margins in Illinois and Texas to lose the electoral college by 303 to 219 electoral votes.

Tuesday, August 19, 2008

The Race Narrows Slightly

A slew of polls, two national and five state, were released today. The state polls included Georgia, Ohio, New York, Illinois and Tennessee only one of which, Ohio, is a swing state. Since it had been more than a month since polls were released in Illinois and Tennessee, the addition of these polls makes the overall results just a bit more robust. I am now working on introducing the results of the national level polls into the model. The results so far suggest that the state level models are quite robust; introducing the national level polls into the model will enable the model to adjust more quickly to significant short-term movements in sentiment at the national level. The model suggests that the race between McCain and Obama tightened fractionally over the past 24 hours:

National Popular Vote Shares
Obama's lead now stands at 2.4 percentage points. The Electoral College vote breakdown remains unchanged with Obama holding on to a 25 vote cushion over the 270 votes required for victory:

Electoral College Standings
In our continuing series on the seven swing states in this election, we look today at the race in Nevada. The state has five electoral votes in the Electoral College. My model suggests that McCain is leading this traditionally Republican leaning western state by a 3.6 percentage point margin.

Swing States - Nevada

Monday, August 18, 2008

Presidential Election Standings August 18th

Since no new polls were released during the past 24 hours, there's nothing new to update at the national level. I will take the opportunity however to note the current standings between Obama and McCain in the swing state of Missouri:

Missouri Two Party Vote Share

Obama has a 2.6 percentage point lead in Missouri according to my model. This does contrast with most polling which suggests McCain has anywhere from a 1 to 6 point lead in the state. If these polls are correct, then Obama is running behind where he should be in Missouri based on polling in bellweather states like Ohio, Pennsylvania and Oregon. And, if that is the case, then Senator Obama has a problem.

Sunday, August 17, 2008

Presidential Election Standings August 17th

Based on my model for each of the 50 U.S. states and D.C., Obama currently leads McCain in the overall popular vote by 2.6 percentage points:

Two Party Popular Vote Share


Obama's lead in the Electoral College is somewhat more substantial:

Electoral College Vote Totals


Seven states appear to be crucial swing states that will decide the election if it remains close on election day. These include Colorado, Missouri, Nevada, New Hampshire, New Mexico, Ohio and Virginia. Colorado, with 9 electoral votes, appears at present to be in McCain's column. I estimate his current lead in Colorado at 2.6 percentage points:

Colorado Vote Share


The map below shows the current electoral map as of today based on my model:

Friday, August 15, 2008

Tamino's Bet - Months Six and Seven

This post is the sixth in a series of posts on a "bet" made by Tamino in January of this year. Tamino's position is that the trend of increasing average temperatures at the earth's surface will continue into the forseeable future. In order to test his assertion that global temperatures would continue to increase between now and 2015 in line with the 1975 to 2007 trend he has set conditions for each year between now and 2015 to confirm or deny the assertion. The conditions of the bet for the year 2008 are as follows:

1. If the average global temperature anomaly, as measured by NASA GISS, equals or exceeds .7350 degrees Celsius, the "still-warming" side will receive one point.
2. If the average global temperature anomaly, as measured by NASA GISS, is less than or equal to .4035 degrees Celsius, the "not-warming" side will receive one point.
3. If the average global temperature anomaly, as measured by NASA GISS, falls between these two figures, both sides will receive zero points in 2008.

Although the GISS global temperature anomaly was .51 degrees in July, the average NASA GISS global temperature anomaly for 2008 through July was .3686 degrees Celsius or .0349 degrees Celsius below the Tamino threshold for the "not-warming" side to receive one point in 2008. In order for the "not-warming" side to win a point in 2008, the average anomaly for the reminder of the year would need to be below .4524 degrees. My own forecast for the remainder of the year is for the global temperature anomaly to average .5837 degrees. Thus, I don't expect that the "not-warming" side will win the point in 2008. However, things are getting very, very interesting with this bet.

Tuesday, August 12, 2008

It's a Tight Race

I've developed a method to estimate the most likely current state-by-state two-party percentage distribution in the upcoming presidential election. The method takes advantage the fact that the two-party distribution of votes in any given state are correlated with the distribution of votes in number of other states. This basic fact means that polling data in, say, California, may be reasonably used to estimate the current distribution of votes in states as geographically disparate as Washington, Vermont, New Mexico and Illinois. This approach enables a relatively large number of polling samples to be estimated for each state and, as long as the polls themselves are relatively unbiased within and across states, the resulting estimations should give a good picture of the actual state of the election in each state.

I have used data from the the 1964 to 2004 presidential elections to calculate correlations between each state. I have chosen the 1964 election as the starting point for the data because, arguably, this election represents the first "modern-era" election. That is, the pattern of vote distribution across states established in that election are remarkably similar to the pattern of vote distribution across states in most elections since then. I have used a correlation cutoff of 0.8 to determine whether the polling data of a state would be used to estimate the percentage distribution of votes in the state under consideration. I have use the RealClearPolitics website as my source for state level polling data. I have used data from polls conducted over the previous 30 days and for each state have simply averaged the polling results across all polls taken in a given state (that is, I assume that each poll lacks systematic bias that would, uncorrected, render the model invalid). Since these averages in nearly all cases add to less than 100%, I then adjusted the average figures to derive each major party candidate's share of 100% of the two-party vote in a given state. These percentage shares then became the base percentage shares for that state in the model. Note that where no polling data was available for a given state in the past 30 days, no prior polling data from the state was used (even if available) to estimate the two-party percentage shares in other states. Because recent polling data is available for most states, the lack of recent polling data in some states did not have a major impact on estimating the candidates most-likely share of the two-party vote in any state (in other words, the model is reasonably robust if polling data from some states are not available).

The percentage shares for each state were calculated in a three-step process:

1. Regression models were estimated for each state against all other states. In the case of California, for example, regression models were calculated between California and each other state (California-Alabama; California-Alaska; etc.). These regression models would allow a poll in, say, Alaska, to be used to estimate the two-party distribution of votes in California.
2. The polling data was used to estimate the two-party distribution of votes in each state. since for each state the results of polls in a number of other states were used, each of these results were in effect separate samples. The average across all samples was then taken as the preliminary two-party distribution of votes in a given state.
3. The preliminary two-party distribution of votes in each state generated in step 2 above was then used to replace the initial polling data. The model was rerun for each state and the resulting figures were then taken as the final estimation for each state.

Note that I did make further adjustments in a total of five states. In the home states of the two candidates (Arizona for McCain and Illinois for Obama) I added five percentage points to the home state candidate to reflect their likely "home field" advantage. I also added five percentage points to McCain in the Appalachian states of West Virginia, Kentucky and Tennessee to reflect a likely resistance to Obama's candidacy from voters that might otherwise vote for a Democratic candidate in those states.

The following maps below show the model's results as of August 12, 2008. The red states are states where McCain is presently estimated to be ahead and the percentages in those states (in black) are McCain's estimated percentage of the votes in those states. The blue states are states where Obama is presently estimated to be ahead and the percentages in those states (in yellow) are Obama's percentage o the votes in those states. A handful of the smaller northeastern states are not listed in the map and are instead shown in the northeastern closeup map below:

The map below gives a closeup for the smaller northeastern states:

For those readers enjoying detail, the chart below shows the estimated two-party popular vote in each state, the percentage shares and the allocated electoral votes (please click on the table for a closeup view):

In the all-important electoral college, Barack Obama leads John McCain by a slender 295 to 243 vote margin. This electoral college vote figure is nearly the same as the estimated figure given in the 538.com website as of today. My model also suggests that Obama has a 2.2 percentage lead in the popular vote, just fractionally over the 2.0 percentage point Obama lead noted in the 538.com website. My own work in setting up this model suggests that Obama's lead has been sublimating away slowly over the past several weeks.

There's not much more lead to be lost before John McCain is in the lead. The media continue to treat Barack Obama as the presumptive president-to-be but the data seems to suggest that this presumption is very much misplaced. Obama's lead in Ohio appears to be only about 1.4 percentage points (or about 80,000 votes) while his lead in Missouri appears to be only about 0.8 percentage points (or about 23,000 votes). Their combined 31 electoral votes would give McCain 274 electoral votes and the lead. On the other hand, McCain now appears to have a 2 percentage point cushion in Colorado and a more than 3 percentage point margin in Nevada.