Can John McCain win? Lots of folks think it is over. Nate Silver calculates today on his fivethirtyeight.com site that John McCain has a 3% chance of winning. Realclearpolitics and Rasmussen put his chances a bit higher, just above 12%. Many pundits have been arguing the contest has been over for a week, and some like Jonathan Alter, believe only a big and so far unseen racist swing vote can bring victory to McCain. Clearly, the idea that voters might reject Barack Obama for reasons other than race is not conceptually possible to someone so in the tank for Obama as Alter.
I think McCain can still win. But it will not be easy, and Obama remains a heavy favorite. In essence, things have changed very little since my last update on the race almost three weeks ago. Barack Obama is in good shape in all the blue states won by John Kerry in 2004. In only four of them, is his lead less than 10%. Obama's lead in New Hampshire (4) hovers around 5%. In Minnesota (10), Wisconsin (10) and Pennsylvania (21), Obama's lead is between 5% and 10%. The national tracking polls, which have tightened the last two days to about a 5% lead for Obama on average, would need to move five points more to put New Hampshire in play, and probably more than that to put any of the others in play. Campaigning in a state and ads can move the state number 1-2% beyond the national tracking trend, but likely not more than that in the last week, especially with both sides contesting every important state. Obama has much more money to advertise the last week, a good position to be in if you are running out the clock with a lead.
Obama overpolled and underperformed in New Hampshire in the Democratic primary in January, and McCain has a strong residue of support in the state, so that state has the potential to be a surprise on election night. The almost astonishing stupidity of John Murtha should cost him his House seat, and has also made Pennsylvania a bit more competitive that it seemed a week ago. Murtha is certainly argument number one for term limits in the House of Representatives. Jonathan Alter is allowed to call voters racist (he is a liberal pundit after all), but for Murtha to call his own district's voters racist defies comprehension.
McCain may have to win without a blue state, or at best, just New Hampshire. Is this possible? Two red states seem out of his reach -- Iowa (7), and New Mexico (5) are in the 5-10% Obama margin category, if not more. I am not sure why the McCain team has spent so much time in these states, unless they have internal polls that show the race is far closer than other surveys. If these two states go for Obama, he would have a base of 264 Electoral College votes. He would then need to turn but one more red state from among Colorado (9), Virginia (13), Nevada (5), Ohio (20), Florida (27), North Carolina (15), Missouri (11), and Indiana (11). While Nevada would bring him only to a 269-269 tie, the Democrats' edge in the House would give Obama the Presidency. The same result would occur were a tie achieved by Obama winning Iowa, New Mexico and Colorado and losing New Hampshire. Neither of these two scenarios is that far fetched, if McCain can gain 4-5% in the national polls in the remaining week.
The Obama team regards four red states as part of their firewall: Iowa, New Mexico, Colorado and Virginia . If Obama wins these four, he wins 286-252, the identical Electoral College margin by which Bush beat John Kerry in 2004. Rasmussen surveys on Virginia and Colorado released Monday night, have Obama up 4% in each of them. That is good news for McCain -- 4% is not a firewall lead. If national polls move 5 points to a 50-50 race, then both states could shift. Many surveys show bigger Obama leads in both states (in the 5-10% range), but analysts with long term familiarity with Virginia politics, such as Larry Sabato, are skeptical that Obama has such a big lead in that state. I think Colorado is in better shape for Obama than Virginia at the moment.
The next category of red states that Obama could pick off are Nevada, Florida and Ohio. Obama currently has small leads (probably 2-5%) in all of them, with Florida probably closest of the three. If the national polls stay where they are and prove accurate (and Obama wins by 5%), he will probably win all three.
The next category of Obama targets include Missouri, North Carolina and Indiana. McCain seems to be ahead in Indiana, and about even in the other two states. If I had to rate them, I think Obama is doing best in Missouri and worst in Indiana of the three. If the national vote winds up as it shows today, a 5% Obama margin, I think he wins Missouri, and loses the other two states. Summing this all up, with his current 5% national lead, Obama wins the Electoral College 349-189. But if the national vote gets to even, McCain has a shot at a 274-264 win. The problem for McCain is that with a 50-50 national vote, many of the red states above will be very closely decided. Obama has a decided manpower advantage in the field, and a better voter turnout operation. This could prove decisive in one or more of the tossup red states in that case, and earn him a small margin of victory. And again, Obama only needs to turn one of the 8 red states to win.
If I had to rate Obama's red state prospects from best to worst I would put them in the following order: Iowa, New Mexico, Colorado, Virginia, Nevada, Ohio, Florida, Missouri, North Carolina and Indiana. The first two I believe Obama will win, even if the national numbers move 5%. The other 8 are not so certain. McCain's blue state prospects in order would be: New Hampshire, Minnesota, Wisconsin and Pennsylvania. New Hampshire would fall in with Colorado and Virginia in terms of the degree of difficulty of winning the state for McCain. The other three blue states would be harder for McCain to win than any of the 8 red states that are likely still in play. I think McCain's high water mark in the Electoral College is 278- losing New Mexico and Iowa and winning New Hampshire.
The attention the McCain camp is paying to Pennsylvania suggests they think there is movement there, and they have a chance. But there is one other reason they may be working the state so hard -- with 21 Electoral College votes, Pennsylvania can make up for losing both Virginia and Colorado, giving McCain a 273-265 victory. . Of course, McCain would still need to win Nevada, Ohio, Florida, Missouri, North Carolina and Indiana, or he could lose Nevada, but only if he won New Hampshire (gaining a 272-266 win) .
There is no doubt that there are many paths to victory for Obama and just a few for McCain. But if undecideds break for McCain, and the last two days' momentum continues, he remains in the game. The release Monday of the 2001 audio tape with Obama arguing for wealth redistribution seems to suggest that the foolish remark he made to Joe the Plumber was not an aberration. "Spreading it around" matters more to Obama than increasing the total economic pie. Most Americans disagree with the Illinois senator on this, and if McCain can coherently focus the electorate on Obama's economic philosophy, that could move the national numbers a few points. The mainstream media shouting out that the race is over, has the effect, by design I think, of driving down GOP turnout, and campaign enthusiasm. There is nothing in recent history to compare with the free media onslaught for Obama this year, and there is no reason why that will change in the last week.
One final issue needs to be dealt with. While the political pundit class and the major network broadcasters are pulling their weight for Obama, I do not think it is fair to argue that all the pollsters are doing the same. I think this is a very difficult environment to be polling. Estimating how much of an increase there will be in turnout by African Americans or young voters is guesswork at best. Jay Cost has shown that the variability in polling results this year is a good bit larger than in 2004.
The Party identification distribution of those polled also has a big effect with near 90% support by both Democrats and Republicans for their party's candidate. Some pollsters weight their surveys to fit their party ID model and others don't. I think weighting is a good idea. Some surveys include cell phone only households, others don't. Obama seems to get a bit of a boost from cell phone only households. Pollsters who do not include cell phone only users need to adjust some of the age weights in their models to account for this.
I trust Rasmussen more than Gallup among the large survey size daily trackers because his numbers have been less volatile all year, and I can compare his national numbers to his state numbers to see if they make sense, and by and large they do. I do not trust polls sponsored by shamefully partisan organizations such as the dailykos.com, whose Research 2000 surveys are used to fire up contributors and activists. Other partisan pollsters -- Strategic Vision, Democracy Corps, and PPP seem to have a bit more credibility, though I trust them less than non-partisan surveys, and their history this year is predictably favorable to one side or the other compared to poll averages. I trust tracking polls more than I do the once every two week surveys by various national newspaper and TV stations or newsmagazines, which tend to have smaller sample sizes in their surveys.
If Rasmussen shows a tie or near tie next Monday in its national tracking poll, plan on a long Election night.
Richard Baehr is chief political correspondent of American Thinker.