Here’s something to feed your jitters: three new polls on Pennsylvania, all showing Obama ahead by just 4-5%.

How should we regard them? And will the electoral fate of Pennsylvania matter in the big picture anyway? First, a rundown of the polls in question:

• Thursday saw the release of an NBC/Mason-Dixon poll that had Obama ahead by just 4 points, 47% to 43%.

Not reassuring: The last time Mason-Dixon polled the state was in mid-September, when it had Obama ahead by 2. At the time (McCain was still riding his post-Convention bounce), that was comparable with what other pollsters were finding: six other polls conducted around the same time ranged from a tie to a 5-point Obama lead.

Possible comfort: The poll was of a mere 625 likely voters, making for a relatively high 4% margin of error.
Possible comfort: Mason-Dixon, Nate Silver pointed out, “has .. had a Republican “lean” this cycle of perhaps 2-3 points. They are quite frequently the most favorable number for John McCain in any given state.”

The Pollster.com trendlines for the state since 4 February, when Internet polls are filtered out and the trend smoothing is set to more sensitive.

• Yesterday a new poll was released by Strategic Vision, which saw Obama ahead by just 5 points, 49% to 44%.

Not reassuring: Second such narrow lead in a row. Not reassuring: Poll had an ample sample of 1,200 likely voters.

Possible comfort: SV is a Republican polling firm. Nate Silver noted that it “has had a Republican lean and some very erratic polling in Pennsylvania all year”.
Possible comfort: SV published concurrent results from New Hampshire (Obama +9) and Michigan (Obama +13), both of which were about 2.5 points more friendly to McCain than the overall trend for those states (according to pollster.com).
Possible comfort: Last SV poll was conducted on 20-22 October, and had Obama ahead by 7. So that’s not much of a tightening. And at the time, five polls done around the same time all had Obama further ahead, by 9-13 points.
Possible comfort: While 5 points is a narrow lead, it’s hard to see Obama losing the state even if the poll is right, since it already has him at 49% of the vote. Just 1 in 7 undecided voters would need to opt for him to make for a victory.

• Today, Rasmussen has a Pennsylvania poll out showing Obama ahead by just 4 points, 51% to 47%. Obama only gets 75% of the Democrats.

Not reassuring: Third poll in a row with such a narrow margin.
Not reassuring: Trend is distinctly downward as Obama’s lead compares to a 7-point margin earlier in the week and a 13-point advantage earlier in the month.

Possible comfort: Very small sample, just 500 likely voters.
Possible comfort: Rasmussen, too, tends to have a Republican lean compared to the average pollster.
Possible comfort: While 4 points is a narrow lead, Obama still already has 50%+ of the vote, so even if the poll is right, he’d still win the state.

By ways of counterindication, the Morning Call/Muhlenberg College daily tracking poll today has Obama ahead by eight points, 52% to 44%. That, too, is down from a 13-point lead the day before yesterday, when it was 54% to 41%, but nevertheless an ample lead; eight percent is more than by how much McCain leads in Arizona and South-Dakota, according to the Pollster.com trendlines.

So will it matter either way? If Obama would lose Pennsylvania, it would certainly be a blow to chances of a comfortable Democratic win and the convincing mandate for Democratic policies that would come with it. But is his widely expected win itself at risk?

At the risk of seeming a one-trick pony in my selection of links tonight, I’ll turn again to Pollster.com and Nate Silver’s fivethirtyeight.com. Prof. Franklin at Pollster has this neat little graph:

Click to enlarge

What’s this?

This chart shows where McCain must gain, and where Obama must hold in order to get to 270.

States appear in order of Obama minus McCain margin. Width is proportional to Electoral Vote. It takes 269 to tie, 270 to win. The yellow states are very close and jumping around a bit with each new poll.

Neat, I thought. And it provides a one-glance confirmation of Obama’s alternatives to Pennsylvania. If he loses the state, the dark blue states defined by Franklin as safe for Obama are no longer enough to take him past 269. Ohio and Florida are two light blue states he defines as leaning to Obama that could easily replace it though.

That’s where another link to Silver comes in. You hear so many scenarios ‘proving’ how Obama will still win even if he loses PA, but his is the most to-the-point, bare bones summary I’ve seen. “Essentially all relevant electoral scenarios involve some combination of these five states,” he explains: Virginia, Colorado, Pennsylvania, Ohio and Nevada.

Silver, whose electoral map seems to be systematically a tad more optimistic than Pollster’s, insists that “by far the most likely scenario is that Obama wins some relatively decisive victory of anywhere from 3-12 points in the popular vote”, which would have him “accumulating somewhere between 300-400 electoral votes”. But what if the polls tighten, and Pennsylvania comes into play?

[..] IF the national polls tighten significantly [..] McCain will [probably] edge out a victory in North Carolina, Florida, Indiana, North Dakota, Montana, Georgia, and Missouri; put those states in the McCain column for the time being. Likewise, New Mexico, New Hampshire, Minnesota, Michigan, Wisconsin and Iowa all appear safe for Obama, even in the case of significant tightening. Put those in the Obama column.

That leaves our five states in play. The victory conditions for Obama involving these five states proceed something as follows:

1. Win Pennsylvania and ANY ONE of Colorado, Virginia, Ohio, or Nevada*
2. Win Ohio and EITHER Colorado OR Virginia.
3. Win Colorado AND Virginia AND Nevada.

(* Nevada produces a 269-269 tie, which would probably be resolved for Obama in the House of Represenatives.)

Now, suppose you think that Colorado is already in the bag for Obama because of his large edge in early voting there. We can then simplify the victory conditions as follows:

1. Win Pennsylvania
2. Win Ohio
3. Win Virginia AND Nevada

That’s basically what it comes down to, although I’m sure each campaign would claim that there are a larger number of states in play.