A couple of weeks ago, I argued that national trends were moving decisively in Barack Obama’s direction, though in a rather predictable way. Those trends have continued, magnified, perhaps, by the financial crisis.

Some eschew the national data, rightly pointing out that we really have 51 separate state elections. That is true, but only up to a point. Those national polls necessarily reflect the views of people who live in states.

Thus, as a practical matter, the Electoral College only comes into play if the national popular vote is close.

As I noted in July, a four-point Obama popular vote win translates into roughly 310-325 electoral votes. In short, the odds of winning the national popular vote by four points or more, and losing the Electoral College, are de minimis.

Pretend for a moment, though, that the popular vote once again gets close, a situation that will affect the state poll results as well. If Obama’s national margin should fall from its current average of about seven points, to two or three points, how will he fare in the Electoral College?

Even in that “tight race” scenario, everything we know from history and polling tells us Obama should hold all the Kerry states, with the possible exception of New Hampshire. That yields 249 electoral votes for Obama.

Since 2004, two Bush states have clearly moved to the Democratic column — Iowa, where Obama leads by an average of over 10 points, and New Mexico, where he leads by about seven. Thus, even if Obama’s margin shrunk in these states by the same amount I am speculating about nationally, he would retain healthy leads there. Putting those two states in the Obama column takes him to 261 electoral votes, nine short of a majority.

There are six states currently in play that could, by themselves, provide those nine votes or more — Florida (27), Ohio (20), North Carolina (15), Virginia (13), Missouri (11) and Colorado (9). While some polls suggest an Obama lead in North Carolina and Missouri, none in either state has given him the kind of margin he would need to withstand a four-to-five-point national drop. Again, I am not predicting Obama will lose those states on Election Day — he may well win them — but not if he wins nationally by just two or three points.

By contrast, there is at least some evidence from Ohio, Florida, Virginia and Colorado that Obama could drop nationally, but still win these states—and that should make them AAA Prime targets over the next 30 days.

Of course, there are other possible combinations. Winning New Hampshire, Nevada and stealing one of Nebraska’s congressional districts would put Obama at 270 without any of the six states above.

Losing Pennsylvania, Minnesota or Wisconsin would complicate matters seriously, though that outcome is just not likely to happen, even if Obama falls to a two- or three-point national popular-vote lead. John Kerry won those states while losing by 2.5 points nationally and John McCain enjoys no special relationship with any of them (as he arguably does with New Hampshire). However, a loss in Wisconsin or Minnesota would not be fatal, as either Ohio or Florida would be sufficient to more than make up for one of the two.

Since there are no sure things, the Obama campaign is wise to keep as many roads open to 270 as possible — and with a significant popular-vote margin, an electoral-vote landslide is within reach. Even if the popular vote gets close, though, the noose is tightening around John McCain and his paths to victory are many fewer and much farther between.

Mellman is president of The Mellman Group and has worked for Democratic candidates and causes since 1982. Current clients include the majority leaders of both the House and Senate.