Pew Research’s report today indicates dwindling support for McCain but I think it’s missing closing week trends.

Zogby, which began as an outlier, showed Obama’s support increased from a low of 47.1 -45.2 on October 7th, to a peak of 52.2 -40.3 on October 22nd. But with the debates concluded and McCain narrowing his closing argument to the longstanding GOP standard of being better on taxes and spending, he’s whittled that down to 49-44.7, a 4.3 pt race as of polling completed yesterday.

Yet a 4% margin in the final week has won every race in many decades except one: 1980. So in itself, that margin is no cause for concern. And tightening in the final week has always been expected by pollsters and election historians. So the trend doesn’t mean it’s time to worry either.

Together, however, both factors made me take notice. So I began checking trends and margins per other pollsters. Here’s what I found:

Gallup still has it at 51-44, but that’s 5 pts closer in a week.

ABC/WaPo has it at 52-45, a 4 pt narrowing.

Diageo/Hotline has it at 50-42, just 1 pt closer.

DKos/R2000 has it at 50-43, 5 pts closer.

And Rasmussen, the most reliable national trend pollster, has it at 51-46, a 3 pt narrowing.

ARG, a Republican pollster, has the race widening by 1 pt, now at 50-45.

So a narrowing between 3% and 5% has occurred with outlier polls removed. And only Zogby is showing Obama below 50%, at 49%. Zogby’s also the only major pollster calling it as narrow as 4.3%. Even the Republican ARG says it’s 5%. But then, so does the very reliable Rasmussen.

That gives me pause because I felt Obama would hold the line at 6%, after the narrowing. But let me add, it draws my concern but not any worry because tomorrow’s as close as McCain is likely to get in the national polling. What’s clear is the following Republican strategies are now being employed:

1) McCain’s pulled his central campaign back from several battleground states. Its focus is down to the three richest in electoral votes: FL, PA and OH.

2) Palin still gets sent to a few others where it’s felt she might have slightly more pull than McCain, but she’s also showing up in PA/OH a lot, too.

3) The RNC has been picking up some of the defense in states where McCain has slipped recently. The most visible of these is MT, which turned blue in the last in-state poll there. Previously popular among Libertarians, both McCain and Palin are finding Ron Paul more popular there. And they’re willing to let MT’s 3 EVs go.

4) The RNC also began running a Jeremiah Wright ad in the big 3 states that McCain must now count on. So the race card is being played as long anticipated, but McCain will try to claim he’s not authorizing it, which is a weak argument at best. Still, if he continues to deliver his economy-only message while the RNC does its best to draw the racist vote in, we have to weigh what the worst case could be if McCain succeeds in winning all three.

So for starters, revisit my state-by-state projections from three days ago. Summarizing, I indicated Obama had 255 safe EVs already. And nothing in any poll anywhere provides any evidence that has changed towards McCain.

I projected that Obama had a sufficient lead to win in NM, CO, NV and VA. I called the very close races for MO , OH and IN as likely for Obama. I rated NH, OH, IN, NC and FL too close to call. And MT and GA were Leans McCain tossups. And ND I called a mystery.

Do the polls suggest any of those have changed in the past three days? Yes. VA has shifted to safe Obama.

The reliable Rasmussen says 51-47, a narrowing of 6 pts in the past 10 days. But SUSA, the most accurate state pollster, now says 52-43, which is only 1 pt narrower in the same ten day period. PPP concurs with SUSA in that period. And Zogby has it as a 7% contest. Without campaigning there, there’s just no way McCain can close a 4% to 9% gap.

Add VA’s 13 EVs to 255 and he’s only 2 votes shy of a victory. Which means McCain HAS TO get PA, or his other option is to win every other remaining state. And despite modest narrowing of the race, there’s no poll showing PA closer than a 9 pt Obama win.

Te only Kerry state McCain still has a shot at is NH, but that only puts Obama 6 EVs shy of a win. Every other undecided state except MT and ND provide the margin of victory in that event.

Which means McCain must win FL, OH, NC, GA, IN, NH, CO, NM, NV, and either MT or ND, to upset Obama. And here’s what’s happening in each of them.

FL: a 5 pt shift here, per Rasmussen, in 7 days. But favoring Obama, who now leads by 4%. A very ominous sign for McCain.

OH: a 6 pt shift here, per Rasmussen, too. Also for Obama, who now leads by 4%. SUSA concurs exactly in the current poll, though showing a 1% narrowing over 13 days. These two results from his three most contested states, has to be discouraging for McCain because now he has 6 days to get at least 4% from both. And neither trend favors him.

NC: the shift to this tossup state and current trend now makes it lean slightly McCain.

GA: still favors McCain but the trend is all Obama’s.

IN: SUSA shifted 7 pts for Obama from 10/1 to 10/22, and has him leading 49-45. Zogby says it’s 50-44 for McCain but with no prior poll, can’t define a trend. Thus it remains uncertain.

NH: Rasmussen has McCain closing from a 10 pt gap to 4. An in state pollster has Obama winning it by a huge margin, however.

CO has narrowed by 1 pt, but Obama retains at least a 4 pt lead and still carries 50%

NV: mixed results, but still looks like a 4% lead for Obama.

MT: an in-state poll now gives Obama the lead here.

ND Too little data to guess.

NM: needs a poll more recent than 10/13 but still looks safe for Obama. And with polls in his home state of AZ now narrowing to 6%, 4% and 2%, with that slippage attributed to the Latino vote, McCain just isn’t going to win the most Latino state in the country.

With Obama getting NM’s 5 EVs and VA’s 13, if McCain gets NH - and everything else - he wins by the smallest possible margin of 270-268. But as I noted, OH and FL have trended away from him and Obama leads by at least 4% in both of those AND CO, NV and NH. Plus he now leads in MT.

Finally comes the kickers. Undecided voters start choosing in the final 5 days, typically. And the night before that, Obama’s half-hour ad is set to run. At the national level, that should stop further narrowing. Those undecideds often favor perceived winners or go with obvious trends. With a 4% gap nationwide, Obama should get AT LEAST 50% of the undecideds. And McCain will only get about a day and a half to counter Obama’s ad before the majority of those undecideds pour in.

If that ad’s a hit, he won’t be able to stop Obama’s closing momentum.

And let’s not forget how many voters have ALREADY voted, before the gaps closed.

So as I said, the race now has me concerned but not worried. Despite his Herculean closing effort, too much remains against McCain to pull off a miracle finish. And there’s a vast amount of economic data due on Thursday and Friday, so there’s little to indicate any positive news after tomorrow’s interest rate cut by the Fed.