Hello professional national poll concern trolls and hand-wringers.
First of all, you're all wrong. Obama is not losing even at the national level. By any definition, he's tied. Even if you go by the least charitable method of measurement (ahem, RCP), with the appropriate amount of cherry picking, he's down by less an a percentage point. That, my friends, is a meaningless difference. And if you go by a fairer definition of which polls count and which polls do not, it is a tied race (all LV models) or Obama is slightly ahead (allowing for some accounting of RV numbers).
Regardless, as Nate has taught us, these national polls have less meaning than the state polls. State poll averages have been incredibly predictive over the last few election cycles. Go check the RCP archives. It's uncanny.
Yes, I would prefer that Obama was a couple of points ahead in the national polling. If he was, we'd be looking at 90-95% probability instead of 70-75%. At 70-75%, we have to accept the fact that there is a chance Obama could lose. It's possible. But if the state polls do not change between now and Monday, the strong likelihood is that Obama will win the presidency, perhaps even with a PV tie or slight loss.
And listen up now, Popular Vote Truthers. Even if Mittens is up by a hair in the PV, that does not spell doom for President Obama. Way more often than you might believe, the popular vote does not line up with the electoral vote. And I'm not talking about just the obvious examples (2000, 1888, 1876), where there was an EV/PV split.
Look at the popular and electoral vote margins in 1960
PV: Kennedy +0.2 EV: Kennedy +94
Look at the popular and electoral vote margins in 1968
Popular vote: Nixon +0.7/Electoral vote: Nixon +110
And let's not forget 2004, where John Kerry got beat by 2.5% in the popular vote, but would have won the election with another few tenths of a percentage in Ohio. Or 1914, when Woodrow Wilson won the popular vote by over 3%, but came
within 3,800 votes in California from losing.
Seems to me, this sort of thing happens ALL THE TIME, where the electoral vote and popular vote do not line up. So why is it so hard to believe that Obama might lose the popular vote by 0.1 or 0.6 or something and win the EV by something like 281-257?
So to sum up, we're ahead. We'll probably win. End of story.