<H/T to Ed Kilgore>

This morning's combination of blog reading and coffee brought me to an interesting post at The Liberal Oasis.  One of our main goals at this site is to empower people to understand how polls, analytics and data really work so that they can do what Bill Scher did so well in this post.

Here are a few nuggets from his post, with my commentary following.

"It's the McCain number that moved up, which was extremely low in those earlier polls because of conservative disaffection. It is not surprising to see some base consolidation that gets McCain into the low 40s.

Still, today's polls have Obama up by 9, 8 and 6 points.

You can claim that he "should" be doing better, but a final victory with that margin would be earth-shifting."

Anyone wondering why the Obama/McCain race is seemingly so close would be well advised to go back and look at the history of our Presidential elections. As far as I can recall, even Reagan v Mondale wasn't decided with such massive margins.  I can't think of a single election, short of FDR's victory in '32, that would have such huge margins.   There's a part of me that thinks that what we're seeing here is the result of a successful attempt by the GOP to raise expectations on Obama's performance so that they can claim that they're not as behind as people would have thought.

Also, hfor everyone repeating the conventional wisdom on Obama and race, here's a special one just for you:

"The evidence? The NYT says:

After years of growing political polarization, much of the divide in American politics is partisan. But Americans' perceptions of the fall presidential election between Mr. Obama, Democrat of Illinois, and Senator John McCain, Republican of Arizona, also underlined the racial discord that the poll found. More than 80 percent of black voters said they had a favorable opinion of Mr. Obama; about 30 [percent of white voters [sic -- actually 31] said they had a favorable opinion of him.

That's highly misleading, as it makes it sound like massive numbers of whites have an unfavorable opinion of Obama.

The complete poll data shows that 1) 31% said they are "undecided" or "haven't heard enough," and 2) McCain doesn't do much better, only scoring a 35% favorable rating from whites, with a similar number also not expressing an opinion.

The article also notes that McCain leads Obama among whites 46%-37%, a 9-point margin. (The other two polls today have McCain up 8 and 7 points among whites.)

But it doesn't tell you that in 2004, President Bush beat Sen. John Kerry among whites by 17 points.

Obama runs at least 8 points better among whites than Kerry, not to mention performing vastly better among Latino voters (39 point lead) than Kerry (9 points).

Also of note is this line buried at the bottom of the Washington Post analysis of its poll: "The candidates are tied among whites who earn less than $50,000 a year, while McCain leads by 10 percentage points among those earning more than that."

Yes, Obama runs better among white working-class voters than other whites.

Not a half-bad step forward for race relations, in my book."

It seems as if certain ideas just get repeated frequently enough that it doesn't matter whether or not they're borne out by data.  Anyone who's worked with numbers for any period of time can tell you that, and it's not just in politics. (One fun one, and that's easily, easily testable is whether or not Super Bowl Sunday has more domestic violence instances than any other day of the year. Hint: It doesn't.)  Maybe it's because this idea corresponds with everything that we already want to believe, it's easy to believe. The simple fact here, though, is that Kerry had a good ole boy at the bottom of his ticket, and still did worse than Obama is currently doing with blue collar white Americans.   This reversal is massive.

Moreover, to all of those people who thought that the massive numbers of blue collar and rural white Americans who preferred Clinton to Obama meant that those white people would not vote for Obama in a general election, I would remind you to remember that you cannot meaningfully extrapolate from a primary in which people express their preference amongst X number of Democrats that they will switch Republican if their preferred Democrat didn't win the primary.  The same thing obtains for Republicans.   In the end, party identification remains the single strongest predictor of voting behavior, and that's that.

Dirty D