The most compelling demographic analysis I've seen lately: Alan Abramowitz's "The Incredible Shrinking Republican Base" on Real Clear Politics.
Short story: White Married Christians–the combined demographic–is and has been the base of Republicans' victories. Ignore income. Ignore gender. Ignore age. If you're white, married, and Christian, the odds are six or seven in ten that you'll vote Republican.

And that demographic is in precipitous decline. It's below 50% for the first time in (modern?) history.

I saw another telling statistic the other day: for the first time ever, among children in the US under five years old, less than 50% are white.
There's a pretty stunning new NPR poll out (PDF) (conducted by one Republican and one Democrat) showing that Pubs actually prefer Democratic policies by wide margins.
This is Very Good News.
But what's amazing is how brainwashed Pubs are by party affiliation, compared to Dems. If their beliefs aren't validated by Herr Comrade Party Leader, they change their beliefs.
The pollsters asked the same questions (do you prefer this position or that one), but in half the cases they identified the positions as being Democratic or Republican.
Here's the difference in support for each party's position when the respondents were told that it was their party's position. (Sorry for the table formatting…it's lined up in the editor…)
Pubs Dems
Economy +20% -9%
Iraq +11 -4
Foreign Trade +20 -18
Taxes +28 -9
Example: Only 38% of Republicans support Republican tax policies.
Tell them it's the "Republican" policy position? All of a sudden 66% support it. Whodathunkit.
The Republicans don't have a branding problem (at least not within their own party). (Though it sure looks like the Dems do.) Quite the contrary. Their problem is that even Republicans don't like Republican policies. They prefer Democrats' policies—except when they're told that the policies they approve of are Democratic policies.
Since emerging as the presumptive Republican nominee, John McCain's primary results have been tepid by any measure.
5/27-ID 70%
5/20-OR 85
5/20-KY 72
5/13-WV 76
5/6-IN 78
5/6-NC 74
4/22-PA 73
4/11-MS 79
Compare George Bush's primary results after sealing the nomination–consistently in the high 90s. McCain has broken 80% exactly once, and his latest–70% in Idaho–is his lowest number yet.
If November is about turning out the vote–and it definitely will be–these results can only be heartening to the Democrats.
Speaking of who Obama should choose as a running mate, David Brooks thinks "He should be thinking about who can help him govern successfully so he can get re-elected."
This reminds me of a scene from The West Wing. Jed Bartlett is trying to decide whether to run again with John Hoynes, with whom he has a very fractious relationship. He decides to go with him, and when Hoynes asks him, rather perplexedly, why, Bartlett writes on a piece of paper, folds it, and hands it to him.
Four words: "Because I might die."
In his NYT Op-Ed today, David Brooks makes a very good point:
McCain has infinitely better grounds than Obama to run as a do-what-it-takes reformer. He has a long record of taking on not only the other party, but his own.
Case in point: McCain voted against the farm bill, a bill that’s uniformly vilified by everyone except the legislators who vote for it, and the lobbyists who cause them to do so. Obama voted for it.
But as for the title of the piece, "Talking versus Doing," it’s not clear that McCain is quite the paragon that Brooks suggests. Brooks quotes McCain, for instance, using words that make my heart sing:
“In all my reforms, the goal is not to denigrate government but to make it better, not to deride government but to restore its good name.â€
That’s the right contra-Reagan message–government is not the problem, bad government is the problem. He’s singing the hymn that Obama should be owning. But McCain’s proposed tax cuts–equal to total nondefense discretionary spending–say something very different: that bleeding the beast and further eviscerating our government is the answer to our problems.
McCain also makes very nice sounds about multilateralism in foreign policy. But his specific proposals suggest a continuation of policies that have resulted in the greatest decline in American power in living memory.
Update June 2012: See data through 2010 here.
People love to cherry-pick statistics to show that the US, or Europe, is winning the growth game. That got me curious: if you look at all the possible growth periods, who’s ahead (most)? Short answer: no clear winner. The results look pretty random. Over the longest periods, Europe is consistently ahead. But some shorter but still lengthy periods give the US the advantage.
Change in real GDP per capita (PPP), difference between EU and US
US growth percentage minus EU15 growth percentage (positive numbers=US ahead)
Starting years at left, ending years, top. Gray is >29 years. Outline is >19 years.

Here’s the same table expressed as a percentage difference.
(US growth %-EU growth%)/EU growth %

All data is from the OECD. I just did the arithmetic.
Update 10-5-2010: The tables above are great for demonstrating the danger of cherry-picking periods, but a graphic provides a more immediately-apprehensible understanding of the long flow of history:

Given the inherent difficulty of measuring GDP accurately, we can say that these are essentially identical. They both basically doubled over those 36 years. (Though the U.S. growth path shows far more volatility.)