To go along with their number from late April, Pew has released a second set of their in-depth crosstabs on the Democratic nomination campaign. Side by side, they serve as a useful indicator of where Obama has lost ground on Clinton. For example, here are the non-regional locations where Obama led in March and April, but now trails in Sep-Oct:
- Seculars: 35 point Clinton shift, from Obama +25 to Clinton +10
- Over $100K a year: 28 point Clinton shift, from Obama +12 to Clinton +16
- 18-29 year olds: 15 point Clinton shift, from Obama +5 to Clinton +10
- College graduates: 12 point Clinton shift, from Obama +7 to Clinton +5
- Men under 50: 12 point Clinton shift, from Obama +3 to Clinton +9
Clinton also gained among other demographics that were once strong for Obama, such as liberals (from Clinton +2 to Clinton +13), and those with some college education (from Clinton +2 to Clinton +16). She has actually gained in every single demographic according to Pew, which is to be expected when a national advantage shifts from 9 points to 19 points. However, every demographic I have listed here experienced a great pro-Clinton shift than the nation as a whole, especially seculars.
As I have written before, all of these major shifts matchup with prominent netroots demographics. The netroots skew male, secular, wealthy, under 50, well educated, and self-identifying liberal. For a while, I pointed to these demographics as the main reason why Clinton performed so much worse in online polls than she did in off-line polls. At one time, Clinton faced a significant, off-line disadvantage among all Democrats who fell into two or more of those categories, so why shouldn't she also face a significant, on-line disadvantage among those same demographic groups? Seems perfectly reasonable. Now, it should be pointed out that in one of the more prominent, non-scientific straw polls, Obama does not lead Clinton by all that much.
Of course, these are national polls, and as such are not the best indicators of the state of the campaign, to say the least. The most important state of all remains Iowa, where the campaign is still much closer than it is nationally. It is worth checking, however, which of these demographics matter the most in Iowa. According to the 2004 Iowa caucus entrance poll, here are some of the key demographic breakdowns:
College Degree: 55%
No college degree: 45%
Here, we can see how Clinton's large edge among self-identified conservative Democrats means very little in Iowa. Her advantage among Democrats without a college degree also means less, and thus partially explains the closer Iowa campaign. Iowa Democratic caucus goers are well-educated and liberal.
Under $50K: 53%
Over $50K: 47%
These are much more helpful demographics for Clinton in Iowa. Polls have always shown Clinton doing better among self-identifying Democrats than among independent leaning Democrats, and among lower-income voters than higher income voters. Both of those are heavily represented in the Iowa caucuses. Also, even though voters under 30 are also well represented in the Iowa caucus (probably easier for them to get out for a full evening), the median age is still 53, and absolutely dominated by Boomers. So, while there is clearly a significant younger contingent that could potentially help out Obama, overall the age of the Iowa caucuses favors Clinton. The Edwards focus on lower-income voters, both in 2004 and 2008, also seems to explain his long-term strength in the state.
If Obama, Edwards or anyone else is going to comeback against Clinton, at this late point they will have to do so in Iowa. I honestly don't know the best way to do that, but clearly whatever the non-Clinton campaigns have done so far isn't working. There isn't a single candidate who hasn't lost ground to Clinton in Iowa since late May, so clearly whatever she is doing right now in the state is working. Whether that trend continues in the biggest, and perhaps only, question left in the Democratic nomination campaign right now.