Rasmussen has new numbers on partisan self-identification that, over the last few years, show both a clear Republican decline and a clear Independent rise (May 2007 results results are based on 15,000 interviews per month, all other results are three months quarterly averages, based on 45,000 interviews):

Partisan Self-Identification, 2004-2007

Date	Democrats	Indy / Other	Republicans
May 07	36.3%	32.9%	30.8%
Mar 07	37.5%	30.7%	31.8%
Dec 06	37.7%	30.0%	31.3%
Sep 06	37.0%	30.7%	32.3%
June 06	36.6%	30.1%	33.3%
Mar 06	36.5%	29.4%	34.1%
Dec 05	36.6%	29.4%	34.0%
Sep 05	36.4%	28.5%	35.1%
June 05	37.2%	27.8%	35.0%
Mar 05	37.3%	26.9%	35.8%
Dec 04	38.7%	24.2%	37.1%
Sep 04	38.0%	26.1%	35.9%
June 04	37.2%	28.1%	34.7%
Mar 04	37.9%	27.6%	34.5%

This corresponds with the long-term trend of increasing voting registration outside the Democratic and Republican folds:

Long-term voter registration trends, % of all registered voters

Year	Democratic	Republican	Indy / Other
2004	36.8%	28.0%	21.7%
2000	36.3%	27.8%	19.6%
1996	35.9%	26.9%	15.8%
1992	36.6%	25.4%	12.7%
1988	37.6%	25.6%	10.5%
1984	40.2%	24.0%	10.2%
1980	40.0%	22.4%	8.2%
1976	41.5%	21.6%	6.8%
1972	44.4%	24.7%	4.7%
1968	45.8%	26.1%	3.2%
1964	49.4%	24.2%	1.7%
1960	48.3%	27.2%	1.6%

(Note: some states do not have partisan registration, thus explaining why these numbers never really come close to 100%)


This appears to be rock-solid evidence documenting the rise of Independent voters nationwide and long-term. However, it is important to keep in mind that not all Independents are the same. Most, but not all, self-proclaimed Independents will, when pushed, identify with one of the two major parties. Interestingly, Pew shows, in the image on the right, that the number of self-proclaimed independents who refuse to identify with one of the two major parties has remained constant in a range of 12-15% of the population since 1990.

So, while the number of people self-identifying as Independents without being pushed has gone up since 1990, after Independents are pushed the number of "true" Independents has remained constant. Further, exit polls from every presidential election since 1976 show that that percentage of self-identified Independents has remained in a narrow range of 23-27% of the electorate since 1980. In fact, Independents peaked as a percentage of the electorate in the post-watergate election of 1976, at 41%. So, while the number of people registering to vote as something besides Democrats or Republicans has increased dramatically since the mid-1970's, the number of voters who self-identify as Independents in presidential elections has now changed much. This probably the result of gradually decreasing turnout among Independents, and the tendency for independent self-identification to drop in presidential election years.

Overall, what this data shows is that while there is, in fact, a slow drift away from the two parties, Independents are not a rising force in American politics just yet. While there is a general tendency toward decreasing identification with, and participation in, the Democratic and Republican parties, Independents have still not turned away from the two-party system. It is possible that the increasing tendency of Americans to register with other parties (or with no parties), and that their increasing tendency to not identify with either major party unless pushed represents the first stage is a shift away from the two-party system. However, until third party candidates start performing at higher levels, the change is not particularly significant and that longer term trend cannot be proven. Right now, most Independents actually are either Democrats or Republicans, they just don't like calling themselves such. In fact, third party performance is actually declining. While that may change in either a few years or a few decades, I for one am not holding my breath.

Pew has just dropped an amazing and important crosstab set on the Democratic nomination campaign. It frustrated me tremendously that they still include Gore in the questioning, but it is still worth a look anyway.

Looking at these crosstabs, it occurs to me that there is an important structural flaw in the inflated Clinton poll theory that I have been overlooking. Although Pew does not offer crosstabs to prove this, I still believe that Obama probably does much better relative to Clinton among voters who are paying very close, or somewhat close, attention to the campaign than among voters who are not paying much attention at all. However, closed primaries in several large February 5th states might cancel out that advantage, since Obama performs relatively better among Democratic-leaning independents who won't be able to vote in closed primaries than he does among self-identified Democrats (in this case, I am assuming self-identified Democrats are more likely to be registered Democrats than are Democratic-leaning independents). Now, I don't know how many February 5th and earlier states this rule applies to, but I do know it applies to Pennsylvania where I live. I think it applies to New York as well, but that doesn't really matter one way or the other, because Clinton will almost certainly win New York. Overall, it might apply to enough states to counteract all, or at least half, of the advantage Obama currently holds high-information voters who are heavily tuned into the campaign. This probably means that Clinton is still ahead even among the universe of likely caucus-goers and primary voters in the Democratic nomination process. It does not, however, change my opinion that the theory needs to be tested.

Leaving that aside, there are some other very interesting cross-tabs here. First, it is useful to note that Clinton's advantage over Obama among liberal and moderates is negligible, but her edge among conservatives is enormous. Second, Obama holds a gaping lead among seculars--and even Edwards is tied with Clinton among seculars--but Clinton holds the edge among all other religious demographics. Many of the other patterns we have seen are also replicated here: younger voters, wealthier voters, highly educated voters, and male voters all skew toward Obama much more heavily than do other demographics. There clearly seems to be a cultural and class based divide between Clinton and Obama supporters, at least right now. It reminds me of something I wrote over four months ago, even before any crosstabs came out revealing this divide:

I think Obama, simply in terms of his demeanor and his biography, strongly appeals to politicos from a new generation and a new socioeconomic class because he strikes them in some sort of gut, intuitive level as being from that class. Multi-ethnic, post-Vietnam, highly educated, raised in a major urban center--these are many of the cosmopolitan, self-creating, forward looking aspects of life for many younger professionals. As much as we may or may not like Bill Clinton, coming from a little town in Arkansas is not a story many Americans can relate to anymore, because we just didn't grow up that way. Even John Edwards's story of growing up in a mill town when the mill closed seems very, very rustic for a northeasterner such as myself, since our mills closed down sixty years ago to move to places like North Carolina. These rustic visions of America simply are not where people are at these days, especially news junkies and activists within the Democratic Party and the bluer parts of America. Those people instead look to places like Harlem, where Bill Clinton now keeps his offices. People moving into the gentrifying areas of Harlem probably like Barack Obama quite a bit, and probably feel some sort of gut-level, identity-based connection with him that they can't even quite put their finger on at this point.

An important question is not just whether either Obama or Clinton's numbers are inflated, but whether either of their supporters are leading indicators for future trends in the campaign. It is possible that they each currently hold advantages in cultural groups that do not communicate with one another very often, and so neither group will influence the other. But it certainly is interesting, and I wish a lot more polling outfits would release crosstabs this like so we can monitor this socioeconomic divide within the Democratic rank and file.

When looking at the baseline electoral potential of the two major parties, the current balance of partisan self-identification might be the most useful metric available. If a roughly equal number of people are willing to self-identify as Democrats and as Republicans, then the two parties have a roughly equal number of potential voters. However, right now, the two parties are not equal at all. From Gallup (emphasis mine): An average of all national Gallup polling in 2006, consisting of interviews with more than 30,000 adult Americans, finds 34% of Americans identifying as Democrats, 30% as Republicans, and 34% as independents. The parties had been relatively even in terms of national strength since 2001. The most recent figures represent the largest Democratic advantage since the Clinton presidency.(...)

The increasing Democratic advantage is mainly due to declining Republican identification, rather than increasing Democratic identification. From 2004-2006, Republican identification declined from 34% to 30%, while Democratic identification increased by less than a percentage point (33.6% to 34.3%). During the last three years, the percentage of Americans identifying as independents increased from 31% to 34%.

The Democrats' advantage expands when taking into account the "leanings" of independents. In 2006, 50% of Americans identified as Democrats or were independents who said they leaned toward the Democratic Party. Forty percent identified as Republicans or leaned to the Republican Party. That 10-point advantage more than doubled the Democrats' 4-point advantage in 2005, and is the largest gap Gallup has measured in any year for either party since it regularly began tracking leaned party identification in 1991. This is the first time since 1991 that a party's support reached the 50% level.Two things. First, the current Democratic advantage is mainly due to movement within the Independent pool. Many people who previously identified as Republicans are now identifying as Independents, and many Independents who once leaned toward Republicans are now leaning toward Democrats. This means that while Democrats have gained, it is not yet a realignment. We will have achieved realignment-level success when this movement within the Independent pool moves all the way toward more people identifying as Democrats. Given the long-term trend of more people to self-identify with neither major party, this might prove exceedingly difficult.

Second, on a more positive note, the rate of change in favor of Democrats appears to be increasing. The last nine months of 2006 saw an average Democratic advantage of 11.5%, and the final three months of 2006 saw a Democratic advantage of 14.2%. This compares with a stable Democratic advantage of 6.6% from July of 2005 through March of 2006. The percentage of Democratic self-identifiers, not including Democratic-leaning independents, rose by 2-2.5% during the final nine months of 2006. This might suggest that Democrats are in fact on the verge of a very real realignment.

State level data shows even more good news: Based on their 2006 averages in leaned party identification, Gallup classifies 33 states as Democratic in orientation (the state showed a statistically significant advantage in Democratic leaning in 2006) and six as Republican (the state showed a statistically significant advantage in Republican leaning in 2006). The remaining 10 states (including District of Columbia, but not including Alaska and Hawaii since Gallup does not interview in those states) are considered competitive, because the leading party's advantage is within the margin of error for that state's data. The overall results show a net gain of six states for Democrats and a net loss of six for Republicans from 2005. The shift since 2003 has been dramatic, when Republican-leaning states outnumbered Democratic-leaning states 20-14. Over the past three years, Republicans have lost their advantage in 14 states, and Democrats have gained a statistically significant advantage in 19 states. Since 2003, the individual gains and losses were as follows:

• States where Republicans no longer have an edge: Alabama, Colorado, Georgia, Kansas, Mississippi, North Dakota, South Dakota, Tennessee
  
  
• States where Democrats have gained an edge: Delaware, Florida, Maine, Minnesota, Nevada, New Jersey, New Mexico, North Carolina, Ohio, Pennsylvania, Washington, West Virginia, Wisconsin
  
  
• States where both took place: Arizona, Indiana, Missouri, New Hampshire, Oklahoma, Virginia
  
  
• States where Democrats lost an edge: Louisiana

With the exception of Colorado and Gore's home state of Tennessee, both of which are now "competitive" and actually recorded slightly more Democrats than Republicans in 2006, every single state in the country that was considered a "swing state" in either the 2000 or 2004 Presidential election now leans toward Democrats in terms of partisan self-identification. In most cases, that is a change from 2003. So, while we have not achieved realignment yet, we certainly appear to be putting the pieces in place. And the fifty-state strategy marches forward.

Generation Y, defined by the census as those Americans born between 1977 and 1994, is the largest American generation since the Baby Boomers. In fact, it is almost exactly the same size as the Baby Boom generation, and may soon be the largest of all. It also now forms the entire 18-29 year voting demographic that we see on exit polls. In 2006, Generation Y made up 12% of the electorate, and broke for Democrats 60%-39%. Democrats also hold an enormous, double-digit lead in partisan identification among this age group. In 2004, that advantage was 39%-28%. In 2006, it had increased further to 41%--28%.

This is important because if someone develops a voting pattern at a young age, that person is likely to continue voting that way throughout her or his life:Continuing a trend that began in the mid-1990s, young voters once again disproportionately identified themselves as liberals and gave a supermajority to Democrats. Unless basic findings of political science have been repealed, these formative experiences of early adulthood are likely to influence electoral behavior throughout the life of this cohort. Generation X (1965-1976), has pretty much passed the age where formative voting experiences are developed. However, it is a small generation, and while it leans Democrat (and is more liberal than older generations, as conservatives only hold an eight-hold ideological self-identification edge), it is nowhere near the level of progressive generation that is still under the age of thirty. In a rather stunning statistic, ideological self-identification among Gen Yers actually slightly favors liberals, despite a double-digit gap for conservatives within the nation as a whole. This is a generation that is also only 61% white, and less than 40% white Christian. In short, it does not cohere with the ideological or identity tendencies of the modern conservative movement at all.

Given its enormous size, if Generation Y grows to voting and political maturity with the same ideological and partisan tendencies it currently displays, it will entirely transform the national political environment by serving as the backbone of by far the most progressive governing majority America has ever experienced. How do we make that happen? The battle can actually be nearly won in less than two years time, if Democrats nominate a candidate loved by young people, and if that nominee becomes President. Consider the following:Speaking as a political scientist.... Generally speaking, the "you get more conservative as you get older" myth really is a myth. People's ideological/partisan identification don't change much after the age of 30. If someone votes for the same party three times in a row, they're hooked for life. It takes some earth-shattering to change after that.

People don't get more conservative as they get older, but they do get more rigid. What happens is that ideology acts as an informational screen - people shield out stuff that is inconsistent with their predispositions (which is why FOX News works). So as we get older, our attitudes get reinforced.

So liberals should NOT get happy if people who are under 30 are on the left, because the young are very volatile. But after thirty, it's smooth sailing. If the 2008 Democratic nominee becomes President, then that person will almost certainly be the Democratic nominee in 2012 as well. If that person is loved by young people, then that will make a long, 2004-2012 run where young voters broke heavily for Democrats, self-identified as Democrats, and self-identified as liberals. Thus, it will match the 1980-1988 run where young Late Boomers broke heavily for Republicans in the three Presidential landslides of that decade. When that generation grew to political maturity, it resulted in by far the most Republican-identifying generation in over half a century, the 1994 Republican landslide, and the general sense of creeping conservatism the country experienced through the 1990's and first half of our current decade. Generation Y holds the potential to do exactly the same thing for America, only in reverse, for the 2000 "teens," 2020's and 2030's. And much, if not most, of whether or not that happens will depend on what Democrats do over the next two years.

Realignments do not have to take place in just a single election. The Republican and conservative "revolution" was a slow climb, starting with the election of 1978 and ending in the election of 2002. In the 1930's, the New Deal coalition was built over the course of several elections, starting really in 1930 and continuing until 1938. Considering the demographic and political characteristics of Generation Y, the period starting in 2006 and ending in 2012 has the potential to forge a long-term, very progressive, solid-blue governing majority in America for a long time to come. Finding a potentially transformational Presidential candidate is an essential part of making this happen. Here is to hoping that at least one Democratic candidate will be able to step up and fill that role.

Using the information form a great diary by democraticavenger, here is the current percentage, and change since 2004, of partisan self-identification among the electorate in states where there were exit polls in 2006. I have grouped the states in ways they may be relevant to 2008.

Base States

• MA: Dems 42%--19%. Shift: Even
• NY: Dems 47%--25% Reps. Shift: Dems +6
• RI: Dems 38%--18% Reps. Shift: Reps +3
• MD: Dems 50%--31% Reps. Shift: Dems +10
• HI: Dems 40%--23% Reps. Shift: Dems +1
• IL: Dems 46%--31% Reps. Shift: Dems +10
• CT: Dems 38%--26% Reps. Shift: Dems +5
• CA: Dems 41%--35% Reps. Shift: Even
• VT: Dems 29%--27% Reps. Shift: Reps +2
• TN: Reps 38%--34% Dems. Shift: Dems +4
• MT: Reps 39%--32% Dems. Shift: Even
• ND: Reps 38%--29% Dems. Shift: Dems +5
• TX: Reps 41%--31% Dems. Shift: Dems +1
• NE: Reps 50%--27% Dems. Shift: Dems +6
• WY: Reps 56%--27% Dems. Shift: Reps +1
• UT: Reps 56%--20% Dems. Shift: Dems +3

Some blue states got a lot bluer: NY, MD, IL and CT in particular. New York is now just as bad for Republicans as Nebraska is for Democrats. Democrats made positive gains in North Dakota, Nebraska, and Tennessee, but those are all long-term projects. Color me a little concerned about California.

Outer Swing States

• WV: Dems 51%--32% Reps. Shift: Dems +1
• NJ: Dems 41%--28% Reps. Shift: Dems +5
• WA: Dems 39%--29% Reps. Shift: Dems +6
• ME: Dems 37%--29% Reps. Shift: Dems +7
• VA: Reps 39%--36% Dems. Shift: Dems +1
• AZ: Reps 41%--32% Dems. Shift: Dems +5

It looks like Washington, New Jersey and Maine have now all cemented their status as deep blue states. I also like the favorable trend in Arizona. Maybe it leaves this category in 2008, and becomes a true swing state. That would be huge, sine Arizona is one of the three states poised to gain large numbers of congressional districts in the next twenty-five years.

Inner Swing States

• NM: Dems 41%--32% Reps. Shift: Dems +2
• MI: Dems 40%--33% Reps. Shift: Dems +2
• WI: Dems 39%--34% Reps. Shift: Dems +8
• PA: Dems 43%--38% Reps. Shift: Dems +3
• MN: Dems 40%--36% Reps. Shift: Dems +1
• OH: Dems 40%--37% Reps. Shift: Dems +8
• MO: Reps 39%--37% Dems. Shift: Reps +1
• FL: Reps 39%--36% Dems. Shift: Dems +1
• NV: Reps 40%--33% Dems. Shift: Reps +3

With the exception of Ohio and Wisconsin, Democrats did not make big gains here. The Nevada trend is worrying, while the Pennsylvania trend indicates that this may move to the "outer swing states" category in 2008. Pennsylvania really was the epicenter of the Democratic wave this year, which could mean very big things in future elections.

Of course, I am not convinced that 2008 will be as close a Presidential election as 2000 and 2004, making the concept of swing states less relevant. We are going to have a lot of trouble if McCain is nominated and, barring a right-wing third-party revolt, if Giuliani gets nominated. On the other hand, I feel reasonably confident that any Democrat will mop the floor with any other Republican nominee. Then again, I think both McCain and Giuliani will have a surprisingly difficult go of it in 2007 with the lights blaring on them full-time. I am just not convinced they are good campaigners and able to deal with media scrutiny outside the comfy confines of Sunday morning talk shows