If you have "no place to go," come here!

Count Whose Vote 2: Independents vs Moderates

ABSTRACT: Based on available exit polling from states that held primary elections, while Obama dominates the "Independent" voter, Hillary Clinton actually does slightly better among "Moderate" voters— and this is even more true in crucial swing states. The data suggests that a more comprehensive review of all such "electability" factors is required.

Data tables and an explanation of the methodology employed can be found here. [scroll down]. Major thanks to Lambert for all his help with this piece!

* * *

Note: This study is based on exit polling data from primary states. With regard to Michigan, all "uncommitted" voters are treated as if they had voted for Barack Obama. "Swing states" consist of those states not identified as either "Safely Democratic" or "Safely Republican" as of 2/20/08 by Rasmussen Reports for which data is available (NH, AR, DE, FL, MI, MO, NJ, NM, WI, VA). 1 There is an appendix that includes information regarding states that held primaries but for which exit polling data is not available (WA, DC), and data from caucus states where exit polling data is available (IA, NV).

The media, and the Obama campaign, have endlessly hyped Barack Obama's appeal to "Independent" voters. A search of Google News at 8:00 AM EST on February 23 for "'Barack Obama', 'independent voters'" turns up 2490 hits for the last month, for "'Hillary Clinton', 'independent voters'" the number of hits is 1299. But almost no attention has been paid to the crucial "Moderate" voter demographic. A "'Barack Obama', 'moderate voters'" Google News search finds just 54 hits during the same month-long period, "'Hillary Clinton', 'moderate voters" also turns up 54 hits.

Yet exit polling data reveals that the "Moderate" demographic is much larger than "Independents". And there is no correlation between the voting patterns of "Independents" and "Moderates". And "Moderate" voter are the key constituency that will be crucial in swing states in November.

"Independent" itself is a meaningless term. It includes the entire ideological spectrum: people on the ends of the ideological spectrum who think the Democrats are too conservative or the Republicans too liberal, people who share some views with both parties but can identify with neither of them, and people whose views are consistent with a party, but just don't want to be "labeled." It also includes those who consider themselves "Moderates," people who think that the Democrat party is too liberal, and the GOP too conservative.

"Moderate," however, is meaningful, especially given the choices that those polled were given ("Very Liberal", "Somewhat Liberal", "Moderate", "Somewhat Conservative", and "Very Conservative"). And while there is doubtless some variation in how each individual defines a "Moderate" (the average Rhode Island "Moderate" is probably more liberal than the average "Moderate" from Utah), that variability is not truly significant—especially when the general election is winner-take-all in each state.

Moderates also make up a much larger part of the electorate than Independents, both in the Democratic primaries, and in the 2004 general election. In the primary states so far, Independents have made up 19.2% of the voters, while Moderates made up more than twice that (39.1%). In the 2004 general election, 26% of voters were Independents, while 45% identified themselves as Moderates.

Yet, ever since the Iowa caucuses, the media has been obsessed with Obama's success among "Independent" voters, and completely ignored the "Moderate" demographic.

The contrast between each candidate's "Independent" support, and their "Moderate" support, is striking. While Obama received 54.2% of the "Independents" in primary states, he received only 46.7% of the "Moderate" vote. But despite far less "Independent" support (36.2%) than Obama, Clinton managed to outpoll him (47.8%) by a small margin (1.1%) among moderates.

And while Obama maintains his strong advantage among Independents in crucial swing states, Clinton (with 49.2%) does even better relative to Obama (with 44.5%) among "Moderates" in the those states. In "safe states" Obama holds a slight lead over Clinton among "Moderates" (Obama 47.8%, Clinton 47.1%).

SKEWED PERCEPTIONS: Super Tuesday and After
But even these numbers don't tell the true story of Clinton's superior appeal to moderates, because of a gross miscalculation by the Clinton campaign. It is clear that the Clinton camp was too focused on November, and ensuring that voter in key states knew who she was, and what she stood for, and depended upon momentum from convincing wins in those states to do well in the states that voted the next Saturday and Tuesday and from there to victory. The Obama campaign, recognizing that this left her vulnerable in "safe" Republican states, put much more effort into those states, and the states that followed Super Tuesday.

The contrast between what happened on Super Tuesday and what happened since then is staggering, and provides stark evidence that it was the campaign strategy, and not any better intrinsic appeal of Obama over Clinton, that has lead to Obama's recent string of stunning victories.

Total Popular Vote Margins
Super Tuesday: Clinton +1.5%
Post Super Tuesday—Obama +22.5% (with DC +23.5%)

Total Popular Vote Margins in Swing States
Super Tuesday: Clinton +9.1% (without Arkansas ). 2 +4.2%)
Post Super Tuesday—Obama +22.4%

All Moderate Voters Margins
Super Tuesday: Clinton +4.5%
Post Super Tuesday—Obama +24.4 %

Moderate Voters in Swing States Margins
Super Tuesday: Clinton +11.5% (without Arkansas +5.9%)
Post Super Tuesday—Obama +23.8%

Super Tuesday State: AL, AR, AZ, CA, CT, DE, GA, IL, MA, MO, NJ, NM, NY,OK, TN, UT
Post-Super Tuesday states: LA, MD,VA,WI
Super Tuesday Swing States: AR, DE, MO, NJ, NM
Post-Super Tuesday swing states: VA, WI)

(Clinton did manage to do better in Wisconsin among moderates (Obama +17%) than she did in the three primary states in the week immediate after Super-Tuesday (Obama +28.3%), which suggests that her campaign is back on track, but still playing catch-up.)

HEAD TO HEAD—Where Both Candidates Ran Credible Campaigns
The data shows that Clinton did far better in the more competitive states (where the popular vote margin was less than 15%, Obama did better in less competitive states (popular vote margin 15% and above) and that Obama's numbers are heavily influenced by "blowout states"--states where the popular vote margin was greater than 28%.

Moderate Vote Margins
More Competitive States—Clinton +9.5%
Less Competitive States—Obama +4.7%
Blowout States—Obama +27.1% (without "home states" of IL and AR, +34.7)

More Competitive States: NH, AL, AZ, CA, CT, DE, MO, NJ, NM, TN
Less Competitive States: SC, AR, GA, IL, MA, NY, OK, UT, LA. MD, VA, WI
Blowout States: SC, AR, GA, IL, VA)

By focusing solely on Obama's strength among "Independent" voters, while ignoring Clinton's strengths among "Moderate" voters in competitive and crucial swing states, the media is skewing the coverage to make Obama look far more electable relative to Clinton than he actually is. Obama's lead in states, delegates, and the popular vote rest primarily among "Independent" voters, those states that are considered "safe" for one party or the other in the general election, and the post Super-Tuesday states where a strategic failure of the Clinton campaign resulted in non-competitive races. Clinton's strengths are among "Moderate" voters in swing states and competitive states.

The general election is far more like Super Tuesday than the long primary campaign—it's a single day event, where every state is winner take all, and candidates devote their resources to ensuring victory in competitive (not just swing) states, while paying little attention to non-competitive states. On Super Tuesday, the Clinton campaign showed that it knew what it takes to win a general election by outperforming Obama in competitive states by appealing to the moderate voters who make up the nearly 40% of voters in a general election. So far, while the Obama campaign has demonstrated a mastery of primary campaign strategy, it has yet to show that it knows what to do to win in November.

As with the rest of his campaign, the basic "electability" argument for Obama rests with two words: "trust me." We are asked to "trust" that Obama understands how the Presidency operates. We are asked to "trust" that Obama will perform adequately in the face of a political crisis--or an international crisis. We are asked to "trust" that Obama can do what he promises, and create the necessary popular coalitions he says he will build in order to overcome special interests, and achieve meaningful change.

And we are asked to "trust" that Obama will emerge victorious over the "maverick", "appeals to Independents" GOP nominee, John McCain, and do so after a months-long, relentless onslaught by the right-wing smear machine.

At this point, it does not appear that either Hillary Clinton or Barack Obama will gain enough pledged delegates to lock in a first ballot nomination, and the super-delegates will determine who will be the nominee. The Obama campaign insistence that the "most pledged delegates and/or most voters and/or most states" entitles him to the nomination reduces a complex decision involving who would make the best candidate and the best President to one that could be made by a simpleton. But the issues and problems that the next President will face will be difficult, complex, and multifaceted in much the same way that the question faced by the super delegates--who to choose as the Democratic candidate--should be. And rather than insisting that the super-delegates refrain from using their judgment based on years of experience, and saying "Trust me, I know what I'm doing", Obama should be insisting that the super delegates consider all the relevant factors before making their decisions.

The purpose of this analysis is not to argue that Hillary Clinton would make the best general election candidate, rather its purpose is to demonstrate that the assumptions underlying the arguments who say that Barack Obama is a better choice for the general election are dubious at best, and that there are numerous legitimate reasons to view Clinton as the better choice.

Neither primary results, nor current polling, should be major factors in the decision that the super-delegates are likely to face at the Democratic Convention that is more than six months away. We are just beginning to see the effort by well-funded right-wing smear groups to define Barack Obama in highly negative terms—and in 2004 we saw how easy for these groups to redefine John Kerry. While its entirely possible that Obama will be able to transcend the coming onslaught of attacks on his character and define himself on his own terms, it's also entirely possible that by September he will be a severely damaged candidate who would face an steep uphill battle to win in November.

One thing is certain—calls for Hillary Clinton to withdraw are dangerously premature. Not only does Clinton remain a highly viable candidate, continuing the campaign will present Obama with the opportunity to introduce himself on his own terms to voters in a number of states that will be important in November—not just Pennsylvania, but North Carolina, Indiana, West Virginia, Kentucky, Oregon, and South Dakota. Keeping the contest alive also keeps the media focus on the Democratic candidates and their issues while paying little attention to John McCain—and as was seen with both John Edward and Rudy Giuliani, when the media doesn't focus on a candidate, that candidate declines in the polls.

APPENDIX—Unpolled Primary States and Caucus States
This analysis pretty much ignores the results from two states (DC and WA) that held primary elections because there was no exit polling data from those states, and thus no way to determine how "Independent" and "Moderate" voters voted in those states. However, the data tables that accompany this article do include totals, based on using the averages for "Independents" and "Moderates" from polled primaries. The differences turn out to be insignificant (doubtless because averages are being used).

Nor were caucus states used in this analysis—for very good reasons. Except for Iowa and Nevada, there is no exit polling data from the caucus states. There are huge variances in how "totals" are reported by states, with some states reporting "raw" participation numbers, and others reporting on the number of delegates election to state/county/district conventions—and this also creates problems. (For instance, twice as many people were estimated to have participated in caucuses in Iowa than in Nevada. But the total delegates "reported" as elected in Nevada is over 10,000, the delegates "reported" from Iowa are only about 2,500 hundred.)

Finally, as has just been amply demonstrated in Washington State, primary results do not reflect overall voter sentiment. (Washington's Democratic primary was a "beauty contest" and no delegates were at stake.) Obama's margin over Clinton in delegates in the Washington caucus was an astonishing 37%. His margin over Clinton on results reported the night of the "beauty contest" (which consisted primarily of people who voted in the polls that day) was only 3.1%. But Washington also has "mail in" voting, and the bulk of these votes had not been included in the totals reported on election night. As of 2/24/08, Obama had a 13 point lead in mail-in votes, resulting in the overall margin over Clinton to rise to 5.1%.

Even more striking are the Republican results in Washington, where delegates were at stake in both the caucuses and primaries. In the caucuses, John McCain got only 25% of the delegates, and Ron Paul got 22%. As of 2/25/08 (i.e. including mail in votes) McCain has 49.44% of the popular vote, and Ron Paul has only 7.59%

Nevertheless, the accompanying data tables do include totals based on entrance polling results from Iowa and Nevada, using estimates of total participants from local papers, and dividing the "raw" vote between the candidates based on percentages of delegates reported as elected. As with the inclusion of the unpolled primary states, the impact on the results as reported in the main body of this piece is negligible at best.

End notes
1. Rasmussen’s list of "safe" states includes some states that will probably be contested (e.g. CA, TN) by both parties for strategic reasons, and thus are different from "battleground" states. The states listed as safe (and their electoral college votes) that were part of the primary data set, but excluded from the "swing state" data set areCalifornia (55), Connecticut (7), Illinois (21), Maryland (10), Massachusetts (12), New York (31), Alabama (9), Arizona (10), Georgia (15), Louisiana (9), Oklahoma (7), South Carolina (8), Tennessee (11), and Utah.

2 "Without Arkansas" numbers are included for comparison purposes, because the apparent lack of any serious effort by Obama in Arkansas due to Clinton’s "home state" advantage skews the Super Tuesday data in Clinton’s favor. In Arkansas Clinton’s popular vote margin was 44.1%, and her margin among "Moderates" was 46%--the highest margins by far in any polled primary state. Obama’s abysmal performance in Arkansas reinforces the point that when a campaign makes little effort in a state, the results do not reflect what would happen in a state where both campaigns had made a serious effort.

No votes yet


intranets's picture
Submitted by intranets on

I find one aspect troubling. Your % of "independent" and "moderate" are based upon exit polling, correct? I worry that those numbers are complete fabrication by exit pollsters to fudge the actual outcome to match the polling data. So, if one candidate does better in actual race totals, and they associate more moderate voter support for that candidate, they will fudge the number of moderate % of voters in the exit poll until the exit poll matches reality.

So the breakdown of number of moderate voters as decided by exit polling is a complete fiction based upon who they happened to poll and how far the exit data was off from the election results.

Submitted by Paul_Lukasiak on

what actually happens with exit polls is that the raw data is massaged to fit into pre-existing demographic expectations before its first released.

So, if pre-election polling says that 54% of the voters will be female, and 58% of the people interviewed by exit pollers wind up being female, they'll make an adjustment when the exit poll is first released, and if the totals coming reflect that 58% of the voters were women, they make a further adjustment.

They'll also make adjustments based on turnout in various precincts. If black turnout is higher than expected in heavily black precincts, they'll adjust the numbers accordingly.

The problem is that while exit polls collect far more responses than pre-election polling, the data is collected in a limited number of precincts that have "predictive" in the past. But 'past performance is no guarantee of future profits', as they say -- and that's why exit polls get adjusted.

Submitted by Paul_Lukasiak on

The Obama campaign, and the media, hype the "Independent" vote using the same exit polling data that I use. What I'm doing is using the exact same data to say "wait a second... aren't you missing something here?"

There is no question that exit polling data is not accurate on may questions -- I mean, when 78% of people tell you race is not an issue, you'd expect there'd not be major differences in how white and blacks voted -- but there are. And its not that these people are necessarily "lying" either, subconscious prejudices play a big role in how people and ideas get perceived, and people don't even know that their decision has been influenced by race.

But the categories I'm using are far less subjective than those kinds of questions...

Davidson's picture
Submitted by Davidson on

Something I can't help but think of: these numbers are Obama at his peak performance level, considering that he's not only getting a total pass by the media (including liberal blogs, radio, etc.), but is also unabashedly glorified--and he's running against the demonized Clinton. And yet she still remains a contender in TX, OH. How then can he compete with McCain in the GE without the media love/Clinton hate?

I doubt superdelegates will give a damn about all this. They'll go with Obama out of fear of any backlash.

Submitted by Paul_Lukasiak on

Davidson, that is why I think its essential that Clinton stay in the race, and deny Obama a lock on the nomination. If she's doing badly, I think she should acknowledge that Obama is the likely nominee, and if come September his poll numbers have stayed high, she'll release her delegates. But she's still going to campaign, because a lot can happen in six months, and the Party (and not Obama and his supporters) should be able to bail out if the worst happens over the summer.

CognitiveDissonance's picture
Submitted by CognitiveDissonance on

Paul, great eye-opening analysis - again! I would love to see the contact data for every super delegate so I could send them the links to both your illuminating posts. This is critical information.

Submitted by Paul_Lukasiak on

I was about to thank you, but say that I'm sure that the Clinton campaign already has all this data and is using these argument...

But then I realized... ya' know, I haven't heard anyone else making these arguments. I mean, I wouldn't be spending so much time doing this crap if anyone else was even mentioning moderate voters, or the way that the post-Super Tuesday debacle has totally warped perceptions of the campaign as a whole, or how Hillary does better in states that will be crucial to win in November.

The closest you hear is that idiot Mark Penn saying that Obama hasn't won any important states -- and that's so not true it discredits what is true.

So, Thank you for your kind words!

bringiton's picture
Submitted by bringiton on

at Wikipedia, Wiki is not good for analysis but an excellent repository of this kind of baseline information.

No contact info, but a decent place to start working on the campaign.

bringiton's picture
Submitted by bringiton on

Fantastic piece of work, Paul; your labor is as always much appreciated.

One semi-trivial detail: At the linked-to “Count WHOSE Vote 2.0” site, right up at the top, is the sentence “Data tables and an explanation of the methodology employed can be found Here” where Here is a link. But that link actually takes one to your excellent but unrelated “AWOL” site instead of down to tables in the “Count WHOSE Vote” document as you doubtless intended.

And now, about Freud. Strikes me that the media’s relative obsession with Independent over Moderate is a gender issue, the gender value of the words themselves. Independent is a strong, forceful word, a political stance that sticks out (so to speak) and connotes masculinity, while Moderate is soft, squishy, gentle and more feminine. As everyone from Russert to Limbaugh knows, masculine opinions are worthwhile and must be paid attention to while female opinions are fickle and difficult for manly men to understand so why bother?

Sometimes a cigar really is a phallic symbol.

Submitted by lambert on

It goes to a copy of the article on Paul's site that includes the tables.

[x] Any (D) in the general. [ ] Any mullah-sucking billionaire-teabagging torture-loving pus-encrusted spawn of Cthulhu, bless his (R) heart.

bringiton's picture
Submitted by bringiton on

The one I meant is not at Corrente but at this site:

which I believe belongs to Paul Lukasiak. Thought he'd like to fix it.

Click on the link at the end of this top-of-the-page sentence:

"Data tables and an explanation of the methodology employed can be found Here"

intranets's picture
Submitted by intranets on


i think it's a good analysis and you are right that EP data is what the media and campaigns use and it isn't wrong to use it. I am saying that EP data is total crap. Once upon a time it was accurate enough to declare elections fraudulent, but now it is somehow acceptable to have statistically impossible swings in exit poll data from either pre-election polls OR the actual vote tally.

They don't adjust EP data to necessarily fit their expected vote pool like the % women voter example you gave. They literally just take their responses and scale them to fit the final tally. So if women said they voted 80% for HRC and they have 40% of the EP as women, they scale the "final" number of women until is matches the final numbers. Then they declare that more women voters turned out, or whatever BS they make up.

There is statistics and sampling and their only excuse now for why the EP is so far off (usually only in swing states) is typically the Bradley effect or the "reluctant respondent" or some other complete made up crap.

Explain to me why the US uses exit polling in foreign countries and it is usually very accurate. Explain to me why citizen exit polling is usually pretty good. All we know for sure is that exit pollsters are doing something fundamentally wrong or people lie to pollsters or there is something wrong with election results.

Submitted by Paul_Lukasiak on

I'll fix it tomorrow...the data tables are at the end of the piece, btw.

BDBlue's picture
Submitted by BDBlue on

In addition to Obama's ability to attract moderates, I'm worried about his maintaining the Democratic base, particularly women.

A lot of us are angry. As we have documented here. And I'm not sure everyone is going to decide to vote for Sen. Periodically Women Use Their Claws to Drink Tea. If a fair number just sit on their hands, he's got real problems. I admit that while I can in no way see myself doing anything that helps John McCain become president, I'm torn about voting for a man so quick to use misogyny and a party so quick to presume that women have no choice but to fall in line anyway.

Here's a post from Melissa McEwan, who seems like she's about had it with Mr. Unity's sexist crap, in which she argues that feminists don't necessarily have to fall in line, that there are other feminist choices. I find it interesting in that she mentions a lot of things that folks have mentioned here when discussing the issues. She doesn't argue that women should vote against Obama, only that it's not feminists' fault if the Left is divided as a result of Obama and other Democrats' sexism, it's ultimately the fault of those who use such divisive language and tactics (in other words Obama and his surrogates and supporters). Her conclusion:

Shaker CE said in comments yesterday, "Knowing that the alternative is worse actually makes it harder for me; it just reinforces that sense I often get from some Dems, including Sen. Obama in this cycle, that they think they can do whatever the fuck they like to me, because I don't have any other option. The worst part? They're right."

They are right, unless we go somewhere else. This isn't a treatise to convince anyone to do so—but it's an explanation for why a feminist might, why it's a legitimate choice, and why, if that means the Left isn't a picture of harmony, it's not our fault.

The reason the Left is discordant isn't because of our standards; it's because there are so many bigots with no benchmark for success but winning—even at our expense.


Vastleft, she, too had difficulty finding "women's issues" on Obama's site and has noted his mushy language on reproductive rights (women "generally" should get to choose whether to have an abortion, wtf?) see

bringiton's picture
Submitted by bringiton on

Must be the pastel background that makes Laniers' more valuable, plus a week and a half later to get out. There is no justice, eh?

Just tried that link and it still sends me to AWOL, maybe my fingers are the problem.

intranets's picture
Submitted by intranets on

Did anyone look over Cal Lanier's numbers? I call total BS. Which is why it is just stupid to use EP data.

There appears to be a major screwup with the tables and maybe not the EP data, but come on...
8-Jan NH
White Democrats 54%
Black Democrats 43%

What? First of all the numbers don't add up so maybe he sorted some table and the columns got out of whack, but really? 43% of voters in NEW HAMPSHIRE are black? Talk about your voter fraud.

Ok so maybe I'm reading the heading wrong and it is saying % Clinton support among White Democrats is 54%... But really, 200,000 black Obama voters in NH??! Again, let's review the census data for NH.
But this doesn't even match the other columns.

Then you have a -10%??!!! WTF.

This is linked to by Slate, and I'm supposed to trust the magic numbers at the bottom of the table? Stupid internet and authoritative blog linking..

Submitted by Paul_Lukasiak on

I see where the problem is...

The NH black democrat numbers are the same as those for Alabama. (which suggests that at some point, he screwed up a data sort).

His numbers for Michigan are also very hinky. I mean, there were 594K votes in michigan, He says that 57% of voters there were white democrats (which is an extrapolated percentage based on 79% democratic voters times 72% white voters --- not something found in the exit polls that I've seen, and given that the african american vote has been much more "democratic" than the white vote, probably completely off base) But his white Democratic totals for clinton and obama add up to only 29% of the total vote, 39% of the extrapolated white vote, and 35% of the extrapolated democratic vote.

Submitted by lambert on


[x] Any (D) in the general. [ ] Any mullah-sucking billionaire-teabagging torture-loving pus-encrusted spawn of Cthulhu, bless his (R) heart.

Submitted by Paul_Lukasiak on

Not in the aggregate, where "all Democrats" are concerned, as far as I can tell. But I think his "category" numbers are highly suspicious, given his method of extrapolating percentages when he doesn't have exit poll results in certain categories (like a specific "race/party" breakdown).

Submitted by lambert on

although butchering the subtotals might speak to the credibility of the total?

[x] Any (D) in the general. [ ] Any mullah-sucking billionaire-teabagging torture-loving pus-encrusted spawn of Cthulhu, bless his (R) heart.

intranets's picture
Submitted by intranets on

I wouldn't count the spreadsheet being right considering the glaring errors, but the totals might be. At any rate, if you take off FL & MI Clinton popular lead is >200,000 votes out of 14,000,000

Submitted by Paul_Lukasiak on

if all you care about is the big total, is to use my numbers from Count Whose Vote 1, and if you must refer to Lanier, just say that his conclusions confirm mine! ;-)