# Pollster.com

## Articles and Analysis

### McDonald: Democratic Dissention: An Artifact of Survey Methodology?

Today's guest pollster contribution comes from Michael P. McDonald, an Associate Professor of Government and Politics in the Department of Public and International Affairs at George Mason University and a Non-Resident Senior Fellow at the Brookings Institution.

A media storyline surrounding the Democratic convention is how a sizable number of Hillary Clinton supporters are backing John McCain over Barack Obama. A recent CNN/ORC poll provides grist for the mill. Twenty-seven percent of self-identified Clinton supporters are reported backing John McCain, an increase from 16% in a similar June survey.

Yet, there are indications that something is amiss in this survey. CNN reports they interviewed 1,023 adults. The organization does not report the sub-sample size of Democrats who support Clinton, but they do provide a margin of error of this sub-sample from which we can infer the number of Clinton supporters. The reported margin of error for Democrats who support Clinton is 7.5 percentage points, which is equivalent to 171 persons assuming a simple random sample. That is 16.7% of all adults in the survey, which when applied to my 2006 voting-age population estimate of 227 million persons means that there are 38 million self-identified Clinton supporters among Democrats in the CNN/ORC poll (with a 95% confidence interval between 20.9 and 54.9 million persons).

As one might recall, Clinton received 18 million votes in the primaries. If she had received 38 million votes, she would be accepting the Democratic Party's nomination on Thursday.

The question arises, who are these 20 million or so self-identified Democrats who support Clinton who did not participate in the primaries? It is difficult to tell without analyzing the survey in depth. While there are many reasonable explanations for the discrepancy between the election and survey results, a plausible explanation consistent with the large percentage of self-identified Clinton supporters who report supporting McCain in a two-way contest against Obama is that the CNN/ORC questionnaire is worded in such a manner that elicits persons who self-report supporting McCain to report that they are a Democrat who supports Clinton for the party's nomination.

The implication is obvious: if these surveys that purport to measure Clinton supporters who will vote for McCain actually measure McCain supporters who would like to see Clinton as the Democratic nominee, the media storyline of Democratic dissention quickly unravels.

Patrick:

Interesting column, but you don't need to know the specific methodologies of these polls to recognize that many Clinton supporters are currently not planning to vote for Obama. The fact that McCain and Obama are running dead even in the national polls (while every generic Democrat vs. Republican poll shows the Dem leading by about 15 pts) already demonstrates it. Personally, I know at least 10 lifelong Democrats who are voting for a Republican (or no one) for the first time in their lives. So there must be millions of these voters all over the country.

As for the Hillary supporters who didn't participate in the primaries, I am one myself. I live in Washington State and was out of the country on the day our caucuses occurred. I did vote in our primary (in which far more people participated and Obama beat Clinton by 3%), but couldn't participate in the caucus (in which far fewer people participated, but decided the delegates, and Obama won by like 30%). So like millions of voters, I was disenchfranchised by an arbitrary voting system. My 78 year old mother, a strong Clinton supporter, also didn't participate because she was too intimidated and knew she could be outnumbered by Obama supporters. Gives you a good idea of how fair the caucus system (which gave Obama the nomination) is compared to primaries.

____________________

mtorrens:

Patrick:
Although interesting, anecdotes (e.g. you know 10 people...) are not a substitute for data. Unless CNN provides more information on its methodology, or additional quality polling backs it up, I think the jury on this narrative (at a significant national scale).

____________________

mtorrens:

The jury is still out...

____________________

douglasfactors:

Patrick, if your mother failed to participate in the caucus, then how exactly could the Obama supporters have intimidated her?

____________________

buckeyepoliticalscientist:

The subsample size of Clinton supporters planning to vote for McCain is actually n=134, as determined by reverse engineering the margin of error (which we know) for a simple random sample estimate (1.96 * standard error) . But Patrick is right--the story talks about people who "support" Clinton, planning to vote McCain. Supporting is not the same as voting in the primaries. The exceptional turnout of this year's primaries notwithstanding, most eligible voters do not participate in primaries. So, once this is taken into account, the 27% figure is entirely plausible, and has been recently confirmed by other polls (see The Columbus Dispatch poll in Ohio and the Buckeye Poll from the University of Akron).

____________________

@buckeyepoliticalscientist:

The formula for the Margin of Error (MoE) is:

MoE = 1.96*(p*(1-p)/N)^.5

Usually, when pollsters report the MoE, they do so for p = .5, in which case ((.5)(1-.5))^.5 = .5 and the formula simplifies to:

MoE = 1.96*.5/((N)^.5)

(As a whole other teaching point, the MoE gets smaller for smaller percentages, but I digress...)

Some rearrangement of furniture gets you:

N = (1.96*.5/MoE)^2

Using this formula, if MoE = .075, N = 171. (You can verify yourself, and I encourge you to plug in MoE = .04 to your calculation, which should return N = 600).

As for the other comments, my point is that the percentage of Clinton supporters in the CNN/ORC poll is not reflective of the percentage of persons who actually voted for her. There is a well-established line of survey research that finds people misreport voting for winners (the classic example is the overwhleming percentage of people who reported voting for Pres. Kennedy following his assassination when he won the 1960 popular vote by 0.12 percentage points). Now Clinton did not win, but something is going on here with these self-identified Clinton supporters that is causing the polls to overestimate her actual votes, even if her hypothetical support in caucus states is factored in.

Whatever is causing this result may be correlated with support for McCain among self-identified Clinton supporters. The challenge to the pollsters who are running these polls is to explan how they are arriving at their estimates and conclusions, particularly since these estimates are driving public discourse. And just because other polls are finding the same pattern does not make it true. There are all sorts of survey methodology biases that plague all polls in a similar direction; for example, vote over-report bias in post-election polls.

____________________

Mark Lindeman:

Michael, I don't see where you got at the distinction between supporting Clinton and voting for her.

Just to fill in some numbers: I think the CNN report says that the MoE for registered Dems is 4.5 points, so N(reg Dems) = 474 or so; and the MoE for (registered?) Dems for Clinton is 7.5 so N(Dems for Clinton) = 171 as you said. I.e., about 36% of registered Democrats still would prefer Clinton to Obama -- or something like that, in lieu of actual question wordings. Take a fraction of that, and you've got around 10% of registered Democrats who would vote for McCain. That seems plausible.

Now, presumably some of these folks not only did not vote for Clinton, but don't really care much about her. Some may indeed be Democrats In Name Only. So maybe you are reasoning that people who didn't bother to vote for her when given the opportunity [factoring in hypothetical caucus turnout] shouldn't really be considered her "supporters." In terms of the prevailing PUMA narrative, that's probably true. But framing this as an "overestimate" of Clinton's "actual votes" doesn't seem right to me.

____________________

Mark, the numbers show that about half of these "supporters" were not actual Clinton "voters", so we are in agreement there. And I agree that 10% of Democrats not supporting their party's candidate is a reasonable estimate from past election polling.

I'm not denying these people do not support Obama. What I suspect is that "Clinton supporter" is proxying for some other reason why these people will not vote for Obama. If true, then these survey results have a different substantive interpretation than how they are being portrayed: a deep seated anger towards Obama because Clinton lost the nomination. These people are not really passionate Clinton supporters; they didn't feel strongly enough to vote for her. If this is the case, why would anything she or Obama say about party unity at the convention change who they are voting for?

____________________

John:

I am sorry Michael, I think I did you a disservice in a previous thread, I didn't realise that pollsters always picked p to be 0.5. Your numbers are, of course, right. Do you know if they do this with all sub-groups even ones (such as AA or evangelical voters) which they know beforehand will have a very high p?
Anyway, I completely agree with you that CNN should clarify what question they used to define who were Clinton supporters (although it might make it less of a story), not to mention releasing their other crosstabs, (45%+ of Adults are registered democrats??)

____________________

Mark Lindeman:

Michael, OK, we agree: the survey result is being coopted to fit the PUMA narrative when it probably means something else entirely (or, possibly, nothing in particular). What we apparently can say about these people is that they self-identify as Democrats (or as registered Democrats -- it isn't clear) and, loosely speaking, they prefer both Clinton and McCain to Obama.

This is freehand, but pooling the 2000-2006 GSS data, 9.3% of self-identified ("usually think of yourself as a") Democrats said they thought there should be laws against marriages between blacks and whites. (Coincidentally, the percentage was the same for "strong" and "not strong" Democrats.) So it's not nutty to hypothesize that racism could account for most of the supposed PUMA effect.

However, there are lots of problems with that line of analysis, beyond being scrawled in crayon on the back of an envelope. Maybe the biggest is that it implies that that the proportion of Dems who would vote for McCain because of Obama's race is fixed. That may be a decent first approximation -- I don't know. Probably there are millions of likely voters who doubt that a black man in general (or Obama in particular) cares about their problems, but are cross-pressured about whether to vote for him. So, when Holland says that the proportion of "Clinton Democrats" for McCain has increased since June, he might be tapping people like this who have heard little about Obama lately except that he was loved in Europe and had a good time in Hawaii. (He could also be chasing survey noise, but hey, work with me here.) Some of them, regardless of their beliefs about race, are persuadable. Some, realistically, are not.

One implication is that it probably doesn't matter much that Hillary Clinton said she is supporting Barack Obama, but the reasons that the Clintons gave for supporting Obama might (or might not) be convincing.

OK, hours at the Department of Wild Speculation are drawing to a close for this morning. Thanks for the thought-provoking posts!

____________________

RS:

Could it be that CNN is over-sampling "Clinton Democrats" to reduce the MOE? On the other hand, I guess not - CNN doesn't know how many CDs are there in the first place.

I think someone wrote an analysis piece on Pollster.com earlier about how the CNN analysis was flawed, using the change in CD-support for Obama between June and August to explain the change in Obama's support between July and August. So the CNN commentary is already suspect...

As for caucuses disenfranchising Democrats - 13 of the 50 states had caucuses, including WA. Together, these states have a population of 33 million - 11% of the US population. The only state where we may have a comparable primary result is WA, where 700k voted - 11% of WA population (not voters, not Democrats). That seems almost in line with the primary voters/total population in the primary states, about 13-14% (with perhaps some depressed turnout as the WA primary did not count).

So it does not seem unreasonable to me that of the 33 million in caucus states, maybe 4-5 million might have voted in a primary (if caucuses did not exist). And considering Obama won the WA primary by 5.5% anyway, it seems unlikely Senator Clinton would have won a larger share of the "caucus-replacing-primaries" to offset the eventual result. But more to the point, these caucus states do not account for the missing 20 million votes.

____________________

Ciccina:

"What I suspect is that "Clinton supporter" is proxying for some other reason why these people will not vote for Obama."

"So it's not nutty to hypothesize that racism could account for most of the supposed PUMA effect."

Insulting. Baseless. Ridiculous.

You are grasping for numbers to back up your narrative that PUMAs are racist. Calling PUMAs racist is the quickest way to say their (our) concerns are not valid, that we have no "real" issues, just hate.

"the survey result is being coopted to fit the PUMA narrative when it probably means something else entirely (or, possibly, nothing in particular). "

In other words, if you don't like the story the data is telling you, feel free to disregard the findings and make up something you'd prefer to hear.

Nice one, guys. This one should have stayed on the back of the envelope.

____________________

RS:

Ummm... The question is, how many of the "Clinton Democrats for McCain" are PUMAs, and how many are racist? And is there a intersection of the two sets, or is this a null subset?
I think particularly toward the end of the primaries (WV, KY...) there was a good chunk of White voters (~22% in WV) who said race was an important factor in their choice between Obama and Clinton, and most of them (>80%) voted for Clinton. Now, given the choice between Obama and McCain, it is reasonable to assume these voters might feel the same way, and vote McCain.
Now, are these voters PUMAs? Or are PUMAs just a tiny, but vocal minority of Democrats? If the latter, then Mark Lindeman could be right - most of the so-called PUMA effect is in reality a product of racism.

But this does not imply that PUMAs are racist. Just that there aren't that many PUMAs as people think. I think that was Mark's point.

____________________

buckeyepoliticalscientist:

Well Michael, ya got me. I didn't use p=.5 to obtain the "maximum" sampling error. TouchÃ©.

I am aware of the literature on erroneously reporting a vote for the winner. Gerald Wright wrote a good piece on that years ago in AJPS. Even if that's what's going on here, we also have to be willing to accept that a Bradley-Wilder effect might be happening too, in which case John McCain's support among supporters of Clinton might actually be understated. But something else is still wrong.

The CNN article says that:

2) the MOE for "all voters" (I assume likely voters?) is 3.5%
3) it is 4.5% for "registered Democrats" (I assume self-identified Democrats, registered to vote in their state)
4) it is 7.5% for "registered Democrats who want Clinton as nominee"

Using the correct equation for MOE:

n=(1.96*(0.5/MOE))^2

we obtain:

n=784 voters
n=474 registered Democrats
n=171 Dems who support Clinton

Mark (in the post above) is correct.

This means there are 239 non-voters (1023-784=239) in the sample. That sounds about right. It also means there are 310 registered Republicans plus Independents (784 voters-474 registered Dems=310), meaning there are 60% (474/784=.6) registered voters who are Democrats in the sample, much greater than their actual number in the population.

1 of 4 things is going on here:

1) the poll was not a SRS

or

2) Democrats were over-sampled (as RS stated above)

or

3) the weights applied to the estimates are screwing up our back-of-the-envelop analyses

or

4) a combination of all these factors,

in which case all this is much mathematical ado over nothing because we don't know the sample design, and thus which formula to use and so we are only guessing at the number of Clinton supporters who plan to vote for McCain.

I do not think it is just the media getting carried away with the PUMA narrative. Look at the Buckeye Poll from the University of Akron. It's a panel study, so the same people were reinterviewed after they were sampled around the time of the Ohio primary. Here, the weakness is we don't have the question wording, but it is instructive that 29% of Ohio Clinton primary voters are supporting McCain now. And let's not forget Ohio's well-earned reputation as an electoral barometer. This is strikingly similar to the 27% reported by the CNN poll over which we are debating.

http://www.uakron.edu/bliss/docs/AkronBuckeyePollSummer_08_.pdf

It is unlikely (though not impossible) that different sample designs, such as UAkron, CNN/ORC, The Columbus Dispatch, NBC/Wall Street Journal, would yield similar results by chance alone, absent a real, underlying dynamic.

____________________

Ciccina:

RS,

"I think particularly toward the end of the primaries (WV, KY...) there was a good chunk of White voters (~22% in WV) who said race was an important factor in their choice between Obama and Clinton, and most of them (>80%) voted for Clinton."

Saying race is important, or is an important factor in your voting decision, and then voting for the white candidate does NOT mean the respondent is racist. You are assuming way too much.

If a female voter indicated gender was an important factor in her voting decision, and then indicated she supported Obama, does it follow that she's a sexist? Of course not. If a male voter said gender was important and then voted for Obama, is he a sexist? Nobody here ever so much as whispered the possibility. So why do you (plural) keep going back to the formulation that any voter who indicates they consider race important but didn't vote for Obama is a racist? The data you have simply does not back that up.

Look at what Mark L. is doing with the intermarriage stat: "This is freehand, but pooling the 2000-2006 GSS data, 9.3% of self-identified ("usually think of yourself as a") Democrats said they thought there should be laws against marriages between blacks and whites. (Coincidentally, the percentage was the same for "strong" and "not strong" Democrats.) So it's not nutty to hypothesize that racism could account for most of the supposed PUMA effect."

I gather Mark L. is saying that since 9.3% of self-identified Dems are racist, its not nutty to hypothesize that there could be a block of racist Dems who won't support Obama. I guess it didn't occur to him that some chunk of the 9.3% who disapprove of interracial marriage may be African American themselves. What happens to the narrative now? Are they still the racists who won't support Obama?

Its fun to take big speculative leaps sometimes, but when you are throwing around labels like "racist" one should be more careful.

____________________

RS:

Ciccina:
Are you serious?

This is just too rich. I suppose it hurts a lot to find out one is part of a fast-dwindling species.

____________________

Mark Lindeman:

Ciccina, sorry to have missed the excitement. Please note that I didn't cite any data about people who said that race was "important" or a factor in their voting decisions, but rather about people who say that it should be illegal for blacks to marry whites. Whether someone who says that must be "racist" may be open to debate, but it's not at all comparable to "saying race is important."

Your speculations about what I want to believe are largely off target, and at any rate irrelevant. How many Democrats would say they support McCain over Obama because they are angry at how Clinton was treated is, in principle, an empirical question. Believe it or not, I don't have a strong prior expectation about the answer. Omniscience is not my gig. Better questions might help.

____________________

Ciccina:

Mark L, I did not say that you cited any data about voters who indicate race as an important factor. RS used that information in his comment at 2:13pm.

He wrote: " think particularly toward the end of the primaries (WV, KY...) there was a good chunk of White voters (~22% in WV) who said race was an important factor in their choice between Obama and Clinton, and most of them (>80%) voted for Clinton." From that I gathered he was drawing an inference about those voters' attitude toward race. My point is that one shouldn't, because the questions are ambiguous.

With regard to: "So it's not nutty to hypothesize that racism could account for most of the supposed PUMA effect" I get that you are positing the racist voters and the PUMA voters as two distinct sets - that racists are being miscounted as PUMAs - but I fail to see why one should hypothesize about racism as a motivating factor at all given the meager amount of information available. Particularly when doing so feeds into such a divisive narrative.

____________________

Mark Lindeman:

Ciccina, OK, I see how I elided your posts. I'm still pretty puzzled by the first one.

Certainly when people who say that race was an important factor in their decision voted for Clinton, we can't infer how many of them are racist. That's old news.

I wouldn't necessarily say that "racists are being miscounted as PUMAs"; I don't know that anyone has actually tried to count PUMAs. "Clinton supporter for McCain" just doesn't do it. Some may be PUMAs, some may be racists, some may be non-racists who think Obama is kind of out there, some may be other things.

Are you saying that I was wrong to point out that some Democrats are racist, because that "feeds into such a divisive narrative," but the PUMA narrative is not divisive? I'm lost.

____________________

Ciccina:

Mark, I don't think its a matter of pointing out that some Dems are racist. Your comment went from "What we apparently can say about these people is that they self-identify as Democrats (or as registered Democrats -- it isn't clear) and, loosely speaking, they prefer both Clinton and McCain to Obama" directly to a measure of racism among self-identified Democrats.

Did you not mean to imply that racism was the most likely motive for voters to say they preferred Clinton or McCain over Obama?

____________________

RS:

Why can't I make the inference that people who say race was important or the most important in making their decision are racist? I haven't seen that addressed much or at all to dismiss it as "old news." But since Mark L says so - and he is usually sensible - I assume it must be so, and request a post or some such to clarify.

There are, however, people who say "race of candidate was most important" - and of the 8% of WVa-nians who responded thusly, 86% voted for Clinton. Considering Whites were 96% of voters, I think I am on safe ground in saying these were Whites voting for a White candidate over a Black candidate with Race as the most important factor in their decision.

These White voters are not like Black voters who have reliably voted for White candidates all these years but now vote for Obama, because at the first opportunity of voting for a viable Black candidate, these White voters refuse to do so on the grounds of race. If such voters are not racists, I don't know what that term even means.

____________________

Ciccina:

RS - what was the exact wording of the question. Was the wording "the most important" or "one of the most important"?

____________________

RS:

22% said Race was important (82-12 to Clinton), and 8% said the most important (86-10 to Clinton).
If Race is one of the important factors, that says something... You can argue over it all you want, but that is a possibility you cannot ignore.

All I am saying is: there are reasons and there are reasons why some Democrats won't vote for Obama. PUMAs who say Clinton suffered from sexism and so want to punish Obama for it; people who support a(ny?) White candidate over a Black candidate; people who think Obama (like Kerry) is elitist and Megabucks McCain is the salt-of-the-earth; people who buy into the myth that Obama has no experience.

Just because some of them say they supported Clinton and are now voting McCain, doesn't make them PUMAs any more than it makes them racist. The point is, no-one knows how many PUMAs there are - are they just a tiny-but-vocal minority, something to keep the media happy? Are they PUMAs for the right reasons (I would argue against any sane reason existing, but I digress), or do they use it to cover darker prejudices?

We just don't know. And just because they are Democrats, doesn't make them free of racial prejudice.

____________________

RS - you can ascribe whatever motives you want, to whichever group of voters you want, for whatever reason. Labeling groups of voters - be they "security moms," "nascar dads," "compassionate conservatives," "pumas," "evangelicals" and so on - is only useful up to a point. You reach that point when you start assuming more than your data can back up.

____________________

Mark Lindeman:

Ciccina, sorry about confusing your two posts. RS's "reasonable to assume" is probably a bridge too far. I do think it's plausible that some Democrats for whom "race was an important factor," and who favored Clinton over Obama, favor McCain over Obama because they prefer white candidates. I also think it's plausible (and of course not exclusive) that some of these people favor Obama over McCain; perhaps they thought a white Democrat would be more electable, but that doesn't mean they prefer a white Republican. Then there are those who said race was important in the Dem primaries, and favored Clinton over Obama, and who favor McCain over Obama for reasons that have little to do with race.

I wouldn't say that "racists are being miscounted as PUMAs" -- I would say that the CNN survey doesn't provide any measure of "PUMAs" at all. "Clinton supporter" just doesn't cut it. Even "Clinton supporter who also supports McCain" just doesn't cut it.

Therefore, the following doesn't work for me: "...I fail to see why one should hypothesize about racism as a motivating factor at all given the meager amount of information available. Particularly when doing so feeds into such a divisive narrative." I've seen more evidence that many Democrats are racist than that many Democrats are PUMAs -- and the PUMA narrative is divisive by definition.

This isn't to say that I prefer "the racist narrative" to "the PUMA narrative," nor that I think Clinton supporters are racists, nor that PUMAs are racists, nor that it's the PUMA narrative that we should avoid hypothesizing about. I don't think that hypotheses are inherently dangerous. On the contrary, I consider hypothesizing an excellent way of putting brackets and question marks around all the narratives.

Earlier you noted: "...some chunk of the 9.3% who disapprove of interracial marriage may be African American themselves." ("Disapprove" seems to be a euphemism here.) I should mention that the results I reported previously were only for 2000 and 2002, because the question was dropped after 2002; probably the level of agreement has dropped a bit more since then. Extending the analysis back to 1998 to get a few more observations, I see 135 white Democrats who supported laws against interracial marriage, and 11 black Democrats. I think that's far from the most important caveat to my earlier comment, but YMMV. (One caveat I meant to add last night was that there are probably some Democrats who said they don't prefer Clinton to Obama, but will vote for McCain.)

____________________

Mark Lindeman:

Sorry to partly repeat myself: I seem to be warring with my browser.

Ciccina wrote: "Did you not mean to imply that racism was the most likely motive for voters to say they preferred Clinton or McCain over Obama?"

Hell, no. Excuse me, but I have no idea how you would get from my words to those words. Try this: racism is one plausible motive for some voters to prefer both Clinton and McCain over Obama.

This meaning-to-imply business is just gonna wear me out. Can I apply for Conscientious Objector status or something?

RS: Three ways in which race could be important to someone who prefers Clinton to Obama and, later, to McCain: (1) Someone doesn't respect and/or trust black people; (2) someone thinks a black candidate is less electable; (3) someone actually wants to prefer Obama based on his race, but ends up preferring Clinton -- and McCain -- based on experience. (By the same token, I bet there are some fairly hard-core racists who will support Obama.)

And in case it isn't clear, I don't think that any single question is a good measure of "racism."

____________________

RS:

Ciccina:
The whole point of my post was to point out exactly that - the Clinton supporters who now vote McCain could be PUMAs, could be Racists, could be other people. We don't know if these are mutually exclusive sets or there is some non-null intersection set.
Mark L. offered one alternative to the prevailing media narrative, which I thought was equally valid, and hence my posts. Maybe I should have just shut-up and let you two duke it out...

Mark L.: Of the three ways you point out, (1) would be racists... Anyway, when the question is "the most important factor", experience would not be in the mix (otherwise Race would not be the most important, no?) - and this was the case for 8% of WVa-nians (who voted 9:1 for Clinton).
As for the remaining 14% - who knows...

In any case, this CNN poll is a bunch of crapola - no cross-tabs, and a selective/misleading write-up.

____________________

JWilly48519:

I don't understand how so many pollsters and pundits can continue to ignore ideology as a motivator in this situation.

Some of Hillary's supporters, looking at her past positions and those of her husband, consider her to be a moderate, whatever positions she recently has advocated in seeking the nomination of a left-wing-dominated party. Those same supporters perceive Barack to be much more left wing. As such, expectations that they should now support him amount to a demand that they abandon their moderate ideology and move to the left.

Pundits, though, continue to treat Hillary's supporters as a monolithic cult-of-personality bloc. Hillary and the Barack team do so as well, in their exhortations that her supporters now support Barack and (in the latter case) their overt insults of those who decline the invitation.

It's all very strange. I'm surprised that anyone with even a smattering of historical awareness could sincerely so mis-characterize the situation. Maybe, though, there's a desperate unwillingness to recognize that in the presence of such an ideological split, choosing a candidate who appeals mostly to one side of the split is victory-dysfunctional.

In any case, for some voters ideology will continue to be more important than candidate-identity. It's been that way since democracy was first tried.

____________________

RS:

JWilly:
On the issues, which define ideology, do you really think McCain is closer to Senator Clinton than Senator Obama?
As Senator Clinton asked: Is your support for me based on the issues, or for me?

That is the message that needs to be sent out - Candidate McCain is waaaay to the right. "No way, no how, no McCain!"

____________________

JWilly48519:

What you or I think isn't relevant. On this website, opinions have to be supported with data.

My conclusion, as argued above, is that the most supportable interpretation of the available data is that a significant fraction of ex-Hillary supporters *do* see a significant ideological distinction between Hillary and Barack, and--judging McCain by his past record, not the positions he's espoused during this cycle to get the nomination of the right-leaning Republican Party--*do* see him as a better choice for a moderate Democrat.

Of course, that argument would suggest that if those individuals became convinced that the old McCain was false and the new one is true, their support for him might waver. But, that wouldn't support a conclusion that those individuals would be willing to move toward Barack. That ideological split is current, and perceptions of it are not likely to change because more information will serve to confirm it.

____________________

Please be patient while your comment posts - sometimes it takes a minute or two. To check your comment, please wait 60 seconds and click your browser's refresh button. Note that comments with three or more hyperlinks will be held for approval.

MAP - US, AL, AK, AZ, AR, CA, CO, CT, DE, FL, GA, HI, ID, IL, IN, IA, KS, KY, LA, ME, MD, MA, MI, MN, MS, MO, MT, NE, NV, NH, NJ, NM, NY, NC, ND, OH, OK, OR, PA, RI, SC, SD, TN, TX, UT, VT, VA, WA, WV, WI, WY, PR