Pollster.com

Articles and Analysis

 

POLL: Zogby IN, NC (5/2-3)


Zogby

North Carolina
Obama 48, Clinton 39

Indiana
Obama 43, Clinton 41

 

Comments
tom brady:

I wish we could see the crosstabs to understand why Zogby seems consistently to poll more favorably to Obama than do the other polls - does anyone have access to this information?

____________________

jac13:

Tom,

Here are some.

In Indiana, BHO/HRC/und:

18-29 - 51.8/34.0/7.8
30-49 - 50.5/34.8/7.7
50-64 - 37.8/47.6/6.9
65+ - 28.3/51.4/10.2
Male - 47.0/38.1/8.1
Female - 39.3/43.9/9.7
White - 35.9/47.1/9.1
Afr-Am - 77.9/7.5/14.6

In North Carolina, BHO/HRC/und:

18-29 - 52.4/35.6/12.0
30-49 - 57.8/32.0/6.2
50-64 - 45.4/39.3/8.4
65+ - 32.3/53.4/8.4
Male - 46.8/42.1/7.0
Female - 48.2/37.2/8.6
White - 33.0/55.5/3.8
Afr-Am - 77.7/5.8/15.7

The undecided percentages among the youngest IN voters and AAs in both states appear unusually high. Otherwise, these look like the demographic breakdowns from earlier states, but there's obviously a different distribution.

Hope this helps.

____________________

jac13:

I meant "youngest NC voters."

____________________

tom brady:

@jac13

It does, thanks. I agree that the undecided African Americans and younger voters seem high. But what I'd really like to see is the breakdown of these demographic groups as % of the survey. For example, does Zogby oversample African Americans, compared to other pollsters? Undersample older women? Something has to explain why Zogby seems systematically to get different results...

____________________

Mike_in_CA:

to tom brady:

I don't think Zogby's results in IN are necessarily "systematically different" or inconsistent with other pollsters. Instead of paying attention to the spread, because there is such a high number of undecideds, pay attention to the LEVEL of support each candidate gets. Obama is consistently polling in the low 40s (much like he was leading up to PA). Clinton's support seems significantly "softer" than Obamas, but, as in past contests, undecideds break for her in the end, she could end up winning by 10 points, assuming Obama gets a disproportionately small number of late-breakers, keeping him to the mid-40s. That's my take on this 43-41 poll. Although, the 41 number is quite low for Clinton.

Also, why 15% undecided 2 days before the election?? Gotta push these people.

____________________

carl29:

For what I see, I think that he is wrong about both. In IN Hillary is high single-digits. In NC Obama is low double-digits. I think that he does not want to blow it, so he is trying to have it both ways. I think he is just kind of guessing.

I think that in PA he did a good job by telling us head on, in his press releases, who those supposedly "undeciders" were. I remember that he said that those undeciders were Hillary's demographic, white, catholic, rural. So, their undecision was about whether they would vote or stay home. These people never considered Obama as an option.

What surprises me is that he is not telling us anything about those "undeciders" in IN. I bet those are the same demographic that tend to favor Hillary.

I think that Obama's ceiling in IN is around 45, just as PA, and Hillary's margin will depend on whether those "ethnic" voters want to mess up with Obama at polls by voting for her. I don't buy those "undeciders" are AA or young people. If there is an age when you think you are sure about everything is when you are young, so forget about "indecision" there.

____________________

tom brady:

Mike_in_CA - So Zogby doesn't push the undecideds, and those undecideds are likely to go for Clinton. Certainly that's consistent with Ohio, Texas and PA, where latebreakers went for Clinton. But without a breakdown of groups, it's hard to extrapolate from the Zogby survey to see where the undecideds are. In NC, for example, I think 15% undecided AA's are likely to break for Obama, 9 to 1. SAme for Indiana - 14% AA's are undecided, but will almost surely break for Obama. But how many are there?

____________________

Jordan Davies:

This time I'm not willing to bet the bank on undecideds turning for Clinton, especially in view of the negatives for her, and Obama's position re the Wright business. Also don't forget about the early voters. I think the difference in NC will be greater for Obama and about the same in IN.

____________________

jac13:

I'm a Zogby subscriber, but no polling expert. Can somebody tell me where I can find the percentages of the varous subgroups in his polling sample? I don't see it in the cross-tabs.

____________________

RS:

@jac13:
Thanks for sharing the cross-tabs with us. The IN numbers don't add up to 100% in the subgroups - e.g. 18-29 is 51.8+34+7.8 = 93.6% Are the rest for "someone else"?

Also, do you know what is the African-American fraction in Zogby's poll? Using your breakdowns, it appears to be 16%, which would be rather high (compared to SUSA, which I trust for demographics). This could explain why Senator Obama's ahead in Zogby's polls.

[I see "tom brady" asked a similar question, and jac13 replied he can't see it in the cross-tabs... Damn you, Zogby!]

Interesting note in Zogby's write-up:
"voters age 35 to 54 - Obama enjoys a 10-point lead. This was a group that went for Clinton in the recent Pennsylvania primary, after leaning toward Obama in the week before the election."

IF that trend repeats, would IN be PA, redux?

____________________

Paul:

In addition to data provided by "jac13", I am adding information on party affiliation. In Indiana, 12% of those polled are Republican and 13% are Independent/Unaffiliated. From Republicans: HRC 31.9%, Obama 28.3%, someone else 32.2% and not sure 7.6%; from Independents/unaffiliated: HRC 31.8%, Obama 44.6%, 16.4% someone else, and 7.2% not sure. Among Democrats (75% of those polled): HRC 44.5%, Obama 44.7%, someone else 2.5%, not sure 8.3%.

In North Carolina, 17% polled are Independent/Unaffiliated: HRC 37.4%, Obama 42.9%, someone else 8.7% and not sure 11%. Among Democrats (83% of those polled): HRC 39.8%, Obama 48.6%, someone else 4.3% and not sure 7.3%.

____________________

Paul:

In response to RS question on jac13 data, the missing percents are someone else, which total 7.9% overall for Indiana. Zogby is using 11% AA in Indiana. Splits in this latest poll for AA are as follows: HRC 7.5%, Obama 77.9%, 14.6% not sure.

____________________

RS:

@Paul:
Thanks.
But assuming jac13's numbers for the White vote are correct, the math doesn't seem to add up (assuming White + Black = 100%):
Obama 77.9*11% + 35.9*89% = 40.5%
Clinton 7.5*11% + 47.1*89% = 42.7%

So... with 11% Black/ 89% White and the above-mentioned vote fractions, the result should be Clinton 43%, Obama 41%... But Zogby says the opposite!

The only way I can get Zogby's Obama 43-41 result is with a 16% Black/ 84% White split. Hence I am puzzled...

____________________

Paul:

In response to "tom brady", Zogby uses the following the race breakdown in Indiana: White 83%, Hispanic 3%, AA 11%, Asian 1%, Other 2%. In North Carolina, Zogby uses the following race breakdown: White 62%, Hispanic 3%, AA 32%, Asian 1% and Other 2%.

____________________

Tom:

Thanks jac13. That is rather interesting information. In PA, a lot of the Us going in were in Clinton key demos, which broke with their demo for her. In IN, and especially NC, it looks like a lot of the Us are in Obama's key demo, especially AA.

The one thing I find troubling about Zogby is the very high number of "other" in his polls. In NC and IN, both are around 5, which is extraordinarily high. The true Us are in line with expectations. If I had to make a prediction, I would say NC is O+10, with IN going C+4. As bad as Obama has been beaten up over the last week or so, this will look like a big win for him, and will virtually seal the nomination.

____________________

carl29:

What puzzles me about Zogby's releases on NC and IN is that he does not tell us where those "undeciders" come from. You know is plain English just the way he did it in PA. I think that is not irrelevant information, given that he deemed it relevant back in PA. This make me think that he is kind of guessing there.

If I had to bet, I will say that those "undeciders" are ethnic whites, probably republicans or DINO, democrats in name only, that haven't quite made up their minds about whether showing up to vote or not. These people's indecision is not about Obama or Clinton; It's about voting for Hillary, whom they don't like, or not voting at all. Believe, this people are not coming out for the democrat in Nov.

____________________

RS:

@Paul:
Much thanks again!

Still I am somewhat puzzled, though - the 83% White/11% Black/6% Hispanics/Asian/other means that for the Obama 43-41 lead, Senator Obama carries the Hispanics/Asians/other category by a 77-18 margin, or similar to his margin among African-Americans.

Of course, 6% of 595 LVs is 36 people, so MOE is very high... Oh, what the heck - the race is tied, 43-41 or 39-45 or whatever. Never mind.

____________________

doriangz:

I am not an expert on polling so I cannot really comment on the validity of those numbers.
What I can do is shed a little light on the "undecideds break for Hillary" meme.
You have to remember that all the states where that meme has been proven were states where she was the favorite heading in.
In other words, the default position in TX, OH and PA was to back Hillary and people who were still undecided because they were torn ended up going back to where they came from.
In NC, clearly the default position is Barack Obama. Sure, there is Wright but a huge number of those undecideds in every poll (not only Zogby) are shown to be AA.
IN is more interesting. It has been great for Obama's expectations that IN has been said to be a toss-up but if you look up polls from a month ago, he was way ahead there. People are more familiar with him because part of the state shares IL media markets. In other words, there is nothing that says Obama is not the default candidate for undecideds in most of IN.
That's what Brian Howey, the David Ypsen of IN, is hinting at in his last column and I am inclined to think we may have a surprise there although the polls seem to show Hillary with the edge there.

We will see. But the Hillary gets the undecideds storyline is largely resting on the particulars of those last primaries (as well as rough media cycles for Obama in the last three days).
Yesterday was a good media cycle for him. Today should be as his performance on MTP was well-received (still waiting for the J-J dinner in IN). If the tone of the campaign is as mellow tomorrow as this weekend's (and I am not sure they are not going to play the terror card again - that Obama may be prepared for this time), then Obama won't have headwinds against him (he is climbing up the daily tracking polls instead of down the previous few times) and being the default candidate, there is a chance he could split those last deciders more evenly.

____________________

Shadar:

Three polls in a row by Zogby showing this tied is making things interesting. If it was a single poll it wouldn't mean much, but the fact that he has shown the same data three times makes this fascinating.

I'd say the polls are still fairly useless though as they have way too high of an undecided count. 2 days out, even a week out, there should be far less undecided voters than this. I would hope that his monday poll will have an undecided count closer to 5-6% at most, if not it will not tell us anything unless we know exactly where those undecided voters are.

But frankly, how the hell can anyone be undecided? You either are inspired and hopeful that Obama will change things or you think we need a steady experienced hand like Clinton. You either hate clinton for whatever reason or you don't. You will vote for a black guy or a woman or you won't.

I really don't see much room for someone to be undecided. But then there were undecided voters between Kerry/Bush as well... the gap between those two men was vast and yet people couldn't make up their mind? I don't see how that is possible. I think the whole "undecided" thing is more a label of "uninformed", as anyone who knows what is going on is going to have a candidate.

____________________

StandardDeviation:

There could be some utility in not pushing the undecideds in terms of being able to accurately predict the final spread.

If the poll is largely accurate with respect to the decided voters, and the undecideds break 50/50 (or 60/40 even), the poll will be within 3 points of the correct margin (even with 15% undecided).

If the poll is largely inaccurate with respect to decided voters, because the weighting is off, biases in getting responses, etc, then having a large pool of undecided voters gives the pollster a 'get out of jail free card'. The pollster can simply claim that the undecideds broke in a way that produced the discrepancy.

It's a win win for pollsters.

____________________

mago:

One datum on the NC demographics. Apparently the state keeps demographic info on early ballots (sorry, I've lost the link) and 39% of early ballots were cast by African-Americans. (We're talking about ~350K ballots here.) If this trend were to hold up overall, it would represent a higher proportion than pretty much any pollster is using, I think.

Btw, the one really odd thing in the Zogby crosstabls is the lack of a gender gap in NC.

____________________

Mike_in_CA:

mago,

to follow-up, NC early voting ended yesterday, and official numbers from the BoE has AA at 40% and whites at 56%. It's possible that Obama's GOTV game has prodded higher numbers of AAs to the polls to vote early, and that the overall % will decline by Tuesday. But with the early voting AA percentage at 40% I simply cannot see it falling back to 32% (as Zogby says) on election day.

If the 40% AA turnout and 56% white turnout (w/ 4% "other" splitting 50-50) holds and whites split 60-40 for Clinton, the final spread is: Obama 61-39, which would be a ridiculous blow-out and the media would pounce all over that.

____________________

Shadar:

If the 40% AA holds in NC then Obama will get a blowout of SC/GA proportions which will def be hard to spin. He will get a massive proportion of the delegates, far more than she got out of PA. In fact even if he loses Indy by 10% he'd prob come out with more delegates from NC than Indy + PA combined.

This all assumes the 40% AA holds, if it drops substantially in regular voting he will be closer to a 10-12% win. But it sure looks good for him atm.

____________________

mago:

IIRC Obama won the early vote in TX but lost the primary on voting day, so I'm not counting any chickens. But it's interesting.

____________________

cinnamonape:

Nobody seems to have picked up on this Rassmussen Poll on May 1st for Oregon.

Obama 51% Clinton 39% Undecided 11%

http://www.rasmussenreports.com/public_content/politics/election_20082/state_toplines/oregon/toplines_oregon_democratic_primary_may_1_2008

There doesn't seem to be much discrepancy b/w Clinton and Obama regarding voting for the other candidate if their candidate loses...but because the Obama voters are a larger % of the total that would actually result in more voter defections.

____________________

tom brady:

Presumably the early voting is being picked up in these polls in NC, yes? So either they are weighting the AA vote inaccurately, or they have some reason to believe that by election day the overall AA vote will drop from 40% to closer to 33-35%. Otherwise we will have a bunch of pollsters with egg on their faces

____________________

Mike_in_CA:

tom,

there's no way for these pollsters to gauge AA turnout, which has exceeded pollster's estimates in a number of states early on in this cycle (i.e. SC, GA, etc.) The excitement in the AA community may be even greater in NC than we expect because Obama has his back up against the wall and is playing the role of "underdog" this weekend, which would lead to more of his supporters coming out to boost him (I would think).

Wait for tomorrow's polls to see if AA turnout numbers have been adjusted upwards in some pollsters' averages. Tomorrow will be a very interesting polling day indeed. A lot has changed since last week IMO.

____________________

Zogby hasn't been very accurate. They have a very obvious scheme. Make it look like Obama is ahead so that more people vote for him because they perceive him to be "the winner" and then, the day before the primary, report actual numbers so it looks as though you were more accurate. Anyone who followed the polls in Ohio, Penn, & Texas can see this. This is one polling firm that I think is more an arm of the Obama PR machine. For all who like an actual picture of the race, I would recommend SUSA which has been the most accurate recently.

____________________

boskop:

zogby. the official media arm of the obama party.
on late night he was embarrasingly clear about his support. his polls, well...you'll see on tuesday and then let's hope, he finally becomes the used car salesman he always wanted to be.

____________________

jac13:

Anthony/boskop --

Any reason you neglected to mention Zogby in PA? He not only picked up the late trend towards HRC, but hit the result right on the head. A fluke?

And do you really think it helps a candidate for a pollster to inflate expectations to a level he cannot meet?

Call me crazy, but it makes no sense to me.

____________________

Mike_in_CA:

anthony and boskop,

The trajectory of the discussion on this post was (finally!) one with intellectual merit, devoid of political shilling, and addressing the actual numbers and possible outcomes and then you come on here and cry foul that Zogby is biasing their results to benefit Obama. (which in and of itself its just absolutely stupid to assume)

Enough enough enough. Discuss the polls or leave.

____________________

boskop:

i think it's really time to hear everyone's take on the 'perfect storm' scenario.

on june 3rd barack obama will not have the delegates. if he is even or close in the popular vote without florida and michigan then we have the perfect storm.

in which case i ask, if the democratic party was as poorly organized and managed to have let this come to pass, perhaps they do not deserve to win in the general election?

no matter what, the thread of any successful candidate will likely unravel in the face of the unresolved two state debacle. there is virtually no way to determine the strength of your party's candidate without ascertaining their viability here.

so the perfect storm looms. if indiana goes to hillary which it looks to do, no self respecting democrat could swallow anyone's win before florida and Michigan's settle.

zogby numbers mean nada. let's figure out florida before june 3rd or it's hail to mccain!

____________________

cinnamonape:

Anthony/Boskop (Zuid-Afrika?)- Actually Zogby was pretty much right on the money in Texas. He called a Hillary win by 3 in the General. And he's had very good predictions w(better than his rivals, for the most part) in Iowa, Missouri (the only one to call that accurately), Nevada, Georgia, and South Carolina. He hasn't surveyed more than a dozen states...so that's actually very good.

I don't think that one can simply ignore his surveys by asserting that he's pro-Obama. How would one explain that Texas result when others were predicting an Obama upset?

____________________

boskop:

when a pollster is as outspoken for one candidate as zogby has been early in in the race, his credibility is in question. you are welcome to check out his interview with jon stewart or colbert, where it was utterly un-professional.

as this is a forum for polling discussion, not to take this into account when there are so many different protocols for pollsters to use, protocols varying upon demograhics, questions, landlines versus cells, automated questions versus real people, length and time of day, then i suggest that you are naive.

pollsters have a remarkable leverage in this nominating madness and they are not above tilting their data to get a false read for a little "self fulfilling prophecy'

if zogby had not been blatantly anti clinton, i certainly would not have been wary of anything this man were to contrive.

____________________

Mike_in_CA:

note to all: don't feed the ego of this boskop guy by engaging in debating his "talking points" re: Zogby's bias and resolving Florida. Talking points are designed so that they are UN-debatable, and if you try to debate them you'll lose.

Let's just ignore, and discuss polling trends re: IN and NC (since that IS what this thread's about!)

____________________

boskop:

MIKEINCA

wow!
if you want to discuss zogby polling 'facts' then resolve all this guy's mea culpa's like "voters want anyone who's not Bush", for kerry's so called win. is that polling data or predictive addiction?

what about his tendency to predict the undecideds completely wrong? he did this with john kerry based on his 'belief' that he knew the pulse of the country, and he did it in new hampshire by saying again completely unprofessionally, "barack obama had another big day" of course he did not.

then after promising he would do no more predicting just polling, he proclaims a 13% win for obama in CA. you certainly must know how correct that was being that you are from CA? hillary won by 10%. do you call being 25% off a reliable pollster?

all i'm saying is a man like zogby has to be taken with a serious grain of salt. if rasmussen and SUSA and ARG all came up with about the same numbers i'd say, hey, obama's on a roll here. gotta hand it to him.

that it's zogby "predicting" again should be read with a warning sign. he has embarrassed himself far too many times to be taken seriously but for some reason never learns.

____________________

Mike_in_CA:

boskop,

Zogby was wrong on CA. Other than that he's been very good this cycle. Yes, he was off by a lot in CA. But why does everyone hold that against him and skew their opinion. ARG is consistently off. A lot of other polling firms have been significantly more wrong than Zogby. But because he screwed up CA, and the media covered it INTENSELY he will never live it down. Talk about bias.

Besides, read my posts above about IN. I don't think that because Zogby has Obama out ahead in IN necessarily means he'll win. His support hovers around 43. Meaning he could end up at 45% with Clinton picking up the lion's share of undecideds. Lets see what happens this week.

____________________

Pat:

I agree with mike-in-ca. We should ignore these people who try to change the discussion to the same BS that is all over other blogs. Lets look at the polls and numbers and leave out personal preferences out of these discussions.

____________________

Does anybody know what the partisan x-tabs are for Zogby? i'd like to know how many GOPers they are saying will vote. It will likely be 10-15%. That could be where the discrepancy is.

I would guess that the GOPers would be splitting about 50/50.

____________________

Claude:

Jesus boskop, how is Zogby helping Obama by coming out 3 days before the Indiana primary and inflating expectations for him? Infact, I can bet that the Obama camp is pretty pissed at Zogby right now. These Zogby polls could be very bad for Obama if the media starts giving them serious coverage.

____________________

adocarbog:

Prediction for NC and IN
sorry it's long but I want to explain how I got to it.
First IN and I made this on 4/29/08
The three latest polls seem to give HRC a decent advantage in IN. However, a closer look at cross tabs paints a different picture so please hang in here with me.

Intro
You must enter an Intro for your Diary Entry between 300 and 1150 characters long.
The three latest polls seem to give HRC a decent advantage in IN. However, a closer look at cross tabs paints a different picture so please hang in here with me.


In any Obama primary where AA voters may make a difference one has to determine if the AA vote has been properly polled.
The important thing is that every 1% that AA vote is larger in a democratis primary than that is almost an entire % gained by Obama and also the same amount lost by HRC. In fact it is a 1.8% swing for every 1% of AA vote.
Also it is helpful to see how Kerry did against Bush in 2004. As a rule of thumb AA are a higher % of the democratic primary vote the worst Kerry did in 04. Look at VA, SC, GA, MD, OH, PA etc.

AA are 8.9% of Indiana population and they were 16.24% of Kerry voters. Kerry lost by 21% in Indiana a heavy red state. The more a state is red the heavier AA percentage is in the democratic primary vote compared to their state population percentage. Why???
Well most AA, by a wide margin, are democrats. Most whites in red states are republicans. Therefore, the more the state is blue the more white voters vote democrat the more AA vote is diluted.
Look at following numbers:
MD Kerry +13% win AA +28% (this means AA 29% of pop and 37% of vote or 28% increaseor 29% X 1.28 = 37%

PA Kerry +2% win AA 40%

OH Kerry -2% loss AA 50%

VA Kerry -8% loss AA +50%

SC Kerry -17% loss AA +89%

GA Kerry -17% loss AA +76%

IN Kerry -21% loss AA (the big mistery to be answered May 6th)

Well let us estimate.

MD If it is merely a 28% (MD example) increase then that would mean 8.9% X 1.28 = 11.4%. This is highly unlikely because Kerry won MD by 13 and lost IN by 21. 11.4% is the bottom range!!!

PA 40% would put IN AA at 12.46%.

OH / VA If it is the OH / VA +50% then 8.9% X 1.5 = 13.35% This is somewhat more reasonable.

GA / SC If it is GA or SC where Kerry lost by 17% only 4% less than he lost IN. then we can expect 8.9% X 1.76 = 15.66% to 8.9% X 1.89 = 16.82%.

So once I calculated these ranges:
Low 11.4% MD
Medium Low 12.46% PA
Medium 13.35% OH / VA
Medium High 15.66% GA
High 16.82% SC

I looked at what % of Kerry voters AA were in 2004. I was amazed thar they were 16.36% or in the high range. Can Obama excite AA more than Kerry. I think so.

SO WHAT % OF AA VOTE IS PREDICTED BY THE THREE POLLS IN MY DIARY TITLE???

SUSA AA are 10% (this is under even the lowest estimate)
PPP AA are 12% (This is essentially low turnout prediction)
Howey-Gauge AA are 20% (This is high + assumes Obama excites AA vote)

SUSA is clearly wrong, plain wrong!!!

PPP is off the mark but most pollsters are chasing their tails anyway lately.

Not surprisingly it is only the last poll that has Obama leading.
So what do you think?
My prediction 800,000 votes cast (+/- 25K) and HRC wins 52%-48% for a net gain of +30,000 votes.

My prediction details:

AA turnout 16%. Obama 90% of that vote. White + others 84% of the vote and Obama gets 40% of that vote. He nets just over 48%. HRC 52%. Close race. HRC does better with whites in blue states while whites in red states (non-deep south) are more favorable to Obama (exception WI and IL). He still loses white vote in most red states but does better than in blue.

For my NC prediction see my diary by adocarbog dated 4/25/2008.
Obama by 58%-42% in NC 1.5 Mil votes cast and Obama netting 240,000 votes. Note that the NC prediction is conservative while IN is neutral.

ALSO THE POLLS PPP AND SUSA SEEM TO ASSUME GOOD OLDER VOTER TURNOUT AND WEAK YOUTH TURNOUT.


North Carolina Prediction made 4/25/08

With 10-11 days to go it would be good to get some predictions on NC out so that we can compare later on May 7th.

Intro
You must enter an Intro for your Diary Entry between 300 and 1150 characters long.
With 10-11 days to go it would be good to get some predictions on NC out so that we can compare later on May 7th.


NC should not be too difficult to predict. We already have results from GA, SC, VA and MD. NC is right in the middle. It is also a growing state. in 2000 it had just over 8Mil people and by 2006 it gre an amazing 10% to over 8.8 Mil. So unlike PA and OH it's not stagnant. VA where Obama did great also grew 8-9% in the same 6 years. MD where Obama did great as well but not the the extent as in VA population grew 6% from 2000-20006.
Georgia population in that time frame grew by 7.7% while GA where Obama did best grew 14% in just 6 years.

So the growing south and south east is favorable to Obama as it is a growing and economicaly vibrant part of the US. On the other hand OH and PA grew by a mere 1% during that period.

Now for the meat:

MD Obama 61% HRC 36% White Vote (WV) was 53% and Obama got 42% of WV
Kerry won MD v Bush by +13 and AA vote in 2004 was 24%. AA are 29% of voters in MD so they undervoted. They were 37% of the Primary 2008 vote. They also undervoted here because that is only a 28% increase over population percentage. This can be explained that MD is a safely democratic state where a lot of WV vote for democrats. 28% increase is the important number to remember as well as Obama getting 42% ow WV.

VA Obama 64% HRC 36%. WV was 61% and Obama won the WV with 52%. Bush won VA in 04 by +8. AA are just under 20% of population in VA and they were 21% of the 04 VA voters. So VA AAs tend to vote in proportion or even slightly higher. In the 2008 primary they were 30% of the vote so the voted 50% over, more that MD 28%. this can be explained by AA voting slighlty more in VA and there also being fewer WV for democrats in VA.

North Carolina went for Bush +12. AAs are just under 22% of the population and they were 26% of the voters in 2004 GE. So AA in NC sure like to vote.

South Carolina is a bit strange because Edwards was still in. AA are 29% of the population. They were 55% of the primary vote so they voted 89% over. Bush beat Kerry by +17 so most WV are republican. AA were 30% of population so they voted in 2004 in proportion. Obama got 24% of WV but HRC got only 36% and Edwards won WV at 40%. It is safe to assume that if he were not there Obama would get at least half of that vote fo a total of 42% of WV.

GA 29% AA population and Bush won 2004 by +17. AA vote in 2008 primary was 51% or 76% over population. Again WV are mostly pro republican. Obama got 43% of WV.

So what does this mean for NC. Likely the AA turnout will be less that +89% in SC but more than +50% from VA. Obama's WV floor will be 40%. He will liley do a bit better than 40% but let's assume 40% in case the Wright thing hurt him some. AA vote may be 65%+. That would put AA at 36%.

So Obama wins 90% of AA for a total of 32.5% and he gets 40% of WV (and others) (40% X 64%) for a total of 25.6%.
Obama's total will be 58.1% at the minimum. to HRC 42% at the maximum.

Turnout will be 100% of Kerry vote are AA will be energized as well as independents and cross over Republicans. Turnout will be 1,500,000 +/_ 100K.

This will be a net + pop vote for Obama of 240,000. I am very confident he will pick up more Popular vote net from NC than she did in PA as well as delegates. Delegates +20 for Obama.

____________________

Jac13

That's my point. There was no "late trend" for Clinton in my opinion. She was always ahead. Same as in Ohio (where I spent lived for 24 years). Everyone knew she was ahead by a decent margin in Ohio despite the zogby polls showing Obama closing or leading until right before the primary.

This is just something i've observed SUSA seems to be much better than Zogby.

____________________

Claude:

I think most of these polls are underestimating the AA turnout. Most polls have the AA sample at 8-11% (SUSA at 10) but if we look at the exit polls from the 3 states surrounding IN (IL, MI, OH) it can be seen that in IL the AA turnout was 1.6x % of their population, in OH the AA turnout was 1.5x the % of their population, in MI(where Obama wasn't even on the ballot) their turnout was 1.5x the % of their population. Since AA are around 9% of Indiana if that pattern holds up the AA turnout should be around 13.5 %. Add that up with the the perception of Obama having his back against the wall and him having a bad week it could further boost the turnout by 1-1.5% which would dramatically alter the final results. I think a lot of pollsters are going to have egg on their face on wednesday.

____________________

Mike,

I thought I was discussing this poll. can I only share my opinions on the polls if they agree with yours? Maybe it will help if I provide a little more explanation of why I feel this way:

Ohio
Zogby Poll 3/3 before Ohio 44 Clinton 44 Obama
SUSA Poll 3/3 before Ohio 54 Clinton 44 Obama

Actual result in Ohio 55 Clinton 45 Obama

Penn.
Zogby Poll 4/19 Clinton 46 Obama 43
Zogby Poll 4/21 Clinton 51 Obama 41
(two days later)

These are just a few instances that i've noticed. Whenever Clinton is favored to win a state their "undecided" total is higher than most other polls.

Example: Indiana on 5/6

Zogby 5/3 Clinton 41 Obama 43 Undecided 8

I am very interested to see how accurate this poll is on tuesday.

Most other polls besides this one have Clinton up by 7-10 points.

Perhaps I'll rethink my position and make my final judgement after the results on tuesday.

There is nothing wrong with disagreement. But if this site is only for people with whom you agree, then it is probably not one I would like to frequent anyway.

____________________

Mike_in_CA:

anthony,

thanks for the reasoned post above. i don't think everyone should agree with me, i just think we should all agree that this forum is for discussing POLLS (i.e. numbers) and not politics -- except as they pertain to polls.

That said, you make all valid points. The thing about this darn IN poll is that right now it looks close, i.e. 43-41, but that leaves a HUGE amount of undecideds. If they broke the same way the rest of the electorate is breaking then it could be close. if they break like they did in OH and PA, then Obama could be stuck at 45%, while Clinton nets a 10% win. It's so hard to read these types of polls when the race "seems" close, but may end up being a 10% deal in the end.

Per Claude, above:

I, too, think the AA turnout % is drastically underestimated, but what I don't understand is why, this late in the game, after seeing state after state with above average AA turnout, these pollsters haven't adjusted? Or is there something else going on that they know better than us ?

____________________



Post a comment




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