« BACK  |  PRINT

RS

MEMBER DIARY

Quinnipiac poll claims Obama over 50 percent in OH, FL, PA — but let’s look at the instant replay

By: Finrod (Diary)  |  August 1st, 2012 at 12:49 PM  | 

RESIZE: AAA

Quinnipiac University has a new poll out of Ohio, Florida, and Pennsylvania:

President Barack Obama hits the magic 50 percent mark against Gov. Mitt Romney among likely voters in Florida, Ohio and Pennsylvania, with wide support for his plan to hike federal income taxes on upper-income voters, according to a Quinnipiac University/ CBS News/New York Times Swing State Poll released today.

This is the first measure of likely voters in these swing states and cannot be compared with earlier surveys of registered voters. Matching Obama against Romney in each of these key states – no one has won the White House since 1960 without taking at least two of them – shows:

* Florida: Obama edges Romney 51- 45 percent;
* Ohio: Obama over Romney by a slim 50 – 44 percent;
* Pennsylvania: Obama tops Romney 53 – 42 percent.

Support for President Obama’s proposal to increase taxes on households making more than $250,000 per year is 58 – 37 percent in Florida, 60 – 37 percent in Ohio and 62 – 34 percent in Pennsylvania, the survey by Quinnipiac/CBS/The Times finds.

They make a big deal about how this poll isn’t strictly comparable to previous polls, because they’re screening for likely voters, and not just polling registered voters. What they don’t tell you in that url (you have to go all the way to the bottom and open up a Word document for each state linked there) is the demographic breakdown they’re using. First for Florida:

(apologies for the wonky formatting, but I can’t get html to recognize a tt tag)

PARTY IDENTIFICATION - Generally speaking, do you consider yourself a Republican,
a Democrat, an Independent, or what?

                            LIKELY VOTERS........
                            Weighted   UnWeighted
                            Percent    Frequency
PARTY IDENTIFICATION
 Republican                 27%         359
 Democrat                   36          373
 Independent                32          393
 Other/DK/NA                 5           52

Then for Ohio:

                            LIKELY VOTERS........
                            Weighted   UnWeighted
                            Percent    Frequency
PARTY IDENTIFICATION
 Republican                 27%         345
 Democrat                   35          397
 Independent                32          392
 Other/DK/NA                 5           59

Finally for Pennsylvania:

                            LIKELY VOTERS........
                            Weighted   UnWeighted
                            Percent    Frequency
PARTY IDENTIFICATION
 Republican                 32%         390
 Democrat                   38          408
 Independent                26          322
 Other/DK/NA                 4           48

Doing the math, the D-R-I breakdowns:

Florida, unweighted: 31.7-30.5-33.4 (D +1.2)
Florida, weighted: 36-27-32 (D +9)
Ohio, unweighted: 33.3-28.9-32.9 (D +4.4)
Ohio, weighted: 35-27-32 (D +8)
Pennsylvania, unweighted: 34.9-33.4-27.6 (D +1.5)
Pennsylvania, weighted: 38-32-26 (D +6)

While Rasmussen doesn't publish breakdowns per state, their national partisan breakdown has Republicans at a low for 2012, but still R +0.9: 34.0-34.9.

So what can we conclude from this poll? Obama breaks 50 percent in Ohio, Florida, and Pennsylvania-- if you put your thumb on the scale to give Democrats anywhere from 4 to 8 more points off the raw numbers, and maybe even more than that compared to what a proper weighting probably should be.

In other words: another biased poll, nothing to see here.

Tags: , , , ,
Get Alerts

COMMENTS

POSTING RULES  |  LOGIN TO COMMENT  |  REGISTER TODAY
  • emptybucket

    the skewedness (is that even a word) of those latest polls and realize they are over polling the dems to get those statistics. Am trying to find the factual chart regarding percentage of overpolling of dems versus the models from both 2008 and 2010. Will post it when I find it, still a bit clumsy on the Internet.

    Polls need to be taken with a grain of salt for the most part.

    • gekster

      The one poll that counts will be on election day.

      • emptybucket

        I find it interesting the millions and millions of dollars being spent by both campaigns on “internal” polls.

        Just speak the truth, stand up for what is Right, live your life as you should and don’t worry about what some poll has to say. AND save that money for a better use.

  • Pingback: read

  • Pingback: free music downloader

  • CrabCakes

    Quick Googling shows that the voter breakdown by party in 2008 and 2010 were as follows:

    Ohio:
    2008: 39% D/31% R/30% I (D +8)
    2010: 36% D/36% R/28% I (–)

    Florida:
    2008: 39% D/32% R/29% I (D +7)
    2010: 36% D/36% D/29% I (–)

    Pennsylvania:
    2008: 44% D/37% R/18% I (D +8)
    2010: 40% D/37% R/23% I (D +3)

    It appears that their model is roughly accurate if you assume that turnout will resemble 2008, which is probably too optimistic for Democrats.

    If you assume that the numbers will look more like 2010 (probably too optimistic for Republicans), then you’d shave 4 points off Obama’s lead in Ohio, 9 points off in Florida, and 3 points off in Pennsylvania.

    Personally, I think that shaving 2 in Ohio, 4-5 in Florida, and 1-2 in Pennsylvania is about right, as I think that turnout will be somewhere between 2008 and 2010.

  • ohiohistorian

    You sort of “split the difference” between 2008 and 2010. What makes you think that 2012 couldn’t go even farther to the right than 2010?

    There is only ONE poll that counts, and that will be Nov 6 this year. Suggest that all take part.

  • PowerToThePeople

    growth in the percent of people who vote republican and a lesser amount of people who turn out to vote democrat since 2010 one would have to assume Obama has done a better job from 2010 till now then he did from 2008 to 2010. One would have to assume that many people have decided that his failed policies have suddenly become correct.

    And for one to assume that out of all the telling numbers out there pointing towards a bitter defeat for Obama both in the once blue states and the many swing states that suddenly Quinnipiac has become anything near a reliable polling agent, one would have to be quite ignorant.

    In less than a hundred days, liberals and their leader Obama are going to be dealt a harsh blow of reality. But by all means, stay delusional.

    Just for the record, Romney takes Florida, Ohio, and either wins or comes very close to winning in Michigan and Pennsylvania. Obama will be a one term loser.

  • APA Guy

    2010 was a dead-even party ID split…and Republicans whipped Dems every which way from Sunday. I have neither seen nor read anything that leads me to believe that the enthusiasm gap has changed for the better for Dems – or that any mitigating factor has swung the LV party ID back to anything remotely like what transpired in 2008.

    These polls are propaganda tools…just like the pew poll that showed Obama 16 points ahead just before the 2008 Election. They are meant to depress Republican turnout. This is why Dems hate Rasmussen polls so much. He polls accurately…and has called the last 2 presidential elections 100% correct.

    Oh, and he has Romney up +3 nationally.

  • http://travismonitor.blogspot.com Freedoms Truth

    It’s time to cut to the chase and point out that they sampled, then slanted that sample in a way that over-weighted the Democrats.

    Republican 27% 359 (30.5%)
    Democrat 36 373 (31.6%)
    Independent 32 393 (33.4%)
    Other/DK/NA 5 52 (4.4%)
    1177

    But we should ignore these polls at our peril. Before we saw biased polls that showed the race even, and we could take comfort in that, thinking ‘aha , Romney is really ahead”. These now show, with sizable bias, equally sizable leads … Obama would be winning in PA and OH even if you take out the slant.

    the ‘de-slanted’ numbers would be Obama still leading:
    Ohio 46-44 Obama leads
    Florida 46-46 tie
    PA 49-44 Obama leads, outside margin of error

    This is worrisome. Romney should be stomping Obama, who has such a dreadful economic record he should be at 20% approval and not electable to be dogcatcher. He’s been awful … but he’s got support. Dont ignore the biased polls, but lets work to prove them wrong.

  • CrabCakes

    Reason for using 2010 as a baseline:
    -It is the most recent election.

    Reason for using 2008 as a baseline:
    -Historically, presidential elections have drastically different (and higher) turnout than midterms.

    I think that its self-evident that at least some voters who aren’t as likely to vote in midterm elections (e.g., young people, minorities) will vote in a presidential election year. I don’t think it is at all likely that as many “unlikely voters” will turn out as did in 2008, though. That’s why i split the difference.

    Even assuming a 2010 turnout, though, the Q poll points to an Obama win if the election were held tomorrow (which it obviously is not). I haven’t seen any data to suggest that the GOP is in an even stronger position than it was in 2010, which is what Romney would need in order to win unless he can flip some Independents or Democrats.

    In any case, my reason for posting was to provide objective baselines for determining where the race stands if you don’t buy the Q poll’s weightings and would like to change them according to your own assumptions about turnout in November.

  • APA Guy

    2006 was a strong midterm turnout year for Dems, but 2008 was stronger.

    If you use that model as a basis to predict turnout – and factor in the increasing intensity on the part of conservatives apropos removing Obama – you can logically deduce that 2012 will be an even better GOP turnout year than 2010.

    Oh, and they’re not MY assumptions. Even Dems concede that our side is more fired up than theirs for this one.

  • gekster

    and since they were not loudly heard, they were irrelevent.

    Tell me, how did that work out.

  • emptybucket

    any case. Are you sure it, or yourself as claimed, are free of bias?

    As usual, too much is assumed with poll results, you cannot leave out the enthusiam of the tea party members and Repbulicans who will go to the election booth in droves in, ah is it 98 days?

  • APA Guy

    …so a “tie” isn’t really a tie when the votes are all counted. The incumbent, particularly in an election year where the economy is terrible, MUST have a significant lead to fend off the challenger. When that is not the case i.e. Carter in 1980 and Bush Sr. in 1992, the challenger wins handily.

  • CrabCakes

    I’m assuming:

    1) that people who tend to sit home during midterm elections tend to vote Democratic at a higher rate than Republican.

    2) that some of those people who sat home in 2010 will show up in a presidential year simply because it is a presidential year.

    3) that the GOP hasn’t increased its number of enthusiastic voters enough since 2010 to offset the additional people who will be at the polls who weren’t there in 2010. (In 2010 63% of Republicans and Republican leaners were “more enthusiastic than usual about voting.” Right now that number is at 51%.)

    I wouldn’t mind if the Romney campaign discounted polls like this and assumed a 35% D/39% R/27% I turnout rate and focused on GOTV instead of flipping Independents. I don’t thing they’d be too happy with the result, though.

  • gekster

    You forget where the Tea Party was dismissed, but on election day showed up.

    The Dems have no such support.

    How is it living under that dark cloud you portrey.

    And I hope you are smart enough to know about ass/u/me/ing things.

  • APA Guy

    1) that people who tend to sit home during midterm elections tend to vote Democratic at a higher rate than Republican.

    This is easily dismissed by even a rudimentary glance at the RV/LV disparities in months and months of polling comparing Obama and Romney. Obama easily wins the percentage of the RV tally that will not turn out. Romney is winning those likely to show up, including Indies.

    2) that some of those people who sat home in 2010 will show up in a presidential year simply because it is a presidential year.

    You are assuming that those people will vote Democrat. See my response to Point 1 for the fallacy of this reasoning.

    3) that the GOP hasn?t increased its number of enthusiastic voters enough since 2010 to offset the additional people who will be at the polls who weren?t there in 2010. (In 2010 63% of Republicans and Republican leaners were ?more enthusiastic than usual about voting.? Right now that number is at 51%.)

    Again, where is this pessimistic view of conservative enthusiasm coming from apropos the primary results…your rear end?

    Only an Obamabot would assume that his side is more enthusiastic than ours. Unfortunately for bots, FACTS tend to get in the way…a reality you will come to realize on Election Day.

  • CrabCakes

    I only meant “objective” in reference to the exit polls themselves, not in reference to any interpretation of them, including mine. If you think that this year will be show a 2% improvement in GOP party ID turnout over 2010, now you have the “objective” evidence of the actual 2010 numbers to determine what that means. That’s all I was getting at.

  • CrabCakes

    I was very disappointed the morning of November 3.

    http://www.realclearpolitics.com/epolls/2004/president/us/general_election_bush_vs_kerry-939.html

  • emptybucket

    issues about the word objective. I haven’t had one single political conversation with a left of center friend in years yet they always use that word objective as if there is no bias in what they are saying.

    Thanks for taking the time to repond.

  • gekster

    You sure you got that right.

    Or were you actually disapointed.

  • gekster

    You did post you were disapointed on 11/3/2004.

    Bush was re-elected, why the disapointment.

  • CrabCakes

    The moderators are kind enough to let me play in the sandbox as long as I don’t crap in it. It’s more fun to talk to people with whom I disagree on most things. Plus, I like having my assumptions challenged.

  • CrabCakes

    Is my sig line only showing up for me?

  • gekster

    It is to promote conservatism and Republicans.

    Try the HuffPo or DailyKos.
    They are that way. <——-

    And tell me, where and when has libeeralism actually worked.

  • gekster

    Tell me when and where liberalism has actually worked.

  • CrabCakes

    Go up and reread the stuff in my original response to you between my name and the sig line. I’ll repost it for you here:

    “The moderators are kind enough to let me play in the sandbox as long as I don?t crap in it. It?s more fun to talk to people with whom I disagree on most things. Plus, I like having my assumptions challenged.”

  • CrabCakes

    To politely acknowledge your question, though, and give an answer that I’d qualify in all kinds of ways if this thread were in fact about how well liberalism works in practice, I’ll just go with the standard liberal reply and go with Norway.

    As this is not a thread about how well liberalism works, that’s all I have to say about the matter here. Also, as you rightly note, this is not a website designed for me to use to convince everyone how awesome liberalism is. That would be, to reuse a phrase, “crapping the sandbox.”

    If you want to talk about polling methodology and partisan weightings, though, I’d love to.

  • gekster

    You want to pay over 80% of your income in taxes.
    You want the government to wipe your butte from cradle to grave.
    You want the government to take away all of your failures.
    You want the government to make sure you won’t be allowed to be all you can be.
    You want the government to run your life because you, according to them, are too stupid to do it on your own.
    You want the government to control your life, because they know better than you.

    So tell me, when the government runs out of other peoples money, where do they get it from then.

  • gekster

    I guess his time on the OWS computer is up.

  • westcoastpatriette

    Perfectly said.

  • tnfriendofcoal101368

    15.0% real unemployment
    23 million Americans unemployed
    5.7 million Americans unemployed for more 6 months or more
    45 million Americans on food stamps
    5.4 million Americans on Social Security Disability

    Our plan – it worked – in creating a Welfare state
    We did build that welfare state.

  • gekster

    and I am hearing crickets from him.

  • gekster

    did your Mom take away the laptop.

    Time to get out of the basement and go to bed.

    (if you think I am making fun of you, well, I am)

  • Finrod

    He’s being honest about who he is and what he believes, so there’s no need to harass him about it. We can be polite to our occasional visitors from the other side without diminishing the strength of our arguments any.

  • tnfriendofcoal101368

    The real question is did you spend more time walking your dog as socialist Barack Obama spent on the economy.

  • gekster

    My dog walks on his own, I don’t have to walk him.
    He can walk without any help from me.

    (to tell you the truth, he takes me for a drag, he drags me along)

  • CrabCakes

    Yes, Republicans poll better among LV’s than RV’s. That’s been a basic fact of polling for decades. I’m not talking about how likely Democrats are to turn out this year versus how many Republicans are likely to turn out this year. I’m talking about how many Democrats are likely to turn out this year versus how many Democrats turned out in 2010 and how many Republicans will turn out this year versus how many turned out in 2010.

    As you yourself note, the GOP was extremely energized in 2010, and Republicans saw an impressive turnout. Even if their enthusiasm were the same this year (which it isn’t; my source for the 63%/51% enthusiasm numbers is Gallup, not my rear end), the Democrats have a larger number of people who simply sat home in 2010. In short, the Democrats have more people who didn’t vote in 2010 but will probably vote in 2012 than the Republicans do because the Republicans got such a high percentage of their voters to the polls in 2010.

    Finally, in 2010 the enthusiasm gap was +19% R. Right now it is +12% R. That’s good for the GOP, but not good enough to expect a repeat of 2010.

    Gallup links:

    http://www.gallup.com/poll/144152/record-midterm-enthusiasm-voters-head-polls.aspx

    http://www.gallup.com/poll/156194/Democratic-Voting-Enthusiasm-Down-Sharply-2004-2008.aspx?utm_source=google&utm_medium=rss&utm_campaign=syndication

  • streiff

    using Gallup as a model we have this table

    In 2004 the enthusiasm gap was D+17. In 2008 it was D+26. If these data mean anything, and I’m not sure that they do, I don’t see how you can say an R+12 foretells anything but a curbstomping.

  • APA Guy

    http://www.gallup.com/poll/111703/final-presidential-estimate-obama-55-mccain-44.aspx

    Why not link to Rasmussen, who accurately predicted the 2004 and 2008 presidential elections instead of a pollster who has consistently oversampled Democrats for years and years?

    Don’t bother…I already know the answer.

  • CrabCakes

    but I don’t think it contrasts with the general point that I’m trying to make, which is that if you’re going to dig into the internals of polls like the Q poll and adjust the samples, you need an objective baseline. If you’re going to argue that a poll oversamples Democrats, then point to turnout in previous elections to make that case.

    If you think 2010 is analogous to this year because Republicans are just as enthusiastic this year as they were in 2010, then you need to find evidence that that is the case.

    On the flip side, if you think that 2008 is analogous to this year because it’s a presidential year, then you need to point to evidence that Democrats are as enthusiastic this year as they were in 2008.

    I don’t think that either of those claims stand up to scrutiny. The way I read the data is that since 2010 both the GOP and the Democrats have seen their enthusiasm drop, but GOP enthusiasm has dropped more, which only makes sense since GOP enthusiasm had further to fall. That means that the GOP can count on a higher than average turnout compared to Democrats, but probably not 2010 levels. If you want to tweak the Q poll sample, then tweak away, but 2010 is probably a tweak too far.

    Finally, as your data demonstrates, there isn’t a 1:1 correlation between enthusiasm and electoral margin of victory. After all, independents vote too, and if they break hard for one party it doesn’t matter how enthusiastic an opposition party is (cf. 2004).

    That’s why I think that the GOP would be better served trying to flip Independents than trying to repeat the turnout advantage that they had in 2010 or, even worse, simply assuming that turnout will look like 2010 and ignoring polls that point to the contrary.

  • streiff

    Jay Cost hits it directly. The Q poll is D+11 in FL. FL has a R Gov and 19 of 25 House seats are R. FL was D+9 in 2008. What makes anyone think you will get a higher D turnout in 2012 than in 2008. If you are saying enthusiasm will be the same the sample should be D+9. Really it should be more along the line of D+3 or 4.

  • CrabCakes

    I certainly don’t think that the Democrats will enjoy the same advantage that they did in 2008 unless something drastically changes the race between now and then.

    I think that it will be somewhere between 2008 and 2010, which is why above I stated:

    “Personally, I think that shaving 2 in Ohio, 4-5 in Florida, and 1-2 in Pennsylvania is about right, as I think that turnout will be somewhere between 2008 and 2010.”

    The Q poll actually has Florida at D +9, not +11. Shaving 4-5 in Florida would put leave us with D +4 or 5, which is only a point different than where you come down.

  • streiff

    why was the poll weighted in this Bizarro World manner? Probably to create a story that would attract media attention. All other polls are showing Romney/Obama neck and neck and they come out with one that shows an Obama blow out in critical states.

  • APA Guy

    We’re in the business of REALITY here, not subjective thoughts and feelings. If you’re interested in spreading feelings as “facts”, the Daily Kos is there stage left.

  • CrabCakes

    I’ve read the arguments in favor of ignoring party ID (i.e., it’s a very malleable category), but I’m not 100% sold.

    In any case, I haven’t seen any indication that Quinnipiac is a partisan hack organization, so I’m willing to give them the benefit of the doubt here unless there is evidence of which I’m unaware. That doesn’t mean that I won’t tweak their numbers for my own personal consumption, though.

  • streiff

    nt

  • tnfriendofcoal101368

    is this was Quinnipiac’s first attempt at a likely voter model and the first time they had partnered with the NY Times to collect the data. They’ll probably come in line with other polling as they tweak the model and iron out the samples they are getting. I agree with you 100% Quinnipiac’s tendency is to play it to the middle.

  • Finrod

    Rasmussen polls for party ID separately and uses that to weight their other polls. I don’t know if any other major pollster does that, though.

  • CrabCakes

    they weight based on fairly unchangeable things lik age, race, gender, etc. but not based on party ID. The do ask about party ID, but they ignore it for the purpose of weighting. That allows them at the end of the poll to report that X% of likely voters are Democrats and X% are Republicans, which it appears is what they did.

    That they came up with numbers that don’t seem right to me tells me that something in their weighting and/or sample is a bit awry, or (less likely, in my opinion) that my assumptions about the party breakdown of likely voters are wrong.