Poll Analysis: Romney creeps up a bit

99.3% probability of winning 0.7% probability of winning
Mean of 323 electoral votesMean of 215 electoral votes

Nine new state head-to-head polls have been released in the contest between Presidential hopeful Mitt Romney and President Barack Obama since my previous analysis:

MIDenno Research14-Jun15-Jun6004.04040tie
UTDan Jones15-Jun21-Jun12222.82668R+42

The unsurprising polls include one from Oregon that gives Obama a +8% lead over Romney, and the first poll of the year in Utah that gives Romney a +42% advantage over Obama.

Colorado puts Obama up by +3.6% over Romney. The four current polls weigh solidly in Obama’s favor, as does the overall polling picture:

The fourth Michigan poll released in the past week has the race tied at 40% each. Combined with two other recent polls, Obama has the slightest lead in the state overall, giving him an expected 68% probability of taking the state (if the election was held now).

The second New Hampshire poll in a week has Obama up by a tight +3% over Romney. (Note that I used the Likely Voter results instead of the Registered Voter results given for this poll.) The pair of polls combined gives Obama an 81% chance of taking the state at this point.

Romney takes the most recent North Carolina poll with a weak +3% lead over Obama. That makes five leads in a row for Mitt, all since mid-May:
ObamaRomney26May12-26Jun12North Carolina

Ohio gives Obama a small +3% lead, breaking Romney’s two poll lead streak. The three current polls in that race give Romney a thin 55% probability of taking the state at this point:

Two new Virginia polls split. One goes to Obama by +7% and the other goes to Romney by +4.7%. The weight of the evidence from the five recent polls is that Obama would take the state now with a 83% probability.


I want to again express my gratitude to Sam Minter, who does his own poll aggregation exercise. Sam systematically compared our poll databases, resolved the discrepancies, and sent corrections to me over the past month. Also, I thank the HA commenter who noticed an error in the Nebraska past elections numbers, and persisted in getting through to me when I missed his comment.

Okay…so now, after 100,000 simulated elections, Obama wins 99,310 times and Romney wins 690 times (including the 63 ties). Obama takes (on average) 323 to Romney’s 215 electoral votes. For an election held now, Obama would have a 99.3% probability of winning and Romney, a 0.7% probability of winning.

Since the previous analysis Romney has gained an expected +3 electoral votes and has more than doubled his probability of winning (now), but that works out to less than a 1% probability overall of winning.

Electoral College Map

AlabamaAlaskaArizonaArkansasCaliforniaColoradoConnecticutDelawareFloridaGeorgiaHawaiiIdahoIllinoisIndianaIowaKansasKentuckyLousianaMaineMarylandMassachusettesMichiganMinnesotaMississippiMissouriMontanaNebraskaNevadaNew HampshireNew JerseyNew MexicoNew YorkNorth CarolinaNorth DakotaOhioOklahomaOregonPennsylvaniaRhode IslandSouth CarolinaSouth DakotaTennesseeTexasUtahVermontVirginiaD.C.WashingtonWest VirginiaWisconsinWyoming

Electoral College Map

GeorgiaAlaskaArizonaArkansasCaliforniaColoradoDelawareConnecticutFloridaMississippiAlabamaHawaiiIdahoIllinoisIndianaIowaKansasKentuckyLouisianaMaineMarylandMassachusettsMichiganMinnesotaMissouriMontanaNebraskaNevadaNew HampshireNew JerseyNew MexicoNew YorkNorth CarolinaNorth DakotaOhioOklahomaOregonPennsylvaniaRhode IslandSouth CarolinaSouth DakotaTennesseeTexasUtahVermontVirginiaD.C.WashingtonWest VirginiaWisconsinWyoming

Here is the distribution of electoral votes [FAQ] from the simulations:

Ten most probable electoral vote outcomes for Obama:

  • 318 electoral votes with a 2.50% probability
  • 323 electoral votes with a 2.49% probability
  • 317 electoral votes with a 2.35% probability
  • 328 electoral votes with a 2.29% probability
  • 324 electoral votes with a 2.19% probability
  • 319 electoral votes with a 2.12% probability
  • 329 electoral votes with a 2.11% probability
  • 335 electoral votes with a 2.05% probability
  • 334 electoral votes with a 2.03% probability
  • 327 electoral votes with a 1.94% probability

After 100,000 simulations:

  • Obama wins 99.3%, Romney wins 0.7%.
  • Average (SE) EC votes for Obama: 323.0 (20.8)
  • Average (SE) EC votes for Romney: 215.0 (20.8)
  • Median (95% CI) EC votes for Obama: 324 (280, 361)
  • Median (95% CI) EC votes for Romney: 214 (177, 258)

Each column of this table shows the electoral vote total aggregated by different criteria for the probability of winning a state (Safe=100%, Strong=90%+, Leans=60%+, Weak=50%+):

ThresholdSafe+ Strong+ Leans+ Weak
Safe Obama153
Strong Obama86239
Leans Obama7878317
Weak Obama666323
Weak Romney191919215
Leans Romney3737196
Strong Romney121159
Safe Romney38

This table summarizes results by state. Click on the poll count to see the individual polls included for the state.

84VotespollsVotesObamaRomney% wins% wins
AL91*754 37.8 62.2 0.0100.0
AZ111752 48.4 51.6 27.5 72.5
AR61*679 36.8 63.2 0.0100.0
CA5511465 61.7 38.3100.0 0.0
CO942713 52.0 48.0 92.5 7.5
CT711239 56.8 43.2100.0 0.0
FL2932674 51.3 48.7 83.5 16.5
GA161*404 43.3 56.7 3.0 97.0
HI41*517 64.8 35.2100.0 0.0
IL201*546 61.5 38.5100.0 0.0
IN111*447 45.0 55.0 7.2 92.8
IA621432 50.2 49.8 54.8 45.2
KS61*442 45.0 55.0 6.8 93.2
KY81*528 45.5 54.5 6.7 93.3
LA81*542 41.1 58.9 0.1 99.9
ME21415 58.6 41.4 99.4 0.6
ME111*488 64.8 35.2100.0 0.0
ME211*421 53.7 46.3 85.5 14.5
MD101*792 62.4 37.6100.0 0.0
MA112973 59.7 40.3100.0 0.0
MI1664718 50.5 49.5 67.8 32.2
MN101904 58.1 41.9 99.9 0.1
MS61*717 40.0 60.0 0.0100.0
MO101455 46.2 53.8 11.4 88.6
MT31372 45.2 54.8 9.7 90.3
NE21553 43.4 56.6 1.4 98.6
NE111*389 45.5 54.5 10.3 89.7
NE211*252 49.6 50.4 47.0 53.0
NE311*284 35.9 64.1 0.1 99.9
NV61450 53.3 46.7 84.3 15.7
NH42851 52.2 47.8 81.2 18.8
NJ141947 62.9 37.1100.0 0.0
NM51*417 57.8 42.2 98.8 1.2
NY291758 62.8 37.2100.0 0.0
NC1521217 48.7 51.3 26.9 73.1
ND31569 42.9 57.1 0.8 99.2
OH1831620 49.8 50.2 45.2 54.8
OK71*448 30.4 69.6 0.0100.0
OR71631 54.4 45.6 93.6 6.4
PA2021204 54.6 45.4 98.6 1.4
RI41*495 59.4 40.6 99.8 0.2
SC91*1833 51.7 48.3 85.2 14.8
SD31*442 44.3 55.7 4.7 95.3
TN111*654 46.0 54.0 7.7 92.3
TX381*460 38.9 61.1 0.0100.0
UT611149 27.7 72.3 0.0100.0
VT31*528 67.8 32.2100.0 0.0
VA1353897 51.1 48.9 83.2 16.8
WA1221386 56.2 43.8100.0 0.0
WV51*373 40.8 59.2 0.6 99.4
WI1032157 52.0 48.0 90.4 9.6

* An older poll was used (i.e. no recent polls exist).

Details of the methods are given in the FAQ.

The most recent analysis in this match-up can be found from this page.


  1. 3

    Roger Rabbit spews:

    Rabbit Math

    Romney is handily beating the bank interest rate on savings; but cigaret smoke, which is popular with 38% of Washington voters, has a 50-to-1 advantage over him.

  2. 4

    Michael spews:

    The unsurprising polls include one from Oregon that gives Obama a +8% lead over Romney, and the first poll of the year in Utah that gives Romney a +42% advantage over Obama.

    Just a little bit of a lead in the Bee Hive state. I don’t think I’ve ever seen that lopsided of a poll in a presidential race.

  3. 5

    Serial Conservative spews:

    @ 5

    Michael, you alluded to MO and wishing it were more of a horserace there. As I read McCaskill’s comments today, the Obama campaign doesn’t seem even to have a HQ in that state.

    What happened to the 50 states thingie that Howard Dean put in place?

  4. 6

    Zotz sez: Summertime... and the livin' is easy. spews:

    Darryl and his data are a treasure.

    Nice to observe you’re not a total fucktwit, bob.

  5. 7

    Serial Conservative spews:

    @ 4, and Darryl, I guess:

    Regarding OR, a poll about a month ago had Obama +4.

    Now he’s +8 in the PPP poll.

    However, in the current PPP poll:


    Obama is upside down with independents with 32% approval and 64% disapproval.

    How does one reconcile that? They disapprove but are going to vote for Obama anyway?

    One other interesting tidbit:

    Obama’s approval among whites is 50% and among ‘other’ (non-whites, I guess) it’s 45%. WTF?

  6. 9

    Serial Conservative spews:

    More on OR and why I’m not convinced it’s as blue as the poll results might show.

    I googled and OR registration is 42% Dem and 32% GOP, or a 10 point spread.

    The PPP poll has 46% Dems and 30% GOP, or a 16 point difference. 25% are independents.

    So, if the poll oversamples Dems by 6 points, and if independents in the poll are one-quarter and if that quarter of the poll participants goes for Romney over Obama by 52/30………

    What are we to think of the PPP poll?

    And what are we to think of Oregon?

    I visit http://www.blueoregon.com on occasion.

    Anyone care to weigh in?

  7. 10

    Roger Rabbit spews:

    @7 “They disapprove but are going to vote for Obama anyway?”

    A plausible explanation is they like the alternative even less.

  8. 11

    Serial Conservative spews:

    @ 10

    See numbers in @ 7 and in @ 9 for independents. See the significant overweighting of Dems vs. GOP.

    In a state where there are 25% independents and those independents favor Romney over Obama by 22 points, Obama’s in a shitload of trouble.

  9. 12

    Roger Rabbit spews:

    @11 You sure seem to need a lot of your own cheerleading. Feeling desperate? It’s not even July yet.

  10. 14


    “South Carolina? Really?”

    Yep. It is an old, but very large poll:

    Marist 4 Dec to 6 Dec 2011 on 2107 RV (MOE 2.1), Obama 45.0, Romney 42.0.

    Unfortunately, there have been no proper polls done since then. (One internet poll has been conducted in the state since then, but that fails the inclusion criteria.)

  11. 15


    Serial Conservative @ 7
    “Regarding OR, a poll about a month ago had Obama +4. Now he’s +8 in the PPP poll….”

    There have been 6 Oregon polls pitting Obama against Romney this election season. The spreads are +8%, +3.6%, +11%, +13%, +8.3%, +12%. In 2004, Kerry beat Bush by about +4%. In 2008, Obama beat McCain by ~16%.

    To me, the weight of the evidence suggests that the +8% found in today’s PPP poll is not unreasonable.

  12. 16


    Serial Conservative @ 9,

    “I googled and OR registration is 42% Dem and 32% GOP, or a 10 point spread.”

    Not sure where your Google search landed you, but your numbers are not even in the ball park. The Oregon SOS publishes monthly voter registration information. For May:

    Dems: 55%, Rep: 44%, Other: 2%

    Of course, the poll did not ask people about their party registration, it asked

    “If you are a Democrat, press 1. If a Republican, press 2. If you are an Independent or identify with another party, press 3.”

    Which is a different question. People who self-identify as independent may still register under a particular party label for strategic reasons.

  13. 18

    Serial Conservative spews:

    Darryl @ 16

    I also used your link and again got GOP 32%. So the Dem/GOP spread according to your link is 40/32, vs. the PPP poll spread of 16 points. Add to that how far underwater Obama is with independents, a quarter of the respondents in the poll.

    Your point about the difference between party registration and the answer to the question is noted.

    Curious, now, as to what you think about the PPP poll in OR considering the actual Dem/GOP spread and the ‘independent’ disapproval of Obama.

    Thanks in advance for explaining your math in @ 16.

  14. 19

    Serial Conservative spews:

    Yet another piece questioning strength of Hispanic turnout:


    The evidence can be drawn from the House primaries that took place in states with significant Hispanic populations over the last month, particularly California, New York, and Texas. In contests from Southern California to Spanish Harlem, Hispanic candidates suffered political disappointments because of low turnout from their own voters.

  15. 20

    rhp6033 spews:

    # 16: I’m finding people being more reluctant to express a preference with respect to telephone polling. Antitodal evidence (i.e., comments from friends and relatives) suggests that a large number of people suspect that their preferences are being recorded along with their name and telephone number, which will impact the number of calls/mailings/donation solicitations they will receive in the future. Also, the number of push/pull polls seem to increase in an election year, causing a lot of people to simply hang up on all polls.

    I don’t belive this is a big problem with legitimate pollsters. But the partison polling (testing out key words, push/pull polling) may be having an impact in participation rates, and such an impact may cause further degredation of polling participation as the election nears. I’d be interested in seeing numbers on the number of hang ups/refusals in comparison with other election years.

  16. 21

    Serial Conservative spews:

    @ 20

    Antitodal evidence (i.e., comments from friends and relatives) suggests that a large number of people suspect that their preferences are being recorded …

    Maybe you and those you surround yourself with are awash in paranoia.

    I think the number of all poll types increases in an election year.

  17. 22


    Serial Conservative @18,

    “Also, Darryl, I went to your link. 827K Dems out of 2.049M total. Maybe I missed something but how did you turn that into 55%? I got 40%.”

    Oops…my bad. I missed the second line of numbers. (Note to self…don’t do data analysis right after Drinking Liberally).

    “Curious, now, as to what you think about the PPP poll in OR considering the actual Dem/GOP spread and the ‘independent’ disapproval of Obama.”

    I think the same thing that I think about all polls. Polls provide evidence based on a small, ideally random, sample of voters, and they are subject to sampling variability (a.k.a. the misnomer “sampling error”).

    But since the sampling variability isn’t partisan, it generally won’t affect the results, but will reduce the certainty with which we can make statements. Hence, my use of probability statements and “expected”, “mean”, or “average” electoral vote totals.

    Serial Conservative, your obsession with polls that go against your partisan desires is amusing. It reminds me of the 2004 election contest. Republicans scoured D precincts looking illegal votes (largely, felons who voted prior to getting their voting rights restored). They were CERTAIN this was evidence that Rossi should have won, and demanded a “proportional deduction” of the illegal votes. They were fooling themselves.

    Even if that method had made it past the Frye test (it didn’t), all the Dems had to do was conduct a similar search in R precincts and come up with offsetting proportions (and they did exactly that).

    A scientist would have called bullshit on the whole thing and suggest a proper examination would have required either: (1) scrutinizing every voter in the state to identify every illegal vote, or (2) the more realistic strategy of randomly sampling from all voters in the state, and check everyone in the sample for being illegal voters.

    Likewise, if you want to scrutinize polls so that you can eliminate “outliers”, you must ignore the partisan direction that the poll is off by. Rather, you must either randomly sample polls or examine EVERY poll, and use a standard statistical method (a Chi-Squared Goodness-of-fit test would work) with a (frankly, subjective) cut-off value to determine those you consider “outliers.”

    My philosophy is to accept every poll that matches my well-defined inclusion criteria, even if it “looks” off. That way, my own partisan feelings can have no effect on the outcome of the analysis.

    To summarize: if I have included polls that favor Obama beyond the natural variability expected from surveying a small sample, I am quite confident I have also included polls that favor Romney beyond the natural variability of sampling. I sleep well at night.

  18. 23


    rhp6033 @ 20,

    “I’m finding people being more reluctant to express a preference with respect to telephone polling.”

    I go out of my way to respond to political polls, if only because I want to hear how they conduct the survey.

    The moment it becomes a push poll or I am offered a “free cruise/house-share/vacation” for my participation in the poll, I set the phone down on the table until the robot hangs up. The number of phone lines is limiting resource for these assholes, so I try to tie ‘em up.

  19. 24

    Serial Conservative spews:

    @ 22

    Thanks for the reply.

    Serial Conservative, your obsession with polls that go against your partisan desires is amusing.

    I think I got that from following Al Gore’s lead in 2000, when he only wanted recounts in the four cherry-picked FL counties. I’m pretty sure some of the Supremes snickered before smacking down the FL supremes in the per curiam opinion. They probably didn’t think it was so funny when the FL supremes didn’t catch the hint and they needed to reinvolve themselves in present circumstances.

    But I digress. It’s natural to celebrate the polls that back your candidate and to question those that don’t. People on this site aren’t thrilled with Rasmussen. I wonder why that is?

    To be fair, I also have noticed irregularities in polls that skew red, such as the We Ask America poll in VA that has Romney up there. He may be, but that polling source was the one that had Walker up 12 points in WI just before the recall, so I’m not all that believing that VA is red. Not yet.

    I’ll continue to suggest that OR is closer than the polls indicate. We’ll see how many times Obama decides to swing through Portland, and to pick up big checks in Seattle as long as he’s in the area, this summer. Maybe Obama will have something nice to say about the Seattle Storms.

    I’ll try to amuse you some more in the future, Darryl. I learned lots from what you do. Thanks again.

  20. 25

    polls spews:

    Darryl, it seems like your lefty breathren like ylb are trying to make more out of your analysis than is there….morphing it into November so to speak.
    But thanks for what you do. It stimulates discussion albeit meaningless since the election is not today.

  21. 26

    rhp6033 spews:

    # 23: Good idea. I especially despise the polls that start off saying they are an “opinion poll”, but pretty shortly thereafter turn into a “marketing poll”, and end up being a sales pitch. Those are the ones where, if there is a human being on the other end of the line, I’m inclined to buy a shrill whistle and blow it in their ears. (I wouldn’t do that, though – those are rather lowly paid workers slogging through telemarketing machines and enduring innumberable abuses and hang-ups, it’s not their fault they are being mis-used by their employers).

  22. 27

    rhp6033 spews:

    By the way, I still get a chuckle out of that “Romney creeps up…” headline. With that insepid, closed-tooth attempt at a smile, there certainly is a “creep up” sensation. Reminds one a bit of a snake, saying “c’mon, Eve, just a little bite of the apple will not hurt. You shall not SURELY die!”

  23. 28


    25 – You’re a miserable stinking liar klown.

    If the election were held as of the date of Darryl’s analysis, Obama would win. Overwhelmingly. That’s all it says to my reading..

    I truly look forward to Darryl’s reading of the latest polls to say that all the way to November..

    May you stay forever IRRELEVANT!

  24. 29

    who run Bartertown? spews:

    Polling, at any time, is irrelevant.

    its just something to keep the lemmings entertained and occupied.

  25. 30

    Roger Rabbit spews:

    Romney must be reading his lousy polls, because he’s getting more shrill every day. After telling Republican governors to STFU about the improving economy in their states, he popped off on health care reform by calling Obama immoral for passing it. Really. I’m not kidding. Romney today called health care reform immoral. He must think that sounding like Michelle Bachmann will get him independent votes. That’s exactly the kind of candidate Obama was hoping to run against.

  26. 31

    Roger Rabbit spews:

    I have a challenge for Republicans. When you’re campaigning this summer and fall, do all of the following:

    1. Tell the 22-year-old college student that it’s immoral to be on her parents’ health insurance.

    2. Tell the 46-year-old single mom that it’s immoral for her pre-existing condition to be treated.

    3. Tell the 55-year-old unemployed worker that it’s immoral for her to get preventive screenings if she can’t afford copays.

    Yeah! You guys run on the immorality of making sure Americans have the health care they need! Follow your leader!

  27. 32

    Roger Rabbit spews:

    “UnitedHealth Group’s plan to continue offering some elements of the Affordable Care Act regardless of the fate of the legislation … will increase its costs by as little as one or two percent, one analyst said.”


    Roger Rabbit Commentary: Question? [rabbit puts up paw] Can I ask a question? Thank you! My question is, if these reforms are a good idea and cost insurers almost nothing, why did they wait until the government made them do it? And doesn’t this prove we need more, not less, government regulation of the health insurance industry? Thanks.

  28. 34

    Michael spews:


    According to the insurance companies #1 & #3 are going to stay in place regardless of what The Supremes say tomorrow.


    why did they wait until the government made them do it?

    Because they’re bastards. They’re keeping affordable healthcare act things in place to avoid getting pitchforks stuck in their backsides when they tried to remove them.

  29. 35

    Dan Brown spews:

    @27… “Reminds one a bit of a snake, saying “c’mon, Eve,”

    That’s what a latter day snake is…