Battleground State Polls: Romney +1 in Florida vs Obama +3 in Colorado, +4 in Ohio, +3 in Virginia — Purple Strategies

The latest from Purple Strategies is up and as always, tons of data for poll junkies.  We’ve blogged their last three releases and this one is a good result for Obama.  The only caveat is Purple strategies uses 12 states as their battlegrounds and 2 are pretty solidly Obama states (Minnesota and New Mexico).  So this will skew the top-line numbers a little Obama ‘s way.  Regardless, Romney-Ryan was leading in the previous survey and now they are trailing.  And the individual state breakdowns are also good news for Team Obama:

  • Obama holds a 5-point lead across the 12 Purple States (49% to 44%), reversing a 1-point lead for Romney-Ryan 1 month ago
  • President Obama now leads among independents across the 12 Purple states
  • This important sizable 5-point spread unsurprisingly mirrors his overall advantage
  • This is the first time lead for Obama among independents across the 12 Purple States in 7 months
  • Silver lining for Romney: 14% of voters say they are either undecided (6%) or open to changing their mind (8%)

Colorado (Obama +3):

  • Obama currently leads 48% to 45%, the same margin he had in the last poll
  • His vote total is down a point, as is Romney’s
  • The gender gap is smaller in this state than elsewhere. Obama leads among men by 1 point, and among women by 5 – a gap of just 4 points

Florida (Romney +1):

  • Romney holds on to a slim 48% to 47% lead in the state, which has tracked toward Obama over the last few months
  • The change is driven by independents, among whom President Obama has a 10-point margin, 52% to 42%

Ohio (Obama +4):

  • Ohio has been one of the more volatile states in our polling, with the lead changing hands almost monthly
  • Obama now leads the state 48% to 44%, despite continuing to trail among independents by 10 points
  • Obama’s strength lies in a more consolidated base, with 90% of Democrats supporting him, compared to 82% of Republicans favoring Romney

Virginia (Obama +3):

  • Virginia remains a key state for both campaigns, and has swung between the two candidates in Purple Strategies polling
  •  Today, Obama leads 46% to 43%, a reversal of Romney’s 3-point lead last month

The economy

  • 34% say the economy is getting better, 5-point improvement from August
  • Forty percent (40%) say it is getting worse (25% staying the same)

Voter perception of performance on the economy is the single greatest predictor of the vote:

  • Among those who say the economy is improving, Obama leads 94% to 4%
  • Among those who say it is getting worse, Romney leads, 86% to 8%
  • The improved (though still low) perception of the economy plays an important explanatory role in the improved performance seen for President Obama across the Purple Poll

Job Approval

  • At 47% job approval, Obama’s rating is as high as it has ever been in the PurplePoll (tied with June and April)
  • Obama continues to struggle to reach 50%, a level which would indicate stronger electoral position
  • Romney still has an opportunity to gain ground



  1. John Fisher
    Posted September 21, 2012 at 1:24 pm | Permalink | Reply

    One small caveat – the poll was taken 9/15 – 9/19 during the Romney video fiasco and before Univision interview and WH contradictions on Libya attacks. I say the good news is R is up in Florida. Purple also has NC going to O which is not going to happen. And I see Rass has the national back to O +1 and Gallup remains even.

  2. Posted September 21, 2012 at 2:42 pm | Permalink | Reply

    what’s the party ID break down of the purple state polls?

  3. Posted September 21, 2012 at 3:44 pm | Permalink | Reply

    There’s 6 states (NC,VA,OH,CO,AZ,FL) in this link, overall Obama winning by 2. D+2 overall, D+5 in NC which is suspect. Also Women are +6 (thought national average was W+1), NC is women +12. Blacks in VA and NC are 20% or higher in the poll, seems high to me, don’t know?

    Basically nothing new, race is tied IMO not counting a republican enthusiasm advantage.

    • Posted September 21, 2012 at 3:59 pm | Permalink | Reply

      Yea, it was silly of them to poll Arizona in my opinion since Obama pulled out of there in March I believe. As for North Carolina, I put no stock in that or any other poll claiming Obama is leading. Just like it is telling Romney is not stumping in Pennsylvania, until Barack Obama is stumping in NC, it ain’t a Battleground. And if those polls were remotely accurate you can best believe Obama would be in NC with all the hoopla of a conquering hero. He isn’t there for a reason.

  4. tim
    Posted September 21, 2012 at 4:59 pm | Permalink | Reply

    New to your blog and enjoy your analysis. Question: I see that Rassmusen has Party ID data that give GOP edge by 4.3 over dems. Is is possible these polls are not taking this into account?

    • Posted September 21, 2012 at 5:10 pm | Permalink | Reply

      Although rasmussen sees a party affiliation of R +4, he runs his national polls at D +1. The different in party weight is the entire basis of my complaints over nearly every poll. Poke around and you’ll see more than a few posts on the subject. The most noteworthy post was this one:

    • sunflower
      Posted September 21, 2012 at 9:01 pm | Permalink | Reply

      tim, ras was on fox recently talking about party ids in polls. (both he and gallup post their monthly and yearly averages for party id.)
      ras weights his polls using as his guide his party id average for the last 6 weeks so his party id weight is changing all the time. the party id for the last poll was D+ 2.
      the average of ras party id for jan thru aug is R+ 2 w/ LV data. (gallup’s is D+ 2 for RV data)
      an interesting note: rasmussen’s party id average for 2008 was D+ 7.2. BO won by 7.3.
      looking at both ras and gallup party id data what we’ve seen since 2008 is a precipitous and steady drop in self identification as democrat.
      using rasmussen’s data: ’08: D+7.2. ’09: D+ 5.4. ’10: D+ 2.3. ’11: R+ .5 ’12: jan-aug: R+2.
      i’ll get you the gallup data tomorrow.
      til then, ttfn

    • sunflower
      Posted September 23, 2012 at 4:14 pm | Permalink | Reply

      tim, here’s the rest of my response from 9.21, specifically the gallup info.

      gallup averages:
      2012: jan-aug: D+2
      2011: D+3.4
      2010 D+2
      2009 D+6
      2008: D+8
      i used this site which lists results of every poll per year and averaged them.

      here’s ras averages:
      Jan-aug 2012 R+ 2
      2011 R+ 0 .5
      2010 D+ 2.3
      2009 D+ 5.4
      2008 D+ 7.2

      and again, ras weights his daily tracking poll for party id using as his guide the party id average from his polls from the previous 6 weeks, so it’s always in flux. Last week is was D+ 2, per ras on thursday

  5. tim
    Posted September 21, 2012 at 5:02 pm | Permalink | Reply

    Also, what do you believe is the realistic party ID advantage for GOP going into 2012? Is their anyway to quantify the enthusiasm factor and apply to polls?

    Just trying to get a idea of the real state of the race.

    • Posted September 21, 2012 at 5:13 pm | Permalink | Reply

      My guess at the moment is D +2. The average over the last 5 election is about D +3.5 and I think antipathy towards Obama and angst over Obamacare will tilt the balance slightly towards Republicans this cycle. For comparison, Jay Cost at The Weekly Standard thinks D +3 and Nate Silver at the New York Times thinks D +5.

      Your second question is much harder and more involved. The short answer to the question is yes (somewhat) but that what top pollsters get paid big bucks to discern. Everyone else is just guessing. Some are just smarter guessers than others.

      • shane
        Posted September 22, 2012 at 12:14 am | Permalink

        If you take the average from gallups graph, that was posted of your blogs, for each election 2000-2010 it works out to D+2. Just presidentials it moves to D+3.

        Which all means the logic is this baby is ties and comes down to.Ohio.and Florida..shocker

  6. Posted September 21, 2012 at 5:05 pm | Permalink | Reply

    this is the greatest blog site I have found this year…thank God…now I can go about educating people as to the skew of these polls. From D+HUGE to even slanting heavy to women and heavy to minority over and above even 08 levels. I have a feeling no sleep for this former political speech writer on election night

  7. AussieMarcus
    Posted September 21, 2012 at 5:41 pm | Permalink | Reply

    Odd numbers…..Obama up among Inds in FL yet losing, while Romney up with Indies in OH yet also behind.

    Plus the NC figures make it questionable..

  8. tim
    Posted September 21, 2012 at 7:09 pm | Permalink | Reply

    Have you seen Nate Silver’s forcast? Why do you think he has Obama @ 70% chance of winning? I know he has a liberal bent, but thought his analysis what more objective than that. Also, any idea why intrade would be @ 70% as well. A bit concerning. Thanks and enjoy your blog.

    • Posted September 21, 2012 at 7:50 pm | Permalink | Reply

      Nate is a good number cruncher but when it comes to his odds making he reminds me of the old axiom: There are three kinds of lies, 1) Lies, 2) Damned Lies, and 3) Statistics. Embedded in each calculation are assumptions he makes. Depending on how you construct your model, you can achieve wild swings in results from small changes in the data. This is why casual observes often say Nate needs to stay in politics because he would get cleaned out on Wall Street or if you could bet against him in Vegas. It’s not that he gets the end result wrong a lot, it’s that his probabilities quite often swing too far one way or the other. The 70% figure you site is a good example. Even if I were predisposed to think Obama will get re-elected I would put the probability no higher than 55%. And this compares to what I think the chances are today which is Mitt Romney 52% chance to win. It’s not only that I disagree with Nate but I strongly disagree with the magnitude by which he believes Obama will win. I probably wouldn’t be at 70% for either candidate until one week out.

      As for Intrade, I don’t have as much info. Some people have argued Intrade often favors Democrats ( in its markets simply because maybe more Demcorats play in that market. May be true. I think it is just the fact that Intrade is hugely dependent on the public information which is always pro-Democrats (ie the liberal media bias). So it always just seems like the Democrats are going to win. And even when the Republicans are going to win by a landslide, Intrade will only show a slight chance for Republicans to win until the last moment.

      A good example of the Intrade bias in in the chance Republicans will retain control of the House. It is a near metaphysical certitude this occurs absent a series of catastrophic events that makes two or three-dozen Republicans suddenly lose races they currently are very likely to win. Yet Intrade only has that probability at 77%, just slightly better than Barack Obama @ 72 who is not above 50% in ANY poll that doesn’t massively over-sample a Democrat turnout that has NEVER happened in American history. Both of those % are way wrong. So if you feel like making a ton of money, buy 1000 of the “GOP keeps the House contracts” @ 77 and sell 1000 of the “Obama wins the Presidency” contracts @ 72. Obama may well win the Presidency in November but he will have at least a few tense moments between now and then pushing his contract price much lower (at least 10-15 points). At the same time the “GOP keeping the House contract” almost certainly will only rise from here making the spread between the two grow much larger. Close the contract any time you think Obama has taken a good dip in the polls and you will enjoy lots of winnings. I am deadly serious about all of that.

  9. Posted September 21, 2012 at 7:43 pm | Permalink | Reply

    70% is actually pretty good for Silver. He had Obama pushing 80% not long ago IIRC.

    Maybe he thinks the race is closing too…..

  10. Tim
    Posted September 21, 2012 at 9:39 pm | Permalink | Reply

    Thanks for the thorough analysis! Would you agree the odds favor Romney based on party ID numbers? If D+2or 3, would Romney win easily?

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: