Monday, October 29, 2012

Obama 303 to Romney 235, John Marek's 2012 Presidential/Congressional Election Predictions

Overall Prediction Summary:

President: President Obama will be re-elected (303 Obama to 235 Romney).
Senate: Democrats will retain control of the Senate (52 Ds, 47 Rs, 1 I) (The I will caucus w/ the Dems).
House: Republicans will retain control of the House of Representatives (236 Rs, 199 Ds).

Marijuana legalization measures: CO and WA will pass theirs, but it will fail in Oregon (and the Ark. medical marijuana initiative will fail as well).

     I've waited longer than usual to make these predictions mostly because I simply haven't had time, but also because the debates started up and the numbers were affected by them. That being said, I must give the same disclaimer that I always give: my predictions are mostly based off of Nate Silver's projections (, though I do also look at and I throw in a dash of gut instinct.

     Why rely on Nate Silver? Because he called the electoral vote in 2008, and he called most (if not all) of the Senate races in 2010 (not to mention that he called the Republican House takeover, though he was off by around 20 seats). His model, which aggregates all of the polls for each state, weighs every poll based on sample size, whether it is a robo call poll or a live call poll (live call polls are better), and he adjusts them for political bias (ex: Rasmussen is adjusted for its Republican bias, and Public Policy Polling is adjusted, though only slightly, for its Democratic bias). Unlike (which I also like a lot), he also includes historical state trends and economic indicators. If you are a regular reader of, then my predictions will not surprise you much (though I have varied with him slightly in the past).

     The right has been running a smear campaign against Nate Silver and anyone else who gives them information that they don't want, and here's a good response from Paul Krugman on the matter:

     Here's a good article on polls:

     Before I start to break things down state-by-state, I would like to point out one major factor that people do not consider when looking at the polls: Registered voter polls vs. Likely voter polls. While likely voter polls are definitely better to look at, the actual election results (in a Presidential year) will fall somewhere between the likely voter polling and the registered voter polling. Why? Because multimillion dollar Presidential and Congressional campaigns will be focused on turning out the "less likely" voters. This typically benefits Democrats more than Republicans because a) there are more Democrats than Republicans in this country, and, even more so, because  b) because the demographics that vote Democrat are less likely to turnout (meaning that they aren't considered "likely" voters and meaning there are more "less likely" voters for Democratic campaigns to turn out). Because of this, my predictions may actually favor the GOP more than they should, though I'm sure none of them will be excited about what I see coming November 6th. 

     Here's a post showing the difference between likely and registered voter (national) polls:

     Here's a good article/graph explaining President Obama's ground game edge, which is necessary since the Democratic Party has to focus more on turning out "less likely" voters (Republican field is not as effective because more of their people will show up to vote w/o needing as much contact). Don't allow me to overstate things though because Republicans do need field, they run it well, and they do have "less likely" voters that they need to turnout.

     One thing to watch for in this election is that there's a chance that President Obama could win a decisive electoral vote while losing the popular vote. Nate Silver puts this at around a 5% chance of happening, but, while I agree that this won't be likely to happen, I think that the chances of it happening are more than 5%.

Presidential Predictions (state-by-state): Obama 303 to Romney 235

     I am only going to focus on what I see as battleground, or at least potential battleground, states (so, if you don't see a state listed, it is because I believe it is already going one way or the other w/o need for comment). I am going to start with the battleground states that seem like they are already decided, and then I will work my way down to the more contested states:

NM: Obama. Stop calling this a battleground state.

MN: Obama. 1 poll comes out showing a 3 point race, and now people want to call this a battleground state? It won't be.

MO: Romney. Not the battleground state it used to be.

PA: Obama. Polls typically show it closer than the result we see on election day, which is something that also happens in NJ.

AZ: Romney. This state is not a battleground state this year (as I had hoped); however, I expect it to be a battleground state 4 years from now (due to demographic shifts).

WI: Obama. The media wants this to be a battleground state, but it is not. Even the group of voters that re-elected Gov. Scott Walker preferred President Obama to Romney (side note: Democrats should never create special elections...the low turnout kills us).

NC: Romney. We were lucky to pull this one off by ~14,000 votes in '08. While the state's demographics are moving in our favor (highly educated young people and Latinos are moving into it at a high rate), it is a little too early for us to claim this state.

NV: Obama. The ground game in NV is unbelievable, and Romney's campaign is having to chase votes all over rural NV while a 75,000 + strong union and Harry Reid's state party are giving President Obama a stronger edge than even the polls will permit (though they favor him too).

OH: Obama. If you are hoping for another 4 years of President Obama like I am, then this is the state that should help you sleep at night. No Republican has ever won the Presidency without Ohio, and President Obama has polled 3 to 5 points ahead of Romney in this state the entire race (before the 1st debate he was 10 points ahead...and early voting had already begun). The latest numbers I've seen in Ohio (based on polls conducted w/ early voters) show Obama with a 54% to 39% lead.

"Polls of people who say they already have voted show Obama with a lead in many of the states. The Obama campaign, which benefited from early voting in 2008, has focused heavily on urging supporters to vote early in this election as well.
Obama leads Romney 54 percent to 39 percent among voters who already have cast ballots, according to Reuters/Ipsos polling data compiled in recent weeks. The sample size of early voters is 960 people with a credibility interval of plus or minus 3.5 percentage points."

IA: Obama. Polls have looked good (though somewhat close) there, and good voting/election laws there should help out the Democratic turnout (voting is very accessible there).

NH: Obama. The polls are surprisingly close in NH, and I believe that this is one of the few battleground states (maybe the only) that has a registered republican voter advantage. That being said, NH Republicans are moderates, and I expect NH to stick with the President.

Top 3 Battleground States: (If I am to be wrong, it will probably be on one of these 3): 

CO: Obama. This state had a strong lean towards Obama early on, but has since become one of the most contested battleground states. That being said, it still leans very slightly towards President Obama, and I think the "less likely" voters will put him over the top.

VA: Obama.You can pretty much take what I said about CO and apply it here as well. Nate Silver (and recent polls) have shown VA to be a more likely pick up than CO for Obama (though there is only a razor-thin difference). Even though I believe both will go to Obama, my gut instincts tell me that we actually have a better shot at winning CO than VA.

FL: Romney. Unfortunately, Romney has begun to pull away in FL, and, even though Democrats had a 37,000 vote lead in the early vote last week, I expect Romney to win FL by a small margin. That being said, the polls have tightened to a point where a good ground game could put President Obama over the top. I hope to be wrong on my prediction for this state.

Here's a good article from Nate Silver on the "state of the states":

Here's a more recent article from Nate:

For more data check out,,, and (though really only looks at recent polls/is not a polling aggregate).

2012 Senate races: 52 D, 47 R, 1 I

     As you see above, I believe that the Democrats will have 52 seats + 1 independent Senator (Angus King) that I expect to caucus with them (technically, Senator Sanders from VT is also an I that caucuses w/ the Dems), and I believe that the GOP will have 47 seats in the Senate.

Here's the breakdown:

NV: Republican (Dean Heller). Unfortunately, ethics scandals will most likely keep Rep. Berkley from defeating Dean Heller. She may be able to ride President Obama's coattails to victory, but I doubt it based on the polling I've seen. This is especially disappointing since this is a state the Democrats should have picked up a seat in.

MT: Republican (Rehberg). It pains me to predict against Senator Tester, but I have to stick with Nate Silver on this one. More outside money has been dumped into MT's Senate race than they've ever seen in MT, and, even though shows Tester leading by a couple of points (because recent Democratic leaning polls have shown Tester ahead), I fear that we will lose this contest. That being said, if positive polling continues to roll out for Tester in the coming days, I may change this prediction (I always update these a day or two before Election a separate post of course).

AZ: Republican (Flake). This future battleground state is giving us a current battleground race for its U.S. Senate seat. Carmona, former U.S. Surgeon General, wasn't expected to do as well as he has, but he's proven to be a good candidate considering how close the race has become. That being said he made a minor last minute gaffe when he told the moderator in his debate that the moderator was "prettier" than Candy Crowley. I doubt that the gaffe will affect him much at all, but you don't want anything to go wrong when you are in a race as neck-and-neck as this one.

ND: Republican (Berg). This was definitely not a state that anyone expected the Democrats to have a chance in, but the Democrats picked a great candidate and Heitkamp has made this one a close race. She's won on a statewide level in ND twice before (Tax Commissioner/Attorney General), and has her behind by only 1 point. That being said, I'm sticking to Nate Silver again on this one for 2 reasons: a) the state traditionally votes Republican, and b) the polls favoring Heitkamp are from Democratic pollsters. I hope Heitkamp's team can prove me wrong on this one come Nov. 6th, but, even if they don't, they still deserve much praise for forcing the GOP to spend money in ND.

NE: Republican (Fischer). Bob Kerrey, a former NE Senator and Governor, did the Democratic party a huge favor by jumping into this race. Though I don't expect him to win, he is forcing the GOP to spend resources in a state that they shouldn't have to spend them in. That being said, it is a red state and all of the polls favor Fischer so I expect this to be a Republican pickup.

MO: Democrat (McCaskill). GOP candidate Akin's twisted views on women has allowed this state to go from being one of the easiest Senate pickups for the GOP to a safe Democratic seat. Ever since GOP MO Senate candidate Todd Akin made the below quote, his polling numbers have been dismal.

“'It seems to me, from what I understand from doctors, that’s really rare,' Akin told KTVI-TV. 'If it’s a legitimate rape, the female body has ways to try to shut that whole thing down. But let’s assume that maybe that didn’t work or something: I think there should be some punishment, but the punishment ought to be of the rapist, and not attacking the child.'"

Here's his screw up on a graph:

McCaskill was supposed to be the #1 target that the GOP could pick off for a gain in the Senate this election cycle; however, Akin completely dropped the ball to the point where the GOP has given up on the race (though Senator DeMint is still wasting money there last time I checked).

WI: Democrat (Baldwin). Are you a right-winger that thinks WI is a red state because they re-elected Gov. Scott Walker in his recall election? Well think again, WI is poised to elect a liberal lesbian to the U.S. Senate! Even better, the Republicans made a safe bet in going with Tommy Thompson, and even he hasn't been able to overtake Baldwin (he blames it on Romney dragging the ticket down in WI).

WI is the perfect example of the difference between 2010 and 2012. Mid-term election typically have a low turnout (meaning the likely voter polls are very accurate), and in 2010 the voting base was older and whiter than the country they were voting in. Republicans made big gains (though Nate Silver accurately called that Democrats would retain the Senate while Republicans would take the House); however, 2012 will not be a repeat of 2010 for one major reason: Presidential turnout.

IN: Democrat (Donnelly). Here is another race where the Republicans should have had an easy pickup, yet they dropped the ball. First, they made a mistake when they nominated a far-right-wing tea partier (Mourdock) to run against a moderate Democratic nominee (Donnelly). Because Indiana is a red state, Mourdock still had a slight (though very slight) advantage...that is until he said this:

“'I struggled with it myself for a long time, but I came to realize life is that gift from God,' he said. 'And I think even when life begins in that horrible situation of rape, that it is something that God intended to happen.' Mourdock, who became visibly emotional while speaking, supports abortion only to save a mother’s life, not in cases of rape and incest."

The polls still have this as a close race (Nate Silver's model has it going to Mourdock, but that is mostly due to the fact that not much polling has been done on the race since Mourdock made that bone-headed comment), but I'm going with Donnelly on this one. The race was already close, and Mourdock's comment will scare undecided/moderate women into Donnelly's camp.

OH: Democrat (Brown). Senator Sherrod Brown has been polling above Mandel the entire election cycle, even though millions of dollars have been spent against Brown. Ohio appears to have made up its mind on its Presidential and Senate races (though I will admit that they have both tightened up as we've come closer to Election Day...just not close enough to warrant a switch in my predictions).

FL: Democrat (Nelson). Nelson has also polled ahead of his Republican opponent this entire election cycle, and I don't see him having as much trouble in Florida as President Obama will have (see above).

VA: Democrat (Kaine). Senator Jim Webb did the Democratic party a huge favor by stepping down (his polling numbers weren't looking too great), which allowed Tim Kaine, a popular former governor of VA, to run. I expect this one to be close (just like VA will be close for the Presidential election); however, Tim Kaine has been polling ahead of Allen for most of this election cycle and I expect him to slightly outperform President Obama in VA.

PA: Democrat (Casey, Jr.). The GOP picked a far right candidate, who has made some strange comments, to run against a popular moderate Senator. Even though has this race within 5 points, I don't see it being a close one.

CT: Democrat (Murphy). This race had Democrats worried for a minute when McMahon appeared to be pulling ahead in the blue state of CT. However, recent polling (even from Rasmussen, a right leaning pollster) has shown a strong lead for Murphy so don't expect the WWE family to take over the U.S. Senate any time soon.

MA: Democrat (Warren). You wouldn't think there would be much competition in this state, but, because of a special election and a terribly-run Democratic Senate campaign (back in early '10), MA elected moderate Republican Scott Brown to the U.S. Senate. However, Elizabeth Warren is about to take back Ted Kennedy's seat for the Democratic party, and I'm looking forward to it. She's great for the party because the U.S. Consumer Financial Protection Bureau was her brainchild (and it has already penalized many credit card companies for consumer abuse/legal violations), and because her messaging is so on target. She reminds us that the wealthy would not be wealthy if it weren't for the publicly educated workforce, publicly built roads/highways/infrastructure, and publicly funded research and development that we have in this country. I look forward to seeing what legislation she'll propose/pass in the upcoming Congress.

Independent (King). Republican Senator Olympia Snowe stepped down because she was tired of her party's unwillingness to compromise, and, in her place, we will get an independent that is likely to caucus with the Democrats in the U.S. Senate. I say this because he is viewed as being more liberal than anything else, and he is way outpacing both the Democratic and Republican nominees in the polls.

For a look at each race on

For a good article on many of the Senate races:

Of course, don't forget to see Nate Silver's Senate predictions: (click the "Senate Nov. 6th Forecast" tab)

U.S. House of Representatives

     Trying to call the individual House races is painful and somewhat pointless since these races are polled less often, and most of the polls that are conducted for them are internal. Nate Silver appears to have abandoned his House forecasts for the same reason (he correctly called that the Republicans would take the House in 2010, but he was off by a lot on the amount of seats they would take it by...though he beat me :p).

     The only one I truly have insight on is the 9th District of TN, where I am the Campaign Manager for Congressman Steve Cohen. Because of my bias, I will not make an official prediction for this race, but I will say that I expect the great Congressman Steve Cohen to win by a comfortable margin. That being said, we have to run like we're 10 points behind because that is what you do on a winning campaign, and because we happen to have a mini-Mitt Romney/self-funder threatening to spend up to $5 million dollars against us.

     Therefore, I am only going to lay out a blanket prediction for the U.S. House of Representatives, and I'm mostly making an educated guess by relying on's forecast:

     According to this forecast, the Republicans have 231 to the Democrats 189 seats, which leaves 15 toss-up races. I expect the turnout in our Presidential election to help Democrats, and therefore I'm putting 10 of the toss-ups into the D column and 5 in the R column, which leaves us with my prediction: 236 Rs to 199 Ds in the U.S. House of Representatives. Being that I've pretty much made a guess at this one, don't be surprised if I'm off by a few :p.

Here's the "Generic National Congressional" poll aggregate:

Marijuana Ballot Initiatives

     Unfortunately, the polling hasn't been very heavy for these initiatives, but there has been enough polling to make predictions. That and Intrade (, the prediction market based in Dublin, Ireland, has made some pretty one-sided predictions in the three legalization ballot initiatives. Because Intrade typically follows, I trust it, and I am going with it/polls I've seen in making my predictions:

WA legalization ballot initiative: Pass.

CO legalization ballot initiative: Pass.

OR legalization ballot initiative: Fail.

AR medical marijuana ballot initiative: Fail.

     WA and CO learned from CA's mistake, and wrote their initiatives to be more mainstream through the hard work of lawyers who used polling/focus groups' information to craft them (including harsher penalties for "stoned driving" and limits on how much marijuana someone can legally possess), and they've done a better job at funding their initiatives (most likely due to better polling data in both states).

     It will be interesting to see how the federal government reacts to two states legalizing marijuana for the first time, but, even if the federal government reacts negatively, it will force the U.S. Congress to finally have a serious national debate on the current state of our drug laws. How marijuana has remained a schedule 1 drug while much more dangerous substances, such as alcohol and prescription pain killers, are legal baffles me.

     So there you have it. These are my current predictions for November 6th. I may adjust these predictions in a future post 1 or 2 days before the election. The main two I could see myself switching on are FL (from Romney to Obama) and MT's Senate race (from Rehberg to Tester); however, this late in the campaign things are usually fairly set in stone.

     Also, after the election, I plan on doing a post-election article discussing these predictions and several predictions I made a year ago (a year ago I thought the Senate would be a toss-up and that the Democrats would take back the House...redistricting of the House districts and some good fortune for Democrats in the Senate races have changed both of those predictions). 

     While 2012 will not be another 2008 for the Democratic party, it will be much closer to 2008 than 2010 was. Presidential turnout will help President Obama and every other Democrat down ticket (in close races at least), and I expect the likely voter models to slightly underrate Democratic numbers/candidates.