Princeton Election Consortium

A first draft of electoral history. Since 2004

Election Night liveblogging, 7:00PM

November 4th, 2014, 7:09pm by Sam Wang


See below the fold for older commentary. The most recent comment will appear up top.

8:40pm: The Upshot has projected counts. For now, use those for your Geek’s Guide. Shaheen (D-NH) around +5% and McConnell (R-KY) around +13%, both ahead of their pre-election polls. Ambiguous for estimating Delta.

8:27pm: Reader Forrest asked me how The Upshot estimates vote share from partial returns. I can’t say what they are doing, but look at Jay Boice’s HuffPollster calculation. Basically take the prior history of the state, county by county (or whatever level of granularity you have available). Then slide over all the counts in past comparable elections, and see how each county would have to break in order to reach a 50-50 tie. Use that as an over/under, i.e. calculate whether a candidate is over/underperforming that expectation. Then do a weighted average across counties. That is an estimator of the margin between candidates.

8:17pm: Virginia Senate race looking close. But recall that in 2012, Romney led Obama until around midnight. This is going to take a while.

7:44pm: Everyone, please note: rural areas tend to come in earlier. That means early returns will tilt GOP. That’s not so useful for projecting results.

7:38pm: Kentucky’s projecting around McConnell+7% at the moment, close to polls. Actually, less sure. Stand by…

Virginia counts are screwy. Might be technical errors on the ground there.

7:06pm: CNN has called Kentucky for McConnell (R). It is too early to say whether he ran ahead or behind the pre-election polls of McConnell +7.5 +/- 0.6%. If he does, it will be by a few percentage points either way.

Commenter Chandra says CNN has McConnell+17%. Don’t pay attention to those raw counts! Uncorrected, that’s of no use in estimating the final margin. I think it will come in super-close to pre-election polls.

Tags: 2014 Election

84 Comments so far ↓

  • Michael

    Feeling a little better about the Old Dominion. Margin is down to 6400 votes and 40% of Fairfax still hasn’t reported. 40% of Fairfax is +12,000 votes for Warner, even with this low turnout.

  • Edward G. Talbot

    Fairfax county results:
    http://www.fairfaxcounty.gov/elections/webreports/resu1114.pdf

    substituting these numbers for what’s on HuffPo, Warner still comes up short.

  • Chandra

    CNN calls NH for Shaheen (D) and the spread is at 2%, same as the polls…

  • Jena

    VA is mind boggling. They just got rid of a few wingnuts & a corrupt Governor. Yet here we go again.

  • Reason

    They are already talking recount like in ’06 here.

  • A New Jersey Farmer

    Roberts now leading in KS.

  • Froggy

    Anyone have a good read on the FL gov results? If the numbers on the HuffPo site are accurate, I’m not seeing much chance of Crist making up the difference in what remains of the South Florida counties.

  • Michael

    VA is going to be ridiculously close. The margin is about 23k right now and remaining Fairfax votes will make that up, but turnout seems very low and if Warner does pull it out, it won’t be by much.

  • Reason

    All that may not matter if VA goes R. Which is shocking.

  • 538 Refugee

    I have to get up early so I’m out. Maybe I’ll wake up to “Dewey Beats Truman”. ;)

    • Reason

      Or Gillispie Beats Warner. :(

    • Mikey Z

      That would indeed be pitbull bites man. I am hoping for Truman, myself, as in the Dems take Iowa, Colorado and, in the shockers of the night, Georgia and Alaska…

  • Edward G. Talbot

    ABC calls NH for shaheen. And Hagan performing closer to her 2008 numbers (she won by 9) than to the baseline 50+1 victory numbers.

    So ironically enough, no real surprises in actual results yet, just margins in some places. If IA or CO gets called early, those are probably the two remaining potential Senate surprises.

  • Reason

    I am a little worried about VA. Even Sabato says he is shocked it is this close now. I cannot look at results as the Va State Board of Elections site is not working.

    • J

      I am surprised at Virginia as well, but the results there are being skewed by the more Republican areas getting their votes tallied first. The Democratic areas in Northern Virginia are slower to count their ballots. Once they come through, I think we will see Warner pulling it closer than expected, but winning by roughly 5-8 points.

    • Reason

      I hope you are right. We are all a bit nervous here. Warner was crushing it in poll after poll here. But then again,so was McAuliffe this time last year. He only won by 2%.

  • Chandra

    FiveThirtyeight is calling NH to Shaheen (D)

  • Chandra

    Looks like Florida is going to the wire (as always!)…Crist (D) is dependent on numbers from Broward and Miami….

    With 77% in for KY, the spread still holds strong at 13% for McConnell.

  • Edward G. Talbot

    75% of Kentucky in, including most of Louisville. McConnell is going to remain in double digits here. Seems like VA will skew similarly towards Republicans compared to the polling. GA, NH and NC sound like at this point they are not, but too early to be sure.

  • SureshM

    Looks like Kay Hagan is exceeding expectations !

  • A New Jersey Farmer

    Cotton projected to win AR.

  • Jacob

    The Upshot has McConnell projected to win by ~16%. That’s…a little bit off from the pre-election projections. Their other figures seem plausible. This has to be an error in their calculation, right?

  • J

    It looks like CNN and Fox News have declared Arkansas for Cotton without bothering to wait for the vote count. Unlike WV, MT, or SD, I would have thought they could have waited a bit before making any “declarations”. Obviously not.

  • Olav Grinde

    Sam, any thoughts on North Carolina?

    With 38 % of the vote counted, Hagan leads over Tillis by 52 % to 45 %.

  • Michael

    Fairfax County is 16% of Virginia’s population. Only 10% of Fairfax County has been counted so far. Kaine won 53-47 in 2012, but if I recall correctly, didn’t move ahead in the raw count until almost 11 pm.

  • Kansas Man

    Sam:

    For Florida, has Broward County votes come in yet. Looks like Crist is losing..

  • A New Jersey Farmer

    Hagen winning big in Raleigh, Fayetteville and Charlotte.

  • Froggy

    It’s early, but things are looking very good so far in NH. Shaheen is performing about as well as in 2008, when she won 51.6-45.2.

  • Kenny Johnson

    The Upshot is currently showing an “adjusted lead” for Warner and Shaheen.

    I haven’t read how they make their adjustments, but I assume they are trying to make projections on the uncounted counties by looking at what’s already been counted.

    • register registry

      It doesn’t seem to very accurate. The upshot projections are changing almost as much as the real results.

  • Lana Frez

    Sam Wang’s prediction may be as good as Dick Morris’s are in 2012 with Mitt Romney. How bout a Bug Sandwich.

  • A New Jersey Farmer

    Checking in on Virginia via the NYT site. Not much in yet in the DC suburbs. Fairfax at only 7% counted.

  • Forrest Collman

    Sam, i’ve long wanted someone to make what the upshot appears to be doing tonight, but i’m frustrated at the lack of transparency as to what precisely they are doing. Do you know? For instance are they using turnout numbers from previous years, or doing something that just scales up the turnout for that county based upon the votes reported this year and the % precincts reporting? Are they dividing the remaining precincts based upon this years votes, historical data for that county, or some combination?

    Why haven’t they calculated even a simplistic error bar based upon poission statistics of the remaining votes and the anticipated percentage…

    I want to see that error bar shrinking over time as the result narrows, and see how early error estimates hold up.

    • Sam Wang

      I can’t say what they are doing, but look at Jay Boice’s HuffPollster calculation.

      Basically take the prior history of the state, county by county, and slide over all the counts to see how each county would have to break in order to reach a 50-50 tie. Use that as an over/under, i.e. calculate whether a candidate is over/underperforming that expectation. Then do a weighted average across counties.

    • Forrest Collman

      Hmm.. yeah that looks interesting and totally reasonable, but they aren’t giving me that calculation in a live updating form, so its not so useful for election night viewing.. or am i missing a link where they are doing that?

  • A New Jersey Farmer

    Hagan and Shaheen ahead. The spreads from urban areas will be coming in soon. By 8:30 we’ll have a good sense from these two states.

  • Chris

    Nothing about CT yet. Incredible how close this rematch is.

  • JayBoy2k

    Sam,
    MSNBC just stated that Warner in VA is still OK because the early voting from rural counties was expected. Also the call for McConnell was made on the weakness of normally democratic areas.

    If the early votes are rural, why is the delta over the previous midterm good for predicting?

  • Chaz

    Am I reading the GA exit polls wrong? It seems to show that Nunn wins based on M/F split.

  • Trey

    If the raw margin reported by CNN etc. is not what we should be looking at, then what SHOULD we be looking at for margins if we’re trying to follow along as instructed in the PEC geek’s guide?

    • Sam Wang

      Unfortunately I think it takes a few hours for raw counts to shape up. See The Upshot for their estimator of win margins? Other suggestions welcome.

  • Edward G. Talbot

    39% in now in KY. 150K-200K votes left from Jefferson, and if they go at the rate of the Jefferson votes so far (which is similar to 2008), that would take the race to around 6 points. But the rest of the state will probably pull it back out to 8-9.

    So not far enough off 7.5 percent to make a conclusion yet. We need to either hit 65% or so without the rest of Jefferson or get the rest of Jefferson before we could really get a better sense.

  • Andy

    Does Huffington’s graphic showing McConnell largely underperforming Romney mean anything, or is it noise?

    • Edward G. Talbot

      Romney won Kentucky by 23, so McConnell underperforming him doesn’t mean much.

      At this point it seems like he will be closer to 10 than the 7.5 the polls predicted, however.

  • Kansas Man

    Barry

    You can always call any southern GOP states easily. All Dem states will be relatively closer. You have to wait till they count northern VA

    • Barry. Rubinowitz

      He’ll win, but probably by 4 or 5 points rather than 9 or 10. In Sam’s list, he was a more probable winner than McConnell, therefore the call of McConnell’s win without the call of Warner’s indicates vote erosion from the polling, which, as Sam has pointed, is a very bad sign for the erodees.

  • Lana Frez

    Warner will win Virginia. He is a good goyin democrat who will vote Yea for the Trans Pacific Partnership, however tonight should be a big night for Senator Scott Brown. Senator Shaheen is done. Maybe it is better that way, now Shaheen can take good care of her diabetic grand daughter. You know, family helps. Tonight Professor Wang will be eating a bug.

  • Kansas Man

    GA will go to OT and if it does Dems have an outside chance.

  • Edward G. Talbot

    Regarding exit polls, they don’t count early voters, so they don’t mean as much as they once did. But those GA numbers. .wow!

    • wendy fleet

      Is that a good (from the progressive pt of view) Wow or a dismal Wow??

    • Edward G. Talbot

      wendy – I didn’t intend it as a good or bad from a progressive point of view, just wow that it was that close.

      I suppose I’d have to say it’s positive since recent polls showed it getting away from Nunn. But the exit polls could still be off by 2-3 points and Perdue could still win outright tonight.

  • Barry. Rubinowitz

    They were able to call KY immediately but not VA – not a good sign.

  • Lana Frez

    Looks like Professor Wang was wrong. Republicans after this election is over are going to be 55 to 45. Trans Pacific Partnership trade deal, here we come. Now TPP can come to reality. Republicans win. Hurray

  • Joe

    What do you make of this? CNN Exit poll of GA: Perdue 48.8, Nunn 48.4….Deal 48.6, Carter 48.3.

  • Chandra

    CNN Exit poll for NC has 3% spread for Hagan (D)… Interesting…

  • Kansas Man

    VA will not change till they start counting North

  • Edward G. Talbot

    Virginia being too early to call doesn’t bode well for overall Dem performance.

    • Sam Wang

      Counts in VA are screwy at the moment, at least where I am. Also, R-leaning areas report first there. 2012 and 2008 were late nights for people watching VA.

    • Edward G. Talbot

      Very true that VA is usually late due to the R-counties reporting first.

    • Suja

      VA will be okay. The population density and blue areas are all up North, and don’t report until late. Fairfax, Arlington, Alexandria, Loudon – that’s where the votes are.

  • Edward G. Talbot

    He won by 6 in 2008 and you can look at how it compares county by county here:
    http://elections.huffingtonpost.com/2014/results

    In most counties, McConnell is outperforming 2008 with 12% in. However, worth noting that almost 20% of the votes so far are from Lexington where Grimes is outperforming Lunsford’s 2008 perf by maybe 1% with 36% reporting. Too early to tell about how close it will be to 7.5%.

  • Joe

    Not surprising. The spread was just too much to overcome. For the polls to be wrong and for her to pull an upset would have been monumental. GA not being called as too close is interesting. No one said who was really ahead.

  • Andy

    OMG! Scott Brown with 60% of the vote in…Dixville. Officially too close to call at 3-2. Although Dixville is a bellwether for greater Coos County.

    • AySz88

      Is that one the ceremonial first-count precinct?

    • Cal State Disneyland

      Dixville found a couple of votes on the floor. Shaheen now up a little bit.

    • Andy

      Was just having a bit of fun so I don’t set my hair on fire. I know this comment is irrelevant to everyone here, but I’m angry at Dems for refusing to stand up and give a full-throated embrace of what Obama has managed to accomplish…as Biden did in Debate #2. I’ll always love him for that.

  • Chandra

    The current CNN % gap between McConnell(R) and Grimes (D) is at about 17% which is way above any of the predictions. However it appears lot of metro areas numbers haven’t come across. May be the numbers may change as time goes by…