We Are Going Live Today

We are going live today, with real data. The map now reflects the actual state of the Senate election in states where at least one reliable poll exists. For DE, HI, IN, MA, MD, MT, NE, RI, VT, and WY, we will use the 2012 election results until a reliable public poll exists. Of those, only IN and MT are competitive.

On the subject of reliable polls, things are complicated. Most pollsters are really campaign consultants. Their business model is getting hired by candidates to help them win elections. Usually they are not at all shy about announcing this. If the pollster's Website has a page that says: "We have helped elect 28 Democrats to federal and state office," you can guess which team they are on. We don't trust any of these partisan pollsters and don't include their results in our database. Their polling may well be pretty good (after all, candidates want to know how well they are doing), but the results provided to the public may consist of the polling results seasoned with a heaping tablespoon of fairy dust. There are also pollsters with a miserable track record (Rasmussen comes to mind). Fortunately, Nate Silver at fivethirtyeight.com has compared previous polls from many pollsters taken just before elections and graded them. We won't use any with a grade below C+ (goodbye SurveyMonkey, which gets a D-).

After chucking partisan and incompetent pollsters, we still have 204 pollsters in our list, most of them unrated because they haven't published any polls just before an election. Some of our 204 are established companies like Mason-Dixon and SurveyUSA, which run polls for media outlets. Others are universities like Quinnipiac and Marist College, which have established track records and know what they are doing. Still, there are a lot of small colleges that are new to polling and which are starting to publish polls. Our view is "innocent until proven guilty"—that is, we assume they know what they are doing until it is clear they don't.

Polling is hard for a variety of reasons. First, response rates are under 10%. Most people just hang up. Call a mother of 5 children at 7 p.m. and get a response? You have achieved a miracle! Call a lonely elderly widow at 2 p.m. Sunday afternoon? Be prepared to stay on the line for an hour to hear what her grandchildren are up to. Differential response rates like this can introduce massive selection bias.

Second, it is not legal for pollsters to use computers to call cell phones. Leaving out cell phones removes many young people from the sample, but having a human call them is expensive.

Third, midterms have notoriously low turnouts. In 2016, 58% of eligible voters cast a ballot. In 2014, it was 36%. Pollsters are not interested in talking to people who aren't going to vote. However, it turns out asking them if they plan to vote or not isn't a good way to tell. The better pollsters have a good "likely voter" screen, which is generally their secret sauce (e.g., "Do you know where your polling place is?")

Fourth, sample sizes are typically 500-1000 valid interviews (= 5000 to 10,000 calls). With such a small sample, the pollster might get too many poor, older, working-class Republican Asian women and not enough young, rich, Democratic black men with college degrees. Statistical techniques can correct for bad luck in the sample, but that requires a good model of what the final electorate looks like. If the pollster assumes the electorate will have, say, 37% Republicans and it ends up having 33%, the poll will be way off. The newer entrants into polling may not have very good models.

This being the case, averaging multiple polls is likely to give a better result than a single poll. But should one average today's poll with one taken 2 weeks ago? How about a month ago? Our algorithm is to use the most recent poll and any others taken within a week of it. Older ones aren't averaged in because a week is a long time in politics. So, hopefully you now see we have been forced to make some choices (which polls to use, how to average them, etc.). Other Websites tracking the Senate may make different choices and thus get different results. All we can say is that at least we have been fully transparent about which polls we use and how we compute our daily number for each race. (V)

What We Have to Offer

There is more on this site than just the score and the map. The menu to the left of the map is worth examining. The "All Senate Candidates" link gives a quick rundown of each race, with photos of the candidates and links to their campaign sites and more.

The "Senate polling data" link gets you to the raw data, so you can run your own analyses (e.g., in Excel). It is available in four formats: A formatted web page with the numbers, a page of graphs, flat ASCII text, and in .csv format.

The "Control of the Senate" link is a time series of how many seats each party was projected to have on every day from Jan. 1, 2018 to today, based on the polls available then. It shows the trend. Two versions are given: The top one considers being ahead of the other candidate (e.g., 46% to 45%) as a win. The bottom one considers it a win only if you are ahead by at least 5 points. Statistical ties (i.e., a difference of under 5 points) don't count here.

The "Tipping-point-state" is an important page, but it needs some explanation. The Democrats can win the Senate if they hold their own seats and dip a bit into red territory, but how far? That is what this page shows. Suppose the Democrats win ONLY Vermont (the bluest state with a 2018 Senate race). Then they will have Vermont plus their 23 holdover senators for a total of 24 seats, as shown in the "Dem seats" column of the tipping-point page. Now suppose they do a bit better and hold their top 10 states according to today's polls (see column 5)—in other words, all the states as easy as California (#10 on the page). This will give them a total of 33. Probably you can now see where we are heading. The states are sorted from most blue (according to the current Senate polls) to most red (column entitled "Lead"). To capture the Senate the easiest way, the Democrats have to win all the states starting at the top until they hit 51 in column 6. This special state is marked by the hand icon. In terms of the competitive states, the Democrats' easiest path is to win the fairly blue states, and also Wisconsin, Nevada, Montana, Missouri, Arizona, and Florida. If they win all those, they can lose Texas and North Dakota. As new polls come in, this page will change every day. For example, if tomorrow a new poll shows Sen. Heidi Heitkamp (D-ND) with a 3-point lead, then North Dakota will be inserted between Arizona and Florida, and Arizona will then become the tipping-point state (and get the hand icon).

The page also shows the easiest way for the Republicans to get an absolute majority of 51, just in case Vice President Pence is off at a foreign funeral in a country where Donald Trump is persona non grata. Here, just read up from the bottom until the "GOP seats" column hits 51 (or 50 if you want to hope no foreign leaders die at inconvenient times).

As an aside, columns 6 and 7 never add up to 100. If, for example, the Democrats win Florida to get to 51, the Republicans have to go further into blue territory because we have just assigned Florida to the Democrats.

If you have a blog or Website and would like your visitors to see the map and score every day, go to the Icons for bloggers page and copy and paste a few lines of HTML onto your page. Just do it one time. From then on, your visitors will see the current map and score without you having to do anything again. With a one-time copy-and-paste operation, your page will automatically always have the latest map and scores every day.

The other menu items are self explanatory. If you like this Website, please tell your family and friends about it. Thank you. (V)

New Woodward Book Is Brutal for Trump

Bob Woodward, of Watergate "Woodward and Bernstein" fame, has a new book about the Trump presidency, Fear: Trump in the White House, coming out on Sept. 11. Presumably that date is just a coincidence. In any event, advance copies are now in the hands of reporters across the country, and every one of them says the book is a killer. The headline on the review written by the Washington Post's Aaron Blake, for example: "The most damning portrait of Trump's presidency yet—by far."

Among the claims in the book:

• Sometimes, staffers take papers from Trump's desk and hide them so he can't sign them. For example, former economic adviser Gary Cohn spirited away a letter that would have withdrawn the U.S. from a key trade agreement with South Korea, potentially jeopardizing a top-secret national security program that allows North Korean missiles to be tracked. "I stole it off his desk," Cohn told a friend. "I wouldn't let him see it. He's never going to see that document. Got to protect the country."
  
  
• Former Russiagate lawyer John Dowd tried to lead Trump through a mock Mueller interview, and it went...badly, as the President stumbled, hemmed and hawed, contradicted himself, and told outright lies. "You are not a good witness. Mr. President, I'm afraid I just can't help you," Dowd said. "There's no way you can get through these...Don't testify. It's either that or an orange jump suit." Dowd resigned the next day.
  
  
• Trump, on Jeff Sessions: "This guy is mentally retarded. He's this dumb Southerner...He couldn't even be a one-person country lawyer down in Alabama."
  
  
• After a chemical weapons attack in Syria, Trump wanted to assassinate Syrian President Bashar al-Assad. "Let's fu**ing kill him! Let's go in. Let's kill the fu**ing lot of them," Trump declared. Sec. of Defense James Mattis said he'd get right on it, and then told his staff that they most certainly would not be invading Syria.
  
  
• Mattis also told several people that Trump's understanding of things is comparable to that of "a fifth- or sixth-grader."
  
  
• Chief of Staff John Kelly, on Trump: "He's an idiot. It's pointless to try to convince him of anything. He's gone off the rails. We're in Crazytown. I don't even know why any of us are here. This is the worst job I've ever had."
  
  
• Trump insisted to subordinates that John McCain took an early release from the Hanoi Hilton, granted to him because his father was an admiral. In fact, McCain famously refused early release.
  
  
• After Trump's post-Charlottesville speech, in which he tried to clean up the "both sides are to blame" remark, he told aides "That was the biggest fu**ing mistake I've made" and the "worst speech I've ever given."
  
  
• When Rudy Giuliani defended Trump following the pu**ygate tape, the then-candidate was underwhelmed. "Rudy, you're a baby. I've never seen a worse defense of me in my life. They took your diaper off right there. You're like a little baby that needed to be changed. When are you going to be a man?"
  
  
• After his staff, including then-Secretary of State Rex Tillerson, tried to explain why it is necessary for the U.S. to keep troops in South Korea, Trump was unimpressed. "So, Mr. President," Cohn said to Trump, "what would you need in the region to sleep well at night?" Trump's response: "I wouldn't need a fu**ing thing. And I'd sleep like a baby." It was immediately after the President left this meeting that Tillerson famously said, "He's a fu**ing moron."
  
  
• Trump carefully studies his tweets to see which ones are most popular. He's concluded that the more shocking, the better. The President also views himself as the "Ernest Hemingway" of Twitter.
  
  
• Steve Bannon, while in the White House, insulted the hated Ivanka Trump by calling her a "staffer." She shouted back, in tears: "I'm not a staffer! I'll never be a staffer. I'm the first daughter."
  
  

Woodward's ultimate conclusions are that the White House is in a state of "nervous breakdown," that Trump's staffers are compelled to engage in "nothing less than an administrative coup d'etat" in order to rein in his worst impulses, and that Trump's attitude and demeanor are reminiscent of Richard Nixon during his final months in office.

Needless to say, the White House is denying everything. Once the juicy bits of the book began to leak out, the press office issued a statement via Twitter:

.⁦@PressSec⁩ responds to Woodward book from the White House in this statement: pic.twitter.com/Jke7OoemJ5

— Jeff Zeleny (@jeffzeleny) September 4, 2018

Trump himself said much the same:

The Woodward book has already been refuted and discredited by General (Secretary of Defense) James Mattis and General (Chief of Staff) John Kelly. Their quotes were made up frauds, a con on the public. Likewise other stories and quotes. Woodward is a Dem operative? Notice timing?

— Donald J. Trump (@realDonaldTrump) September 4, 2018

The already discredited Woodward book, so many lies and phony sources, has me calling Jeff Sessions “mentally retarded” and “a dumb southerner.” I said NEITHER, never used those terms on anyone, including Jeff, and being a southerner is a GREAT thing. He made this up to divide!

— Donald J. Trump (@realDonaldTrump) September 5, 2018

Trump also got Woodward on the phone to harangue him, and to complain that he was not even asked for an interview. Remarkably, the President allowed Woodward to record the phone call (transcript here). It really should be read in full, but it's clear that Woodward tried very hard to get a Trump interview, and that in some cases (like, when Sen. Lindsey Graham, R-SC, passed Woodward's request along) Trump forgot about it, and in other cases (like, when Kellyanne Conway was asked to pass Woodward's request along) staffers neglected to abide by their promises, and then lied about it. In short, if a five-minute phone call can substantively confirm the general tone and tenor of the book, then this five-minute phone call did the job.

This is, of course, the third high-profile exposé of the Trump White House in the last six months. The previous two, by Michael Wolff and Omarosa Manigault Newman, potentially had some credibility issues because of their authors' pasts and the way in which some of the information was obtained (Wolff under false pretenses, Manigault Newman through surreptitious recordings). Woodward, however, is the dean of political journalists, and there is nothing shady about how he got his information. Further, we now have three books that all tell the same basic story. A story that is also consistent, incidentally, with daily reporting out of the White House. As we know well by now, there is a certain segment of the public who won't believe anything negative about Trump, but for anyone else, this has got to affect their thinking as they cast their midterm ballots. (Z)

Massachusetts Goes to the Polls

Voters in Massachusetts cast their primary ballots on Tuesday night. There wasn't likely to be too much drama in a state that is as blue as the Bay State, and is something of the headquarters of the Trump resistance (well, East Coast headquarters, at least). Still, there was one interesting development.

In the Senate race, Elizabeth Warren (D) was renominated. If she hadn't been, it would have been pretty embarrassing, since she was unopposed. She will now move on to crush State Rep. Geoff Diehl (R). In the governor's race, incumbent Charlie Baker (R) fought off a not-too-serious challenge to claim the right to run for a second term. He will face former Deval Patrick cabinet member Jay Gonzalez (D). Given that Democratic turnout was more than double that of the Republicans on Tuesday night (488,000 to 242,400), Baker has to be at least a little nervous. That said, the handful of polls that have been done give Baker a hefty lead over Gonzalez (30-plus points), so maybe he won't be so nervous, after all.

The top-of-the-ballot races, then, were somewhat bland. In the House races, however, is where the drama was. While there were some fairly high-profile progressive challengers (Brianna Wu, Tahirah Amatul-Wadud, etc.) who went down to defeat at the hands of incumbents, there was also one huge upset. Ayanna Pressley, the first woman of color ever to serve on the Boston city council, knocked off 10-term Rep. Michael Capuano (D) in MA-07, and it wasn't even close (59% to 41%). Capuano is actually pretty lefty, as is apropos for a D+34 district, but Pressley is even leftier. Given that she knocked off an established incumbent, is avowedly progressive, and is a woman of color, there will be many comparisons drawn to Alexandria Ocasio-Cortez in New York. The most important of these, however, is probably this: Both show that progressives can win in deep blue districts, but do not tell us much beyond that.

As we move into September, primary season is winding down. Delaware is up tomorrow, and then in the next week a few more of the original thirteen colonies will take their turn: New Hampshire (Sept. 11), Rhode Island (Sept. 12), and New York (Sept. 13). Once the Empire State's primary voters have their say, then it will be full speed ahead to the general election on November 6. (Z)

Former Senator Jon Kyl Appointed to McCain's Seat

Gov. Doug Ducey (R-AZ) went for short-term gain at the possible expense of long-term pain with his choice of former Arizona senator Jon Kyl (R) to replace the late John McCain. Kyl served 18 years in the Senate (1995-2013), so he can skip the orientation course for newbie senators. Ducey was in a bind. Arizona has plenty of fire-breathing right wingers, who wanted a fire-breathing right winger in the seat of McCain, whom they loathe because once in a blue moon he went mavericky. On the other hand, Arizona is trending purple and picking a fire-breathing right winger would be the surest way to turn the state blue. Also, Ducey promised McCain's widow that he would appoint someone acceptable to her. Kyl is highly respected in the state and plenty conservative, but he is no fire-breathing right winger. Ideologically, he is about in the same place as McCain, on the whole, but without the mavericky part. His appointment won't get much criticism and he can obviously work well with Majority Leader Mitch McConnell (R-KY), since Kyl used to be the whip (the #2 position).

Now, here's the rub. Kyl (76) has promised he will serve until Jan. 2, 2019. After that, no promises. Maybe he will resign on Jan. 2, maybe he will stay in the Senate until the special election in November 2020 to fill out McCain's term (which ends in Jan. 2023). He has made it clear that he will definitely not run in the 2020 special election. This means there will be an open seat in a presidential election year, when Democratic turnout is high. So by picking Kyl, Ducey will avoid brickbats now, but will be giving the Democrats a good shot at the seat in 2020.

What happens in Jan. 2019 may depend on what happens in Nov. 2018. If Rep. Martha McSally (R-AZ) loses her race to replace Sen. Jeff Flake (R-AZ), Kyl could resign and Ducey could appoint her in January, knowing that she will run in 2020. However, appointing someone the Arizona voters just said they don't want won't look good. Besides, if McSally couldn't make it in 2018, is there any reason to think she can make it in 2020 when the electorate will be bluer? (V)

Kavanaugh Hearings Even More Theatrical Than Expected

Brett Kavanaugh has finished his first day of being grilled by the Senate Judiciary Committee. It was quite a show, according to those who were there. At the beginning, a number of women appeared in the gallery wearing costumes from the book/show "A Handmaid's Tale" to suggest what will happen to women's rights if Kavanaugh is confirmed. Later, Fred Guttenberg—father of one of the students killed at Parkland High School—approached Kavanaugh to introduce himself, and pointedly got the cold shoulder. "I guess he did not want to deal with the reality of gun violence," Guttenberg tweeted.

As predicted, Democrats did a fair bit of posturing. They complained (quite rightly) about the fact that the White House waited until the last minute to release 42,000 pages of documents. The folks running for re-election in 2018/president in 2020 also got their two cents in. "I cannot support your nomination," said Sen Cory Booker (D-NJ). "I am concerned whether you would treat every American equally," declared Sen. Kamala Harris (D-CA). "If you are confirmed after this truncated and concealed process, there will always be a taint, there will always be an asterisk after your name," opined Sen. Richard Blumenthal (D-CT).

Also as predicted, Republicans did a fair bit of deflecting and/or hewing to the party line. "This is shaping up to be the hypocrisy hearing," said Lindsey Graham. "Stop the charades," demanded Sen. Ben Sasse (R-NE). "I believe this fight is nothing more and nothing less than an attempt by our democratic colleagues to relitigate the 2016 presidential election," was Sen. Ted Cruz's (R-TX) take. Naturally, Donald Trump also felt the need to weigh in, via Twitter:

The Brett Kavanaugh hearings for the future Justice of the Supreme Court are truly a display of how mean, angry, and despicable the other side is. They will say anything, and are only....

— Donald J. Trump (@realDonaldTrump) September 4, 2018

....looking to inflict pain and embarrassment to one of the most highly renowned jurists to ever appear before Congress. So sad to see!

— Donald J. Trump (@realDonaldTrump) September 4, 2018

People can decide for themselves about most of Trump's assertions, but the "most highly renowned jurists" is particularly...interesting, given that Congress has—over the years—received visits from John Marshall, Oliver Wendell Holmes, Charles Evans Hughes, Louis Brandeis, William Brennan, Thurgood Marshall, Hugo Black, Joseph Story, and Learned Hand, among others.

In the end, the most important story of day one is this: Kavanaugh didn't say anything stupid. In fact, he didn't say much at all. Unless that changes, then all of this is pretty meaningless. (Z)

North Carolina Will Use Gerrymandered Map in November

The map that North Carolina uses for its congressional districts is illegally gerrymandered. That has been established, and affirmed, in court several times. However, the Republican folks who run the state have fought tooth and nail against that conclusion, and have also done everything possible to drag the process out. And on Tuesday, they scored a victory, when a federal court ruled that there is no longer time to draw a new map, and so the illegal map will be used for this year's elections.

It is not terribly easy to figure out the exact impact of this decision, so we can only make a crude estimate. At the moment, 10 of 13 representatives from North Carolina are Republicans. That is 77%. In 2016, Donald Trump claimed 50% of the vote, compared to 47% for Hillary Clinton, which implies that something like half of the state's voters lean Republican (even though, like many purplish states, North Carolina actually has slightly more registered Democrats). Anyhow, the presidential result suggests that fairer maps would result in 7 Republican and 6 Democratic representatives. This doesn't account for the impact of incumbency (which should favor Republicans) or a potential blue wave (which should favor Democrats), but our crude back-of-the-envelope math says that the best guess is that Tuesday's ruling probably cost the blue team three seats they otherwise would have claimed from the GOP in November. (Z)

Generic Poll: Democrats Up by 14 Points

In a new WaPo/ABC poll, registered voters prefer a generic Democrat to a generic Republican by a margin of 52% to 38%, a 14-point gap. Experts believe that if the Democrats win the national House vote by 7 or 8 points, they will capture the House. With a 14-point margin, it is very likely the House will flip.

The generic House poll has been all over the map this year. Furthermore, Generic Democrat vs. Generic Republican is not on the ballot in any of the 435 House districts. A voter might well prefer a generic Democrat, but think the actual one in his or her district is a crook, or vice versa. Finally, the election is still 2 months away, which is a lifetime in politics.

Virtually all Democrats (96%) prefer a generic Democrat and virtually all Republicans (92%) prefer a generic Republican. However, among independents, 50% prefer a Democrat and only 32% prefer a Republican, which could spell trouble for the GOP. That said, the most important factor in the midterms is turnout. Republicans, especially older white Republicans, are excellent voters. Come what may, they always vote. Democrats are extremely picky voters. It's raining? Don't vote. Have a dentist appointment that day? I'll vote in 2020. Need to do grocery shopping? Eating is more important than voting. Among Latinos, a group that skews strongly Democratic, midterm turnout is typically 20%. Among 18-29 year-olds, 25% would be a huge turnout. So it probably will come down to how well the Democrats turn out their base. (V)

Today's Senate Polls

Every day the last item will be the newest polls, sometimes with comments and sometimes without. For example, Marist separated the likely voters from the registered voters, and we are using the likely voters numbers. Also, Marist asked the question twice, once with the Libertarian Party and Green Party in the list and once without. We will use the numbers from three- and four-way races where available since the presence of minor parties can have an influence by taking more votes from the Democrats or from the Republicans, but we will not report on the numbers of the minor parties. History shows that people tell pollsters they will vote for a minor party, but in the end, most don't. The only candidates not belonging to a major party that we will report are Sens. Angus King (I-ME) and Bernie Sanders (I-VT), since both are going to cruise to landslide victories. In the scorekeeping, we will put them in with the Democrats, since both of them caucus with the Democrats. (V)

State	Democrat	D %	Republican	R %	Start	End	Pollster
Florida	Bill Nelson*	47%	Rick Scott	47%	Aug 29	Aug 30	St. Pete Polls
Missouri	Claire McCaskill*	44%	Josh Hawley	40%	Aug 25	Aug 28	Marist Coll.

* Denotes incumbent

Email a link to a friend or share:

---The Votemaster and Zenger