Tonight, Alabamians will cast their ballots in what’s probably the most consequential election of the year for national politics. Republican Roy Moore and Democratic candidate Doug Jones are in a tight race for the Senate seat currently held by Luther Strange (who was appointed to replace Attorney General Jeff Sessions but lost the special election Republican primary to Moore). Republicans have had trouble getting some of their major policy initiatives through the Senate (e.g. repealing and replacing the Affordable Care Act), and a Jones win would make it more difficult for them to pass policy. Moreover, if Democrats want to retake the Senate in 2018, they need to defend all of their seats, and win three more seats. Arizona and Nevada are competitive, and Alabama could provide them with that third win.
This election is important, so here’s a quick, data-driven guide that’ll get you up to speed on the polls and some of the basics of Alabama politics.
What do the polls say?
There’s a lot of uncertainty in exactly where public opinion is on this race. RealClearPolitics gives Moore a 2.2 point lead, but that doesn’t guarantee a win. Polls can miss. In the Virginia gubernatorial election (about a month ago), Democrat Ralph Northam went into the race with a 3.3 point lead in the polling average and ended up with a 8.9 point lead—a 5.6 point error. In 2014, Senate polls overestimated Democratic candidates’ margin by four points on average and in 2012 they overestimated Republican Senate candidates’ eventual margin. For a more thorough overview of polling error and related issues, I’d recommend this piece, but the point is clear: Moore’s narrow edge in the polls could translate into a victory for either candidate.
Moreover, it’s hard to model the electorate in this race. Nobody knows exactly who is going to show up and vote in this election. Everyone in the world of elections analysis has emphasized this point, but pollster Mark Blumenthal of SurveyMonkey might have demonstrated it most effectively. He applied various assumptions about turnout to his data and came up with wildly different results, ranging from an 8-point Jones win and to a 9-point Moore win. That, along with the differences between automatic and live caller polling, helps explains why the polls currently in the RCP average vary so widely, showing everything from a 9-point advantage for Moore to a 10-point edge for Jones.
There’s some evidence that Roy Moore has gained ground since the Washington Post initially alleged that he had improper sexual contact with teenage girls while he was in his 30s. Moore gained ground in multiple polls after the Thanksgiving holidays, and some (thought not all) polls show him gaining ground since then. Emerson put Moore ahead by +3 in a late November poll, but they now have him ahead by +9. Change Research showed Moore’s lead growing earlier this month, though that seems to have leveled off in their final poll. Gravis and Strategy Research have also shown an increase in Moore’s lead since their last poll.
The verdict here isn’t unanimous. Fox News, a well-respected pollster, found an 8-point advantage for Jones in the aftermath of the Washington Post story, and now they show Jones up by 10 points. Additionally, a number of pollsters (e.g. Monmouth, the Washington Post, Trafalgar (a Republican firm), YouGov) haven’t publicly polled the race more than once, making it impossible to see if there’s a trend within those pollsters.
Moreover, the direction of the polls only matters so much. It’s possible that Moore gained ground since the allegations but hasn’t recovered enough to win. Or maybe Jones gained ground because of the Post story, but it didn’t matter much because the most Moore-friendly polls were predicting the electorate correctly the whole time. These aren’t the only possibilities, but they drive the point home: the position of the polls matters more than the direction. And the disagreement about where the race is indicative of the fact that it’s a true toss-up—which is an amazing fact considering we’re talking about Alabama.
Moore is a scandal-ridden candidate, so how is this race close? The answer: Alabama.
In most states, a Republican like Moore running in these conditions would be all but finished. The president’s approval rating is low enough to put normal Republicans in normal states in danger—and Moore isn’t a normal Republican. Multiple women have accused him of sexual misconduct or assault. (In addition to more pedestrian accusations of corruption, such as the $180,000 salary Moore took for part-time work from a “charity” he founded that also employed his wife and two of his children.) But this isn’t most states—this is Alabama.
Most elections analysts (myself included) have already explained in one way or another that Alabama is very conservative and very inelastic, but it’s key to understanding the state’s politics, so I’ll quickly review. Basically there are a lot of very reliable Republicans (many of whom are white evangelicals), very reliable Democrats (many of whom are African-American) and few third party voters in Alabama. That adds up to a situation where Republicans have an advantage and elections bounce around in a very narrow range, as both Democrats and Republicans stay loyal to their parties.
This graphic shows the 2016 two-party vote share in every state as well as the standard deviation of the last five presidential election results. In other words, states where results vary more are higher up on the graphic, states that vote more Republican are further to the right.
Alabama is the orange point near the corner—it’s a very Republican state where results don’t vary much from election to election, despite national conditions varying widely (think of the difference between Bush’s 2004 re-election and Obama’s 2008 landslide win). This inelasticity shows up in presidential approval polls as well as election results. Morning Consult recently measured Trump approval in each state and compared their January and September numbers. Alabama was one of the states where Trump’s approval had declined the least, dipping only from 62.2 percent to 59.1 percent while Trump’s approval rating in the median state dropped by 6 points.
In other words, Alabama is one of the best possible states for Republicans. The fact that Moore is in danger of losing this contest shows that he’s a problematic candidate.
If you want to keep diving, there are great pieces out there about what a winning coalition might look like for Jones in Alabama, in-depth looks at the polling and demographics, guides on interpreting the results, and great reporting from on the ground. But hopefully this quick look at the data helps you understand some of the wonkier details and watch the results more effectively tonight.
