What Tuesday’s election might tell us about the future of politics

President Trump on stage

Turnout in Tuesday’s election could be another signal that the old rules of politics don’t apply, Professor Laura Stoker thinks.  (Photo via Gage Skidmore/Flickr)

If you want one simple way to understand what’s happening in American politics before Tuesday’s election, Laura Stoker, a UC Berkeley associate professor of political science, has your answer: The generic ballot.

“It’s probably the number one best predictor of the midterm outcomes,” said Stoker, who specializes in understanding changing political attitudes and behavior.

The generic ballot is a simple question put to voters across the country almost daily: If the congressional election were held today, would you vote for the Democrat or the Republican? Aggregate all those results, as sites like FiveThiryEight do, and you can get a pretty good sense of voter sentiment, Stoker said.

Structural disadvantages for Democrats like super concentrations of liberal voters in urban areas and gerrymandering, mean that Republicans will likely retain control of the House if Democrats get 54 or 55 percent of the two-party vote. But if they get more, the scales can start to tip in a dramatic way.

Election 2018

Professor Robert Reich doesn’t think Dems can take back the House

“Every percentage point above seven or maybe eight in the Democratic vote advantage has an increasing consequence for the balance of power in the House,” she said. “More votes bring more seats, but at a 6- to 7- point advantage, it starts to speed up a little bit. In the 8 point and above range, every additional 1 percent should bring well more than 1 percent of the seats.”

(The current estimate on FiveThirtyEight is that Democrats lead by 8.5 percentage points.)

But things are never that simple, Stoker adds. A big source of uncertainty is who will actually show up at the polls. Pollsters always struggle to estimate turnout, especially as both Democrats and Republicans make pitches designed to encourage unlikely voters — the ones polls sometimes ignore — to vote.

a mug shot of stoker

Laura Stoker specializes in in understanding changing political attitudes and behavior.

“All signs point to higher turnout in 2018, but the real question is whether that turnout will advantage one party or another. Shifts in the party’s demographic bases are adding to the uncertainty,” Stoker said.

Democratic voters are getting younger and Republican voters are older, and older voters are more likely to vote. But at the same time, highly educated voters are increasingly supporting Democratic candidates while less educated voters are supporting Republicans, Stoker said.

On top of that, growth in the gender gap since 2016 adds uncertainty. Are the women seemingly turning against President Trump and the Republican Party likely to vote —and vote Democratic — on Election Day?

“That’s one of the big question marks for me,” Stoker said.

Stoker, and an array of other experts, will discuss Tuesday’s election during a Friday panel, “Blue Wave or Red Wall?” which runs from noon to 1:30 p.m. in 820 Barrows Hall. In advance of that talk, Berkeley News sat down with Stoker to discuss what matters – and what doesn’t – five days before the midterm election.


Let’s talk about the generic ballot. Why is that so important?

It’s probably the number one best predictor of the midterm outcomes. In particular, we know that Democrats in general face a structural disadvantage because of two things. One is the way that Democrats are concentrated in districts and states, and the way district boundaries are drawn, i.e. gerrymandering. We know that they need a certain advantage to overcome that. That’s why you always hear that they need anywhere from a 4 to 7 percentage point advantage in the generic ballot to even win 50 percent of House elections.

This election is deeply nationalized.”

– Laura Stoker

This election is deeply nationalized. For that reason, Trump’s approval rating is also a great predictor of election outcomes, more so than say what Reagan’s approval was or what Clinton’s approval was.

Exactly what’s going to happen depends on how big that Democratic advantage is in voter turnout. And since our polls are national and there are 435 races, it’s possible for any number of things to happen. Democrats could win by 5 percent of the vote and still not get a majority depending on exactly where those votes are in the 435 districts.

Every percentage point above seven or maybe eight in the Democratic vote advantage has an increasing consequence for the balance of power in the House. As you go from 50/50 to a 5- to 6-percentage-point Democratic advantage, the seat gain is linear, but at a 6- to 7-point advantage, it starts to speed up a little bit. At the 8 point and above range, every additional 1 percent brings well more than 1 percent of the seats.

Can the generic ballot tell us anything about the Senate or the races for governor?

No. Everything we know about Senate races comes from statewide polling. Same with governors. The generic ballot asks about House races and it works because there are House elections nationwide.

I think a lot of people feel burned by what they thought were bad polls in the 2016 race. Are you thinking about that? Why can we trust polls in 2018?

The claim that the 2016 national polls were wrong is exaggerated. They had Clinton ahead and she won by three million votes! Still, people give too much attention to national polling. National polls are not good when the election of a president depends on an electoral college victory, on what is happening in the states. We need more high quality polls in the 50 states, or at least the battleground states.

A lot of uncertainty doesn’t come from polling being broken but from having no clue who’s going to vote. It’s about this turnout model. Remember when any poll is reporting their results they’re doing it for registered voters or likely voters. And there’s always this big gap between your guess as to who’s going to vote and who actually votes. If you’ve got a bad likely voter model you’re going to have bad predictions.

We have very little basis for making guesses about this. It turns out that early voting is not terribly predictive of anything, though it is up this year compared to the last midterm election. Pollsters routinely ask about voter enthusiasm in order to glean whether one or another party is going to have a turnout advantage. There’s no question that enthusiasm is up. And there’s not a huge Democratic advantage there. There’s a bit of a one, but nothing huge.

a "keep america great again" sign

Voter turnout will the deciding factor in Tuesday’s election, Stoker said. (Photo via Gage Skidmore/Flickr)

What are you watching in terms of voter behavior and long-term trends? What could we learn on Tuesday?

There has been a reorganization of the party bases in terms of the highly educated and the less educated. The less educated Americans are increasingly gravitating toward the Republicans and the more educated Americans are increasingly gravitating toward the Democrats.

That is a fundamental change that could erase the long-term disparity in turnout between Republicans and Democrats. Democrats have almost always been disadvantaged by lower turnout of their base than the Republicans. That difference is typically much bigger in midterm elections — though less so when Republicans are in power, as now.

But my guess is that that’s changing and that means the Democratic disadvantage might be a thing of the past.

Another thing I’ll be watching is the voting behavior of young people. The general pattern has been for young people to come into the electorate and be more supportive of whichever party holds the presidency. But, young people started gravitating to the Democrats in 2006, under Bush, and grew even more pro-Democratic under Obama. What’s very unusual is that has continued under Trump. If anything, we’re seeing an exacerbation of the age divide under Trump. That’s new.

I’m also going to be paying attention to whether the Democrats succeeded in mobilizing young people to get out and vote. Obama did in 2008. It can be done.

We talk a lot about suburban women voting against Trump.

That’s one of the big question marks for me. We’ve seen in the last three years that there’s a massive growth in the gender gap between Democrats and Republicans.

Women make up 50 percent of the population and often make up more than 50 percent of the voters. When you start seeing a massive shift in that gender divide, the question is whether we are seeing a new era of party dominance.

What that could be suggesting is maybe a 5-point edge in the starting point for Democrats, or maybe even a 6-point edge. Going back to that generic ballot, if the starting point isn’t 50/50, but the starting point is 55 to 45 or something like that, that’s a big change. And I think that might be happening.

Reporters interested in speaking with Laura Stoker can reach Will Kane at willkane@berkeley.edu