How a 21-year-old college senior became the breakout star of 2018 election forecasting

Primary election night Twitter is about the hype. It’s a hotbed of instant, and at times misguided, analysis prognosticating a historic election in November.

Special elections are no longer just about filling open seats — they are referendums on Trumpism. Primaries aren’t dull affairs to anoint candidates — they are early tests of voter enthusiasm. The 2018 midterms are about the potential of a “blue wave” and Democrats sweeping back into power.

Veteran election watchers are wary of this trend. The Cook Political Report’s Dave Wasserman warns these kinds of tweets could easily veer into “wish-casting,” combining objective forecasting with desired outcomes.

Then there’s G. Elliott Morris.

Morris is a kind of extremely online election watcher tailor-made for a year when Democrats are looking for a blue wave. With more than 22,000 followers, Morris has become a familiar name to seasoned pollsters, national campaign reporters, and #resist Twitter alike. Pinned to the top of his Twitter page is a graph of his 2018 projection model; it’s a straight line shooting up toward a Democratic House takeover in November.

His personal blog, the Crosstab, is chock-full of posts with nearly the same conclusion: All things point to a wave. His forecast model predicts Democrats have a 60 percent chance of winning the House majority back in November. To be clear, Morris says his model operates in a “paradigm of uncertainty.”

He’s been retweeted by Neera Tanden, the president of the Center for American Progress, a major liberal think tank. And unlike some of the more seasoned and cynical election watchers, Morris has a sort of fresh eagerness about him — an earnestness in his analysis that makes the Nate Cohns of the world roll their eyes a little.

That makes sense. Morris is a 21-year-old who just graduated from the University of Texas Austin. He starts his first real job, as a data analyst with the Economist, in July.

“I guess the trope is I fell in love with statistics and math,” he said in a recent interview. “I guess I’m too far gone now.”

Between censuring campaign reporters for their hot takes based on single polls and predicting election results in late-night primaries, Morris tweets about his undergraduate finals and pulling all-nighters with tubs of ice cream. When he says he’s never seen anything like this before, it’s likely because he was around 12 when the Tea Party wave hit in 2010.

A year and a half ago, the polling community was grappling with how things went so wrong. Few predicted Donald Trump’s 2016 win. As the new president reveled in pulling out a win that no one saw coming, pollsters went through a serious reckoning.

“Despite what some analysts are claiming, the polling error in 2016 was not normal, it was not predictable, it’s not something we should say that happens sometimes,” said Drew Linzer, director and chief scientist of the polling website Civiqs. “It was a systematic error across the entire industry.”

Linzer and his colleagues in the industry say the real problem was that pollsters (and political scientists, pundits, and campaign reporters) were so sure that Trump wouldn’t win that they didn’t add caveats to their predictions of a Clinton victory.

“The polls in 2016 weren’t so much the problem; the interpretation of the polls was the problem,” said Wasserman.

Fast-forward to 2018, where the big political story is energized Democrats. Midterms are typically a good year for whichever party is out of power, but there seems to be a unique backlash building to Trump, due to his seemingly never-ending parade of scandals and historically low approval rating. The tide appears to be turning; the fervor around a Democratic comeback narrative is too big to ignore.

As Morris points out, the generic ballot shows Democrats in a good position to take back a majority in the House, and the Trump backlash has the potential to amplify this significantly.

It’s not just the generic ballot that looks good; Democrats have pulled out a number of startling wins in special elections including Doug Jones winning the Alabama Senate race and Conor Lamb winning a special congressional election in a Pennsylvania district that Trump won by nearly 20 points.

Morris says that matters. He writes:

First, most generic polls are polls of registered voters, but special election voters aren’t the same as all the registered voters nationwide; according to the current numbers, they’re likely about 10 points more Democratic-leaning. Congressional polls that poll likely voters, or those who are enthusiastic about voting, are much more Democratic. So a polling average that doesn’t include them will underestimate Democrat’s advantage in the national political environment.

His forecast for November takes into account congressional election polling and special elections to Congress, and state legislative elections. At the beginning of May, those three metrics together had him predicting Democrats beating Republicans on the generic ballot by 7 to 13 points — potentially enough to flip control of the House.

Wasserman, a veteran pollster and election watcher, said the polls in 2018 have so far been accurate; they lined up well with the results of 2017 and 2018 special elections. He says he still has faith in polls and the ability of the polling aggregate “to point us in the right direction.”

Right now, that direction is pointing to a wave.

“To me, the hallmark of polling in 2018 has been the enthusiasm gap between the parties,” said Wasserman, pointing to another wide enthusiasm gap between college-educated voters and non-college-educated voters. “What that tells me is Democrats are going to get their biggest gains between suburban districts. I think it’s safe to say we’re going to see higher turnout than in 2014.”

But Wasserman has a big caution for people in the polling community, including Morris: Don’t mix what the numbers tell you with your personal political beliefs.

“There are an awful lot of pundits who are blatantly partisan and view positive developments for Democrats as positive things for the country,” he said. “I do think that can lead to wish-casting rather than forecasting.”

After all, Wasserman says, “Twitter is a platform that can allow people to claw their way into the field of political analysis from their dorm room.”

Morris knows he’s still learning.

“I’m still young, so I draw a lot of lessons from my professors,” he said.

And there isn’t a blue check mark next to his name — yet. “It’s not a crime that I’m not verified ... but it sure would be nice,” he tells Vox. But he’s mastered the kind of Twitter that gets him noticed in a year when Democrats are eager to read positive notes about their electoral prospects.