Alex J. Rouhandeh

By Alex J. Rouhandeh

National Political Correspondent

0

Share

Newsweek is a Trust Project member

See more of our trusted coverage when you search.Prefer Newsweek on Googleto see more of our trusted coverage when you search.


Seven days after winning the Democratic Primary in New Hampshire, Congressman Chris Pappas had important news to share: “BREAKING: The Cook Political Report just moved New Hampshire’s Senate race to Toss Up.”

Never mind that the state hasn’t voted for a Republican presidential candidate since 2000. Never mind it’s currently represented by two Democratic senators, Jeanne Shaheen and Maggie Hassan. Never mind four of the nation’s premier quantitative race raters favored Democrats. Cook—one of the top qualitative raters in the business—declared Pappas endangered, so he got to work.

Speaking to camera, Pappas posted a vertical video with a screenshot of the rating on the screen. “The Cook Political Report just changed the rating for our race to toss-up. That means this is a highly competitive contest,” he said. “I hope I can count on your help. Go to ChrisPappas.org, stand with us and let’s win in November.” Of course, when one visits ChrisPappas.org, a fundraising pop-up appears.

Read More on News

“I’ve never seen that before,” Adam Carlson, an election pollster and founder of Zenith Research, told Newsweek. “I’ve seen all the fundraising emails being like, ‘We just polled the race, there’s a dead heat,’ but I’ve never seen it on, like, social media before.”

A Newsweek graphic.

...

Ratings have long been important to elections. They’re a key metric by which donors decide what races are worth funding, journalists decide which races are worth covering and prediction markets decide the worth of payouts. But with the rise of those markets and increased eyeballs on downballot races in the age of President Donald Trump, these ratings have become a greater topic of discussion and scrutiny as political observers debate how informative they really are and what methods should best be used to calculate them. Since the 1980s, independent firms have rated election races on a scale that generally fluctuates from Solid Democrat/Republican to Likely D/R to Lean D/R to Toss-Up. Alongside The Cook Political Report, Inside Elections and Sabato’s Crystal Ball produce these ratings. Each differs in approach but, at their core, the ratings are derived qualitatively, drawing on vote history, candidate strength, national mood, polling and interviews.

What’s injected controversy into the current moment is, as Carlson noted, the increasing prevalence with which the ratings are cited, coupled with discrepancies between raters. However, it’s not just differences among the three raters for New Hampshire and Michigan that have drawn attention. Debate in the rating community is also heightened by the fact that a relatively new group of quantitative raters has come out with very different predictions this cycle.

The Forecast Divide

Quantitative raters use statistical, data-driven models to make their forecasts, relying more heavily on polls and the “generic ballot,” which measures general support for Democrats and Republicans. Leaders in this space include FiftyPlusOne, Silver Bulletin, Split Ticket and VoteHub. These four entities uniformly predict a rosier environment for Democrats this cycle. While the three qualitative raters generally place the Senate elections in Alaska, Maine, Ohio and Texas in competitive territory ranging from Lean Democratic to Tilt Republican as of press time, the four quantitative raters have those four races as leaning Democrat.

“When you talk to quantitative forecasters, what you’ll hear us speak a lot about is that this just looks like a really blue environment from these generic ballot polls,” Zachary Donnini, head of data science at VoteHub, told Newsweek. “Some of those polls are registered voter polls, so when we do our model adjustments for likely voters, these polls get even bluer. So, we’re just very much infatuated with this national polling, and I think that’s pushing the model in that direction.”

Donnini said VoteHub’s model aims to pinpoint those who are most likely to show up on election day. The model gives less weight to polls of registered voters and likely voters in favor of those who will show up, which, in midterm years without Trump on the ballot, has proven to be Democrats, Donnini said. When that reality is put against the generic ballot, which has Democrats up eight to nine points, quantitative forecasters end up with a very blue environment.

Nate Silver of Silver Bulletin, G. Elliott Morris of FiftyPlusOne and Lakshya Jain of Split Ticket all argue that qualitative raters—also referred to as “expert raters”—are not weighing the generic ballot heavily enough. This doubling down on their positions has added pressure this cycle for raters to deliver. Nathan Gonzales, editor and publisher for Inside Elections—which offers the reddest prediction of the three experts—suggested too much weight is being afforded to the national mood.

“I think the reason why our ratings are lagging behind the narrative is that when you look at data in these individual races, these races are very close,” Gonzales told Newsweek. “It’s closer than the president’s job approval rating, the generic ballot, the direction of the country, right track-wrong track. The individual races are still close, so our ratings are trying to reflect a balance between where is a race now and where do we think it’s going to end up.”

Related Story

How the Data Center Dividend Can Make Shiny Happy People

How the Data Center Dividend Can Make Shiny Happy People

Gonzales noted that while expert raters do not base their ratings on data-driven models, they do utilize data. Some polling data is provided to raters that is not made public. Each of the raters publicly identifies their entities as independent and nonpartisan, making decisions accordingly. However, that does not mean they will ignore credible insights uniquely provided to them.

Kyle Kondik, managing editor of Sabato’s Crystal Ball, said Democrats in likely or safe Republican races this cycle have been arguing that they should be in more competitive categories. He said there has been a flurry of internal Democratic polls this cycle showing impressive polling leads. Republicans, meanwhile, have released few internal polls. When it comes to assessing the numbers for ratings, Kondik told Newsweek these internal surveys often overestimate performance. So, unless raters have enough solid information about which direction a race may land ahead of election day, they keep it a toss-up.

Ultimately, despite the different forecasts, this remains a friendly competition. Donnini complimented the expert raters for being more cautious in reacting to polls, resulting in less frequent rating changes, which he called a “huge benefit” to their offerings. Kondik and Gonzales both said it is beneficial to the space to have more entities coming in and offering innovative ways to assess elections.

What’s deemed valuable, though, will depend on who offers the most accurate picture of what takes place on November 3, and in today’s swift-moving economy, whichever entity can prove that now could reap benefits until proven otherwise.

“All of us are telling the same story,” Kondik said. “I think if this thing ends up being like a blowout and Democrats blow through 235 in the House, you could say the models were faster on that than the experts were. I think that’d probably be a fair assessment. If it’s not that, well, maybe the...kind of more steady approach of the [experts] was better reflective of what ultimately happened. It also may be the day before the election that the forecast isn’t that different.”

View 0 Comments

Request Reprint & Licensing

Submit Correction

View Editorial & AI Guidelines