FiveThirtyEight founder cheers polls’ ‘most accurate cycle ever’, says no GOP bias despite Dem complaints

Another big red wave that wasn’t seemingly left Nate Silver pleased with the work of fellow pollsters as he took to social media to boast of the “most accurate cycle ever” following the 2022 midterms.

The frustration felt by conservatives who understandably expected their candidates to waltz to Washington on the wings of President Joe Biden’s numerous failings extended into leadership battles at all levels. Despite attempts in the Senate and with the Republican National Committee, the only appreciable victory came in the House where a multi-day vote for speaker ended with potentially impactful concessions to holdouts.

Now, four months removed from the November elections, FiveThirtyEight founder and editor-in-chief Silver bragged about the work of peers in his profession as “the aggregate did NOT have a Republican bias in 2022.”

“Our new POLLSTER RATINGS are out and two big headlines…Polls had their most accurate cycle ever* in the years covered by our comprehensive long-term database. *Technically, tied with 2004,” he wrote sharing a graph depicting the aggregate findings going back to the 1998 midterm elections.

He went on to share another graph that focused on the margin of bias toward one party or the other and argued, “Polls in aggregate did NOT have a Republican bias in 2022. Technically they had a slight *Democratic* bias per our method, but it’s close enough to zero that ‘polls were unbiased’ is probably the right way to think about it.”

“I guess there’s not really a nice way to put this, but a lot of people aren’t interested in accuracy and just want narratives and vibes,” Silver went on. “They’re wrong about how the polls did in 2022, but then again they’re wrong about most things. The polls did great & congrats to the pollsters!”

In the updated pollster ratings, FiveThirtyEight explained, “Our ratings are letter grades that we assign to each pollster based on historical accuracy and transparency.” For example, Trafalgar Group whom many considered to be the most accurate pollster for 2020 was given a “B” with a predictive rating of +0.1 and a mean-reverted bias of R+2.4.

Meanwhile, The New York Times/Siena College received an “A+” with a -1.2 predictive rating and D+1.0 bias. Yet Gallup, with -0.1 and R+0.6. only got a “B+” for their efforts.

Of course, Silver’s reporting showed his own selective biases as he focused on the good numbers from the aggregate and paid less heed to the fact that the combined “weighted-average” of polls accurately predicting the winner was a combined 72 percent across the federal races in 2022, the lowest over the past 24 years. The average had not surpassed 80 percent since the 2010 midterms when TEA Party candidates made their first effort to combat then-President Barack Obama’s agenda.

(Image: FiveThirtyEight)

For FiveThirtyEight’s Nathaniel Rakich, that reality wasn’t at all problematic as he contended, “Polls’ true utility isn’t in telling us who will win, but rather in roughly how close a race is–and, therefore, how confident we should be in the outcome.”

Kevin Haggerty

Comment

We have no tolerance for comments containing violence, racism, profanity, vulgarity, doxing, or discourteous behavior. If a comment is spam, instead of replying to it please click the ∨ icon below and to the right of that comment. Thank you for partnering with us to maintain fruitful conversation.

Latest Articles