A nomadic Twitter user only knows the most accurate American political pundit by his username — @Cityafreaks — or by the dime-sized profile picture that accompanies his written thoughts and, contrary to popular belief, is not Tony Soprano.
The anonymous University of Georgia senior and election aficionado who runs the account chose an actual Italian felon — media tycoon, financial criminal and three-term Prime Minister Silvio Berlusconi — to represent his persona.
The photo is from a year before his tax fraud conviction. Berlusconi is silhouetted against a foggy mountain, flashing an impossibly white smile while his glass chalice of wine perfectly catches the sunlight. The star-shaped reflection bisects the cup, contrasting his glistening teeth and his cartoonishly waxy skin with his jet-black Russian state parka and beady eyes.
Standing to his right — cropped out from City’s profile picture — is an uncomfortable, yet stylish, Vladimir Putin. His cherry red ski jacket jumps off the screen as the brightest object in frame. His head is cocked away from the Italian, and his eyes, hidden behind midnight black aviator sunglasses, are trained on the ground.
City chose the image for its depiction of modern power, and when another poor wonderer sheepishly confused Berlusconi for James Gandolfini’s character, the political connoisseur ripped him, replying that — clearly — “his ass has NOT read about Eurocommunism.”
This serves as an excellent introduction to @Cityafreaks and his eccentric online presence. Even the most avid politics fans stumble onto his profile, and the obscure cultural references and recurring tendency to wrap niche political opinions in Twitter memes leaves them scratching their heads.
One example is from six days before the midterms, when he tweeted four cartoons of anthropomorphic lone wolves and the caption “How it feels being a special election truther in November 2022.” This is a reference to his contrarian — but ultimately correct — assertion that Democratic overperformance in multiple August special elections was evidence that polls predicting Republican gains were systematically undercounting a motivated Democratic base in the wake of Roe v. Wade’s reversal.
His pinned tweet, written the day after one of these special elections, opaquely reads: “approvalcels seething over dobbschads.”
City is a “dobbschad,” a leader of the self-proclaimed “alpha” or “chad” school of punditry. The central tenant of their beliefs is that June’s Dobbs v. Jackson Women’s Health Organization Supreme Court ruling, which overturned nearly 50 years of Roe v. Wade, would motivate high turnout among the enraged Democratic base. Additionally, aggressive bans on abortions by red-state Republicans would sway independent voters — predominantly suburban women — into voting blue. The outcome, in this prediction, would be a much more neutral year than the blood-red wave predicted by the other camp — the approvalcels.
This dominant majority of “incels” argued Joe Biden’s low approval rating was a sign that independent voters who backed him two years ago would break heavily against him at the polls, blaming him for high inflation and other “kitchen table” economic issues.
By November, this was an overwhelming consensus among the pundits and mainstream media. If you casually googled an election night cheat sheet while watching the results roll in, you almost certainly read about how the last two times Democrats went into a midterm with unified control of Washington — as they did this year — they lost 63 seats in the House of Representatives in 2010 and 54 in 1994. You probably also saw that Biden’s approval rating was lower than Barack Obama’s and Bill Clinton’s when they received those respective shellackings. You definitely read some vague lines about how these “historical headwinds” would make Democrats unlikely to hold either chamber of Congress this election.
City and the other “dobbschads” rebuttal? This is not a normal election, and you can’t treat it like one.
It turns out they were completely right.
In many ways, the crash-and-burn flare to this year’s punditry is the culmination of a judgment national media first made in the early hours of Nov. 9, 2016. Reeling from Hillary Clinton’s shocking defeat, they decided the American electorate was stupid and Donald Trump’s culture war populism was a brilliant and effective strategy.
Like children visiting the zoo on a hot summer day, the beltway journalists flocked to diners in Pennsylvania, truck stops in Ohio and state fairs in Wisconsin. They desperately looked for a way to explain the phenomena of white working-class voters and lifelong Democrats who swept Obama into office twice, abandoning the party to back a New York billionaire.
The lesson they took back to D.C. was liberals were out of touch. They were “woke” and unconcerned with the economic issues that “real Americans” care about.
Prominent Democrats such as James Carville — Bill Clinton’s geriatric White House advisor who, 30 years ago, coined the famous slogan “It’s the economy, stupid” — chastised the party for abandoning his winning mantra in favor of an unpopular socially liberal strategy.
“Some of these [Democrats] need to go to a ‘woke’ detox center or something,” he told Judy Woodruff last year.
“Woke” quickly became a right-wing buzzword — featured in attack ads on corporations, politicians, or just anyone slightly to the left of Attila the Hun as conservatives geared up to punish more “elitist” liberals at the polls.
If anyone called the GOP culture war strategy into question after defeats in 2018 and 2020, it was brief and relatively muted. After all, despite their aggregate losses, Republicans beat four incumbent senators in 2018 and significantly overperformed expectations in 2020, leaving President Joe Biden with razor-thin majorities in both legislative chambers
So, as the president went to Congress seeking wins in the face of a shrinking agenda, Republicans regrouped and turned their gaze on the next major target — election day 2021 and the Virginia Governor’s chair.
The Commonwealth has shot blue over the past two decades, and in 2020, Biden carried it by double digits — the first time a Democrat had done so since Franklin D. Roosevelt in 1944. But Democratic support was more tenuous than national groups may have figured.
The struggle for Republican strategists in the Trump era is maintaining their landslide margins and turnout in rural areas while winning back defecting suburbanites turned off by the “Make America Great Again” message. Marrying the “country club” wing — embodied by traditional conservatives such as Mitt Romney — with the nationalistic and anti-establishment Trump side may seem like an impossible contradiction since the populist rhetoric that alienates educated, suburban voters, is exactly what drives overwhelming rural turnout. However, Virginia might have been the most favorable environment for the strategy.
This is largely due to the unique demographics of Northern Virginia and the high propensity of ex-government employees currently living in the Washington, D.C. suburbs. Many of them worked in the Bush administration or for various Republicans in Congress. They may have voted for Biden in 2020 but are still conservatives at heart, and the “right kind” of Republican could win their vote.
That Republican happened to be Glenn Youngkin, who upset former Democratic Gov. Terry McAuliffe by 2% in the traditional bellwether for the upcoming midterms. Youngkin outperformed Trump’s margin by almost 12% and, in a flash, analysts looked to the Old Dominion and saw the makings of a clear Republican romp in 2022.
Youngkin, a private equity manager turned politician, kept Trump at arm’s length but never openly crossed him. He managed to walk a fine line on the former president’s key issue — the validity of the 2020 election — by supporting vague calls for “increased election oversight” without climbing too deep down the rabbit hole. His strategy for winning suburban voters was attacking one key issue: education.
He dressed non-threateningly, donning a fleece vest, button-down and jeans that would have fit in at a high school soccer game. This “concerned parent” persona was central to his campaign’s message of “look what the Democrats have done to our kids.”
He promised to ban mask mandates, but other than that, he kept his policies — and boogeymen — as ambiguous as his statements on the past year’s election. As the campaign reached the stretch drive, he hammered home his closing message — sensible Virginians like him should be appalled “critical race theory” and “woke gender ideology” are so pervasive in our educational system and, in his commonsense view, educators should just be teaching our kids to love this country.
Despite little evidence that this was actually occurring, it seemed to work as a strategy. He won back substantial ground with affluent independents in Northern Virginia, and this new culture fight turned heads among pundits.
After his victory, the media — just like in 2016 — was desperate to understand this new line of attack prior to the midterm elections. In their search, they elevated the Manhattan Institute’s Christopher Rufo as an intellectual conservative trying to build a post-Trump — but still Trumpian — Republican party.
Rufo was the first politico to use the phrase “critical race theory” outside of a graduate-level sociology class and became a leading advocate for banning transgender high school students from playing sports. More importantly, he pushed these as political slam dunks and, over the past 18 months, has been the subject of major profile pieces in traditionally liberal media outlets such as The New York Times, The Washington Post and The New Yorker.
These journalists visited his home, spoke with colleagues and friends, sat down and wrote thousands of words about this man, his beliefs and his strategies without truly grappling with the biggest question — was the culture war crusade a winning strategy? To @Cityafreaks, the answer is clear: No.
Rufo is making a business decision — promoting his brand to win favor among Republicans and engaging in historical revisionism at the same time.
CNN exit polling from 2021 divulges that only 24% of voters said education was the issue they were most concerned with, and “common-sense” Youngkin won this minority by just six points. This election was not a mandate for a conservative social agenda — it was a referendum on an unpopular president.
In a state where Youngkin won by double digits just a year earlier, 53% of voters disapproved of Biden and rejected his party by a margin of 93-7. Biden’s standing hasn’t magically reversed since then — in fact, it’s declined further. CNN’s exit polling from this year tallied his unfavorability at 56%. But this time, Democrats doubled their support from this group — only losing 86-14.
That was the difference in the midterms — voters were grappling with different issues and ultimately decided there was a sticking point that outweighed their frosty feelings towards Biden. It was the abortion question.
Missing the forest for the trees
The two outcomes also validate 25-year-old Lakshya Jain, who postulates that pundits are so often wrong because “we get so few elections, and everyone is trying to predict the last one.”
Jain, a software engineer and computer science lecturer at the University of California-Berkeley, started following politics out of bewilderment at the unsubstantiated narratives spun by Beltway “experts.” He sought out data as an omnipotent counterbalance to Washington’s habitual self-promotion and a direct window into the American electorate. He began building models and his tweets about trends and shifts have earned him nearly 33,000 followers.
But raw statistics independently are not enough, he said. While data scientists see numbers as just numbers, election modelers cannot afford to be this naïve. There is a seller’s market for being wrong — so long as you’re telling consultants what they want to hear.
It isn’t just your guileless grandmother getting duped by the skewed numbers, either. These polling groups walk away with hefty contracts from (apparently) some of the top political minds in the country.
Every cycle, the powerful conservative Super PAC “Our American Century” raises millions of dollars from wealthy donors such as former Republican National Committee finance chair Steve Wynn, pharmaceutical tycoon Fred Eshelman and Donald Trump’s personal action committee. This year, they shelled out over $2.5 million to boost Republican Mehmet Oz in Pennsylvania’s U.S. Senate race and handed over $2 million apiece to Don Bolduc and Tiffany Smiley — Senate hopefuls in New Hampshire and Washington, respectively.
It didn’t work. It wasn’t even close.
Oz lost the open seat by nearly five points, a de facto landslide in the bellwether Keystone State. Bolduc, a Stop the Steal promoter, could barely stop Sen. Maggie Hassan from clinching a double-digit win in New Hamshire, despite winning her last race by just 2,000 votes. Smiley, running in deep blue Washington state, was trounced by five-term Sen. Patty Murray to the tune of nearly 15 points.
While these expenditures seem head-scratching, especially considering how close margins were in states like Wisconsin, a less flashy disbursement — an $89,000 check made out to the Atlanta-based Trafalgar Group — helps explain them.
The Trafalgar Group is an anomaly only possible in the Trump era of irrationality. In 2016, they took a curtain call as the only pollsters to correctly call his upset victory, and founder Robert Cahaly leveraged the win into influence among the Republican establishment and big-money contracts. Even though they incorrectly predicted five battleground states in the 2020 election, their margins were closer than many big names in the industry, and prestigious forecaster FiveThirtyEight named them one of the most accurate pollsters of the cycle.
This is the problem with pundits’ inability to look below the surface-level data — the most important question asked about Trafalgar should not have been what their numbers were, but how they got them.
The group is notoriously opaque when it comes to sharing its methodology. All they vaguely admit to is a data adjustment to account for “social desirability bias.” This, theoretically, allows them to catch the “shy Trump voters” the polls missed in 2016.
But after taking a cursory look at this group, it is obvious Trafalgar doesn’t have a special sauce — they just expect Republicans to overperform every time and reverse-engineer results to show this. While the experts should have understood this, these kangaroo polls intoxicated them — because polling has consistently underestimated Republicans since Donald Trump rode down his golden elevator in July 2015.
It didn’t this time, however, and Trafalgar’s credibility collapsed like a house of cards on a windy day. They predicted statistical dead-heats in the three blowout races their client so heavily invested in, consistently over-projecting Republican support in every corner of the country.
It wasn’t just them, either. Spurred on by low barriers of entry and a potentially huge payday, Republican polling firms flooded the averages with GOP-friendly numbers that weren’t consistent with underlying data.
This was the problem Jain encountered when he built the election model for his amateur political website, Split Ticket. The polling averages did not line up with the prior and concrete data points from the abortion-salient special elections that crafted @Cityafreaks outlook. Observing a diversion in the numbers from traditional, non-partisan polls and the opportunistic hacks, Jain made the decision to split up these numbers into two separate averages and, when he did, he got data that resembled the future results.
This perspective allowed him to build a model which, by his own admission, was not perfect but more accurate than the professionals at FiveThirtyEight. He stressed that polling is the essential foundation of modeling, but the key is good polling.
The key is also to have a broader view that expands beyond the calculator, as Jain does. He understood the electoral evidence that City used to base so much of his argument.
When pro-abortion activists rolled to victory in blood-red Kansas and Democrats overperformed in special election after special election, it should have been clear that the record-high turnout and deep unpopularity of the Republican position would not just go away.
When the dust ultimately cleared, Democrats won the nearly 30% of voters who declared abortion their most important issue by over 50 points. The Democrats, who would only ever talk about taxes and the budget if figures like James Carville had their way, were bailed out of certain electoral defeat by a social issue.
Picking up the pieces
I think the reason so many people hate politics is because it reduces everything to a number, to a narrative and to something simplistic when nothing about the last six years has been simple. Why is it so hard to predict elections? Maybe it’s because we’re not data points – we’re humans with real and complex emotions. The best that pollsters can do is give us tomorrow what we had yesterday—ignoring that we change as humans and societies.
When pundits promote Robert Cahaly and Trafalgar as many major news outlets did, they’re acting like everything is the same as it was six years ago — back when we played Pokémon Go and laughed at Brad Pitt’s character in “The Big Short” for wearing a mask on a plane.
This isn’t even the same country it was seven months ago. When the Supreme Court struck down Roe v. Wade, it was the culmination of a 50-year conservative effort to enact their radically unpopular agenda through the court system. For the first time, abortion became a salient political issue for mainstream voters — not just Christian activists — because real people suffered real consequences. That’s not something that will go away by the next election, either.
The Twitter kids are more than modest about their achievements. As Jain said, he’s just “an engineer who tweets.” He’s able to avoid serious scrutiny because he doesn’t have to appear on cable networks and have past comments ripped apart.
@Cityafreaks, for his part, called it quits. His accuracy grew his follower count from 5,000 to 27,000, and his once-niche community quickly became overwhelming and toxic. He deactivated his account several weeks after the election when he was doxxed by someone in his comment section. The party was over.
Both amateurs know getting it right once doesn’t really mean much in the grand scheme of things. If only everyone who promoted Cahaly and Rufo understood that too.
My advice to the pundits is the same as my father often has for me — have more humility. Don’t act like you know everything when you don’t. The numerous political and social concerns of one person, let alone 209 million American adults, are too much for any one brain or computer to ever fully understand.
Graham Brown is a Sophomore studying Political Science at the University of Wisconsin-Madison. Do you believe political pundits were inept at predicting the 2022 Midterm Elections? Send all comments to firstname.lastname@example.org.