National Journal Column from Vice President Natalie Jackson: Mitt Romney’s ‘unskewing’

Tuesday, October 31st, 2023

With all the excerpts and news coming out of the new biography of Sen. Mitt Romney by McKay Coppins, including his “going out in a blaze of glory” style of honesty about his fellow Republicans, I can’t help but think this is truly the end of an era of politics. Of course, Romney hasn’t said he’s done with politics, but taking a scorched-earth approach to your own colleagues—no matter how deserved it may be—doesn’t exactly say “I’m planning a future in this area.”

I have a soft spot for the senator because the 2012 race was my first foray into election polling. After spending the first 27 years of my life in rural Texas and Oklahoma, suddenly I was in New York, working for the Marist Poll and doing battleground-state polls for NBC News and The Wall Street Journal. Not only was it my first cycle, it was also a really good year for polling. Marist nailed every race we polled.

Recently I’ve been nostalgic for that era—not just because it was when I first felt the adrenaline of electoral work, but also because it was the last time the presidential race didn’t feel like an existential crisis for the country.

It’s worth a trip back to that cycle to see just how much politics and polling have changed since 2012. Back then, we thought “binders full of women” was peak male obliviousness. And we thought the ultimate saying-the-quiet-part-out-loud moment was a leaked video of Romney saying at a fundraiser that 47 percent of voters would vote for President Obama no matter what because they were dependent on the government and paid no income tax.

Romney wasn’t wrong about a large proportion of Americans not paying income tax, although the dependencies part of the statement was not factual—it was rhetoric to fire up Republicans. And he certainly wasn’t wrong about a divided electorate in which 40-some percent will vote for each candidate no matter what.

It feels quaint now, but it was quite the controversy back then. Even Romney himself later noted how much the comments hurt his campaign—which looked reasonably strong at the time. Throughout much of the fall, national polls had indicated that Romney could win, but battleground-state polls showed Obama with the edge.

It was in that context that a conservative commentator got semi-popular by creating a website called, arguing that media polls that looked overly rosy for Obama were intentionally “skewed”—including some that I worked on at Marist. The site’s author would then recalculate what the polls should have said in his view, usually by recalculating what the outcome of the poll would have been if more Republican voters had been interviewed. Thus, the premise that the mainstream media was misleading audiences by skewing polls toward Democrats became widespread even before the polling misses in 2016 and 2020 that underestimated Republican support.

In the end, Romney’s appeals to the Republican base worked. He secured 93 percent of Republicans’ votes in the 2012 election but only got to 47 percent overall, losing by more than 3 million votes. Obama ran the table in battleground states, resulting in a considerable Electoral College win. The polls had been correct just as they were.

That said, of course, we shouldn’t take polls as concrete even when they have been right—they are fuzzy estimates at best, because of all the sampling issues. But in 2012, telephone polling was still generating good estimates of voters and the industry had not yet crashed into Donald Trump.

When Trump came into power four years later, the media polling complex crashed and burned in the same way Romney’s presidential hopes had by indicating Trump would lose, only for him to win narrow victories in enough states to take the Electoral College. Trump has permanently altered everything in politics, including polls struggling to measure his support accurately.

In both cases, the seeds of trouble were already there: Romney had defeated several more-extreme candidates on his way to the Republican nomination. Technology and faltering response rates were already affecting polls. However, something about Trump exacerbated both—the Republican Party is still his to lose despite a litany of controversies that would have sunk anyone else, and public polls haven’t been able to capture Republican support accurately when he’s on the ballot (granted, that has only happened twice, but the industry is 0-for-2).

I don’t know what the future holds for Republicans or polling, but I can definitely say I have more hope for the polling world than Romney seems to have for the Republican Party. His retirement is another reminder that the relatively stable politics of 2012 no longer exists, nor does the environment we polled in back then. We’d all be well-served to adopt Romney’s realism.

– GQR Vice President, Natalie Jackson


This column was originally published on October 24, 2023, on and is owned by and licensed from National Journal Group LLC.