A Columbia University sociologist, Musa el-Gharbi, says we shouldn’t trust poll results that claim to show millions and millions of Republicans are crazy (although lots are):

According to a number of polls and surveys, significant majorities of Republican-aligned voters seem to believe the big lie that T____ was the rightful winner of the 2020 US presidential election and, consequently, the Biden administration is illegitimate.

Taking these data at face value, a growing chorus insists that we’re living in a “post-truth” era, where members of one political party, the Republican party, can no longer tell facts from falsehood. As a result of the Republican party becoming unmoored from reality, the narratives typically continue, America is drifting headlong into a fascist takeover or a civil war.

Fortunately for all of us, these dire predictions are almost certainly overblown. We are not living in a “post-truth” world. We are not on the brink of a civil war. The perception that we are is almost purely an artifact of people taking poll and survey data at face value despite overwhelming evidence that we probably shouldn’t.

For instance, in the wake of the 2016 election, T____ claimed to have had higher turnout at his inauguration than Barack Obama did. Subsequent polls and surveys presented people with pictures of Obama and T____’s inauguration crowds and asked which was bigger. Republicans consistently identified the visibly smaller (T____) crowd as being larger than the other. A narrative quickly emerged that T____ supporters literally couldn’t identify the correct answer; they were so brainwashed that they actually believed that the obviously smaller crowd was, in fact, larger.

Of course, a far more obvious and empirically plausible explanation is that respondents knew perfectly well what the correct answer was. However, they also had a sense of how that answer would be used in the media (“Even T____’s supporters don’t believe his nonsense!”), so they simply declined to give pollsters the response they seemed to be looking for.

As a matter of fact, respondents regularly troll researchers in polling and surveys – especially when they are asked whether or not they subscribe to absurd or fringe beliefs . . . [“well, the world is flat, isn’t it?”].

However, many academics and pundits do not seem to be in on the joke. Instead, post-2016, a consensus quickly emerged from credulous readings of polls and surveys that America is facing an epidemic of “fake news”, which was leading people to believe things that were obviously false, and to vote for unsavory political candidates. Some of the initial studies on this topic were blatantly prejudicial in their design; other widely shared studies were ultimately retracted.

As more reliable data began to emerge, it turned out that, contrary to the initial hysteria, “fake news” stories were viewed by a relatively small number of voters, and infrequently at that. Most of those served pro-T____ or anti-Clinton “fake news” by social media sites already seemed firmly committed to voting for T____, or intractably resolved against voting for Clinton (which is why the algorithms served them this niche content to begin with). That is, “fake news” is unlikely to have changed many, if any, votes. It is not a plausible explanation for the 2016 electoral outcome nor T____’s support more broadly.

Even people who share “fake news” stories typically never read (or even click on) them. That is, people are not sharing the content because they read the stories, grew convinced of their factual accuracy, and are genuinely trying to inform others. Instead, people typically share these stories based on their headlines, for a whole host of social reasons, while recognizing them to be of questionable accuracy see here, here, here, here and here for more on this).

It should not be surprising, then, that correcting misinformation seems to have virtually no effect on political preferences or voting behavior; misperceptions are generally not driving political alignments to begin with – nor are they driving political polarization.

Contrary to narratives that have grown especially ubiquitous in recent years, Americans are actually not very far apart in terms of most empirical facts. We do not live in separate realities. Instead, people begin to polarize on their public positions on factual matters only after those issues have become politicized. And even then, polarized answers on polls and surveys often fail to reflect participants’ genuine views. Indeed, when respondents are provided with incentives to answer questions accurately (instead of engaging in partisan cheerleading), the difference between Democrats and Republicans on factual matters often collapses.

In other cases, apparent disagreements about factual matters often turn out to be, at bottom, debates about how various facts are framed and interpreted, or disputes about the policies that are held to flow from the facts. That is, even in cases of genuine disagreement, there is typically less dispute about the facts themselves than about what the facts mean – morally or practically speaking.

All said, measuring misperceptions is a fraught enterprise – even when it comes to banal and politically uncontested facts. Attempting to draw inferences about “incorrect” views on matters tied political, moral and/or identity struggles is a far more complicated endeavor. These are not data that lend themselves to being taken at face value.

Similar realities hold for the data that purportedly show we’re on the brink of a new civil war.

There is strong evidence that many of the surveys and polls indicating support for, or openness towards, political violence hugely overstate actual levels of support in the American public. Likewise, data that purport to show high levels of partisan vitriol may be misleading.

In general, behaviors are often a stronger indicator than attitudinal data for understanding how sincere or committed people are to a cause or idea. The number of people who are willing to rhetorically endorse some extraordinary belief tends to be much, much higher than the subset who meaningfully behave as if that claim is true. The number of people who profess commitment to some cause tends to be much, much higher than the share who are willing to make sacrifices or life adjustments in order to advance that cause.

The big lie is no exception. Both the low levels of turnout and the relatively low levels of violence are extraordinary if we take the polls and surveys at face value.

Event organizers were expecting, “hundreds of thousands, if not millions” to take part in the January 6 uprising. This would be reasonable to expect in a world where tens of millions of Americans literally believed that an apparently high-stakes election was stolen out from under them. Even if just 1% of those who purportedly believe in the big lie had bothered to show up, the demonstrations would have been hundreds of thousands strong. Instead, they only mustered 2,500 participants . . . 

The lack of casualties was also striking, even when one considers injuries and indirect fatalities. After all, the former president also enjoyed strong support among people who are armed and formally trained in combat, such as active duty and veteran military and law enforcement. A large number of other T____ supporters participate in militias, or are private gun owners.

Yet most January 6 participants did not bring firearms, and those who were armed did not discharge their weapons – not even in the heat of the violence that broke out. . . . 

In a world where 74 million voted for T____, and more than two-thirds of these (i.e. more than 50 million people, roughly one out of every five adults in the US) actually believed that the other party had illegally seized power and now plan to use that power to harm people like themselves, the events of January 6 would likely have played out much, much differently.

Indeed, had even the 2,500 people who assembled on the Capitol arrived armed to the hilt, with a plan to seize power by force, committed to violence as “needed” to achieve their goals – things would have gone much, much differently.

Instead, most participants showed up expecting T____ would provide them with definitive evidence for his claims of electoral malfeasance, and then unveil some master plan to take the country back. This didn’t happen. . . . 

There was an even small number . . . who showed up to the Capitol with a clear intent to forcibly overturn the election – who equipped themselves for violence, researched the congressional proceedings and the layout of the building, developed and executed a plan, etc. These are behaviors consistent with a sincere belief in the big lie, and a strong commitment to doing something “about” it. . . . 

Of course, even tiny numbers of genuine extremists like these can be extremely destabilizing under the right circumstances. Had Oath Keepers breached the Capitol instead of being repelled (even as Q-Shaman, Confederate Flag Guy et al wandered the building aimlessly) … January 6 could have played out much differently.

Nonetheless, there is a huge difference in talking about identifying and disrupting small numbers of highly committed individuals willing to engage in revolutionary political violence versus tens of millions of Americans genuinely believing the election was fraudulent and being open to violence as a means of rectifying the situation. Those are very different problems. Orders of magnitude different.

The good news is that the second problem, the tens-of-millions-of-Americans problem, is not real. It is an artifact of politicized polling design and survey responses, followed by overly credulous interpretations of those results by academics and pundits who are committed to a narrative that half the electorate is evil, ignorant, stupid, deranged and otherwise dangerous [well, they did vote for a person who has no redeeming qualities and are poised to do it again — that doesn’t imply being smart or well-informed].

In fact, rather than January 6 serving as a prelude to a civil war, the US saw lower levels of death from political violence in 2021 than in any other year since the turn of the century. . . . This is not an outcome that seems consistent with large and growing shares of the population supposedly leaning towards settling the culture wars with bullets instead of ballots. This turn of events does not seem consistent with the notion that tens of millions of Americans – including large numbers of military, law enforcement and militia members – literally believe the presidency was stolen, elections can no longer be trusted, and the fate of the country is on the line. . . . 

In truth, most Republican voters likely don’t believe in the big lie. But many would nonetheless profess to believe it in polls and surveys – just as they’d support politicians who make similar professions (according to one estimate, Republican candidates who embrace the big lie enjoy a 6 percentage point electoral boost as compared to Republicans who publicly affirm the 2020 electoral results).

Within contemporary rightwing circles, a rhetorical embrace of the big lie is perceived as an act of defiance against prevailing elites. It is recognized as a surefire means to “trigger” people on the other team. A demonstrated willingness to endure blowback (from Democrats, media, academics, social media companies, et al.) for publicly striking this “defiant” position is interpreted as evidence of solidarity with, and commitment to, “the people” instead of special interests; it’s taken as a sign that one is not beholden to “the Establishment” and its rules. That is, the big lie seems to be more about social posturing than making sincere truth claims.

For many reasons, this situation is also far from ideal. But it’s a very different (and much smaller) problem than partisans actually inhabiting different epistemic worlds and lurching towards a civil war. Glass half full.

Unquote.

This view is consistent with one of mine: if millions and millions truly believed abortion is murder, there would be a lot more resistance, including armed resistance, to abortion (despite the Supreme Court’s Roe v. Wade decision). Talk is often cheap.