The secret life of votersA new study shows that Americans agree on more than they’re willing to admit

Abartender at a high-end restaurant in south Brooklyn, Alessia Gonzalez doesn’t feel free to speak her mind at work, at least on political issues. The neighborhood, Bay Ridge, is a Republican stronghold, and customers are often decked out in MAGA hats.

Gonzalez, 27, says that if she were to discuss her political views, particularly her belief that Israel is committing genocide against Palestinians in Gaza, she’d be judged by coworkers and customers or face even worse consequences. “I feel like I could be fired,” she says.

When politics comes up at work, Gonzalez stays mum. “I feel the need to sort of put on a mask and not necessarily agree with them or show them any kind of support, but at the same time not express my own opinion so that I don’t make anyone uncomfortable or make anyone look at me differently,” she says.

Fernando, a 30-year-old from Colombia who works in advertising in New York, is far more supportive of Israel, and he too withholds his views of the conflict around his coworkers. He worries that one colleague, who is outspoken in her opposition to Israel’s treatment of Palestinians, would have more questions than he’s willing or prepared to answer. “I don’t say anything about it, because my thoughts are sometimes very opposite,” says Fernando, who asked for partial anonymity to protect his relationships.

Withholding one’s real views, or what public-opinion researchers call “self-silencing,” appears to be widespread in the US. As part of a new study, the think tank Populace and the research firm YouGov conducted a survey in the US in which 58% of respondents said they thought most people don’t feel comfortable expressing their honest opinions about sensitive issues and 61% said that in the past year they themselves had “avoided saying things that they believe because others might find them offensive.” The study, “Social Pressure Index: Private Opinion in America,” of more than 19,000 respondents found evidence that a majority of all demographic groups — age, gender, race, income, and political party — had self-silenced in the past year.

“Social norms have made it costly to express views that one considers to be controversial, and therefore people withhold those views,” says James Gibson, a professor of government at Washington University in St. Louis who’s studied public opinion for more than 40 years and was not involved in Populace’s study.

Not everyone agrees that all private views, regardless of how objectionable they are, should be voiced in the public square or even at home. The old saying “no politics at the dinner table” seems to have existed since there have been politics to talk about at dinner tables. Pushing certain views that violate basic social norms out of the discourse can be a powerful form of accountability and a way to shape public debate for the better. But critics of self-censorship argue that when too many people aren’t saying what they actually think, it inhibits the vigorous public debate necessary for a healthy society. “If people are self-censoring, deliberation is compromised and that’s very, very damaging, in my view, to democracy,” Gibson says. The study found that the more people self-silence, the less they trust other people.

In the final sprint of another bilious and barbed US presidential election season, where the final polls show Vice President Kamala Harris and former President Donald Trump in a deadlock, it’s clear the country is deeply divided. Public-opinion experts say polarization perpetuates self-censorship by creating an environment where the perceived costs of dissent are higher. “Your brain assumes the loudest voices repeated the most are the majority,” said Todd Rose, the CEO of Populace and a former professor at Harvard’s Graduate School of Education. People fear repercussions for expressing opinions that differ from what they perceive to be the majority viewpoint. In turn, dishonest public debate makes us believe that we’re more divided than we really are.

“Self-silencing can lead to a lot more than just saying nothing,” Rose said. “It can end up leading to this place where public consensus looks very different than private consensus.”

But the more we know about what we actually believe, the likelier we may be to meet in the middle.

America seems to be suffering from what the German political scientist Elisabeth Noelle-Neumann coined in the 1970s as a “spiral of silence.” Noelle-Neumann argued that people are reluctant to share their opinions on controversial topics when they subconsciously perceive that the majority disagrees with them. As a result, majority views come to be seen as dominant over time, shrinking room for conversation and compromise.

In reality, we very well might agree on a lot more than we admit to today.

One thing liberals and conservatives might secretly agree on: that they don’t trust the government, at all.

To try to measure people’s private views, the Populace researchers used what’s called the list-experiment method. This involved dividing respondents into two groups. One was presented a series of lists each containing three statements (e.g., “Class size significantly affects student achievement”). The other was presented with a series of identical lists but with a fourth statement added — these extra statements were the ones that the researchers were actually interested in testing (e.g., “I support school choice in public education”). For each list, respondents were asked how many of the statements they agreed with, an indirect line of questioning meant to encourage honest responses. The researchers could then compare the two groups’ answers to estimate how many people actually agreed with the statements being tested. Finally, the researchers conducted a separate, “public” survey in which a different set of people were asked directly whether they agreed or disagreed with the statements being tested.

Take abortion, as one example. Populace found that Protestant, Roman Catholic, Jewish, and Muslim respondents in the list experiment were all more likely than comparable respondents in the public survey to support the idea that “abortion should be legal in most cases.” While just 39% of Protestants in the traditional survey agreed with that statement, Populace estimated that 54% agreed in the private structure. Among Muslims, the difference was 43% agreeing with the statement versus an estimated 66% who privately agreed. Overall, Populace found that 55% of respondents publicly said abortion should be legal in most cases, compared with 63% thought to have privately agreed. (Both figures were far higher than those in polling from Pew and Gallup, which included question wording with a wider range of possible responses than the agree-disagree format of the Populace study.)

Democrats and Republicans were also more privately aligned on the question of defunding the police. While 27% of Democratic respondents publicly agreed with the statement “I support defunding the police,” just 3% privately supported the movement. At the same time, just 1% of Republicans privately supported it and 6% publicly supported it.

The survey also found that Republicans and wealthier people were more likely to privately feel that society isn’t fair than they publicly acknowledged. Rose posited that Democrats tend to be more publicly concerned with unfairness, particularly when it comes to marginalized groups. Republicans define themselves in opposition to that. That dynamic is on display in this election as, for example, former Trump attacks Harris for defending a federal law that permits government-funded gender-affirming healthcare for prisoners and migrants. But privately, they acknowledged that they agree. While 50% of Republicans publicly said “we live in a mostly fair society,” just 11% privately agreed.

Another thing liberals and conservatives might secretly agree on: that they don’t trust the government, at all. Democrats were far more likely to differ in their private and public responses on this: While 36% publicly agreed with the statement “In general, I trust the government to tell me the truth,” just 5% were estimated to have privately agreed. Similarly, 2% of Republicans were thought to privately trust the government to tell the truth, while 14% said so publicly. Other surveys have similarly found historically low levels of trust in major institutions like the government and the media across the political spectrum.

Rose is particularly concerned about low levels of trust — of the government, the media, and other people. A lack of social trust causes people to resent each other, look for scapegoats, and become susceptible to demagoguery, he said. “By almost any measure, social trust is a phenomenal predictor of the health and vibrancy of democracies,” Rose says.

Not everyone is convinced that American society is suffering from a complete breakdown in societal trust. Gibson argues the long-established survey measure of whether we trust each other — agreement with the statement that “in general, most people can be trusted” — is too vague to be accurate. “You’ve got to build in some context to let people know what the question is. You can’t allow their imagination just to go crazy.”

Overall, certain groups — including Gen Zers, political independents, and college graduates — were more likely to produce dramatically different results in the public and private surveys, Populace found. Gibson speculated that this might be because these groups tend to be more aware of social norms, and sensitive to the costs of violating them.

Gen Zers, 72% of whom said they had self-silenced in the past year, might be influenced by their heavy use of social media, which “makes it very easy to manufacture false consensus,” Rose argues. They also have a particularly keen awareness of being harassed or canceled for voicing unpopular opinions. For wealthier and more-educated people, the costs of offending customers, an audience, a fellow country-club member, or a business partner might be higher. “In a society where reputation matters a lot, in an attention economy, there’s actually a lot more incentive to hold views that are consistent with what you think people want to hear,” Rose says. The risks of social ostracism might not be as acute for lower-income people and those with less education, as they’re often already socially and economically excluded. And ironically, the Silent Generation (born from 1928 to 1945) reported self-silencing the least.

Engaging in public debate or even just a dinner-table discussion can be contentious and uncomfortable. But Rose and Gibson argue that the only way to build a less polarized and self-censoring society is through hard conversations.

“As we start to find the moral courage to be honest with each other, respectfully,” Rose says, “I think you’re going to see the revealing of a common ground that we can start to anchor in and start to try to solve some of the real problems we face as a country.”

That revelation, however, will have to wait until after Election Day.