Lately, Sunshine Hillygus has been hearing the same question from some of her politically active friends. They’ve been writing postcards to voters in swing states and knocking on potential voters’ doors, but they want to know if they’re channeling their energy toward the right things: What should they be doing, they ask her, if their goal is to influence the outcome of the election?

“The thing that I say,” Hillygus, a political scientist at Duke and the co-author of the book The Persuadable Voter, told me, “is that you can have the biggest impact by contacting people that you know.” Biggest, of course, is a relative term—the results of the election will not hinge on one changed vote—but, Hillygus says, people are generally more receptive to appeals from those they know and trust.

This month, I interviewed more than 20 people who had tried to convince a family member to vote for a particular presidential candidate, or to vote at all, in the 2020 election. Their tones and approaches varied, and so did their results: I heard from a woman whose grandparents met her tearful plea with cold indifference, as well as from a man whose mom ultimately caved because this year, his birthday falls on Election Day.

Every relationship has its own particular dynamics, but there are some basic principles of persuasion that apply whether a message is delivered by a family member or a television ad. One crucial factor is “the amount of careful thinking that the recipient does about the content of the message,” Jon Krosnick, a professor of communication, political science, and psychology at Stanford University, told me. When presented, for example, with five reasons to vote for a particular candidate, “what I’m going to say [to myself] is, Can I think of reasons why each of these five are wrong? And if I can’t, that’s the situation in which persuasion tends to happen,” Krosnick said. This presupposes that they have the bandwidth, background knowledge, and motivation to engage in “careful thinking,” which is necessary for a “lasting, meaningful change” to a political opinion.

The results of these all-important internal deliberations, though, can be hard to verify. I told Krosnick about a success story I’d come across, in which a 28-year-old Washington, D.C., resident named Brennan Suen had convinced his 94-year-old grandmother, a lifelong Republican voter living in Arkansas, to vote for Joe Biden. Suen, who is gay, fears for his future and that of other LGBTQ Americans during a potential second term of Donald Trump’s presidency, and while he was anxious about how his grandmother might respond to hearing that, he called her and expressed himself, at one point through tears. “What was amazing,” he told me, “was that as soon as she knew about [my concerns], she prioritized me—my safety and my emotional well-being—above whatever party loyalty that she had.”

The story moved me, and I was curious what Krosnick made of it. “I would caution you as a journalist not to say, That’s a story of a victory. That’s a story of an alleged victory,” he said. Since Suen’s grandmother knows that she can cast her vote confidentially, Krosnick noted, “it may be far more important to her to maintain harmony with her grandson by saying, ‘I hear you, you’ve convinced me,’ than it is to actually change [her] behavior.”

In Suen’s case, several variables point to a genuine success—he told me that his grandmother has been supportive of his sexuality in the past and, despite her party identification, is not “extremely political” nor “immersed in the Fox News bubble”—but it’s aggravatingly hard to know for sure whether one’s efforts to sway a vote actually worked.

Natasha Suri, a 28-year-old who works in advertising in Austin, Texas, found a clever way around this problem. After her dad committed to voting for her preferred candidate by mail, she insisted on photo evidence.

As people around the country attempt to change the minds of their family members, it is worth examining why, exactly, they feel compelled to do so.

The simplest explanation is that people prefer a certain electoral result, and they’d like to do all they can to bring it about; further, many of them can’t stand the thought of a loved one supporting a politician they fear will harm the country. Krosnick, though, interprets these conversations as an attempt to feel a sense of control in what many perceive to be a perilously high-stakes political contest. “Trying to persuade other people is one of the only pathways that people have to deal with whatever anxiety they may feel” about the election, he told me.

The thing about that pathway, though, is that on an individual level, it’s highly unlikely to make any difference politically: No one dad or grandma is going to swing the election. So the curious fact at the heart of all these attempts at persuasion is that while politics dominate the conversation, what people are really doing is indirectly renegotiating their relationships with each other; in this context, votes become less a political instrument and more a symbol of what people are willing to do for one another. “The probability that you’re going to affect the outcome of this election? Zero,” Krosnick said. “The probability that you’re going to affect the outcome of the next chapter of your life with your [family member]? Pretty high.”

Indeed, some people I interviewed who unsuccessfully tried to persuade a loved one found that their relationship had been damaged in the process. Cal Hudson, who is 25 and lives in Honey Brook, Pennsylvania, tried convincing her grandfather not to vote for Trump by calling him on her own birthday, thinking that he’d give her some time on a special occasion. “I made sure I stayed calm [and] didn’t swear,” she told me. But at the urging of her grandmother—who “runs the show” and overheard the call, Hudson said—he abruptly hung up. Her grandparents haven’t responded to her texts or calls in the three weeks since.

Despite the possibility of devastating outcomes, many people choose to have these conversations anyway. Perhaps another reason they do so is because Americans’ families, especially their extended families, provide them with an ample supply of people whose views differ from theirs.

I wondered if Americans’ extended families might even be the main source of political diversity in their lives, but this does not appear to be the case. Meredith Rolfe, a political-science professor at UMass Amherst, told me that, based on the findings of a handful of different studies, any given American’s extended family (excluding their spouse) seems to be about as similar to them politically as their close friends, and more similar to them politically than their co-workers and less-close acquaintances. Meanwhile, any given married American’s spouse tends to be more like them politically than people they know outside their household.

But a key difference between family and non-family is that “people are more likely to keep discussing politics with their ‘close ties’ (including relatives and spouses), even when they disagree,” Rolfe wrote to me in an email. This might be the more important factor in explaining how often Americans find themselves disagreeing with family members.

Many Americans seem motivated enough on their own to keep trying to change their family members’ minds, and on some level politicians sense that the power of these conversations could be harnessed somehow. In September, Representative Alexandria Ocasio-Cortez encouraged her supporters to “use your relationships” and talk with loved ones about voting. “There is someone in your life that only you can get through to … maybe you’ve got an uncle in Ohio, maybe you’ve got family in Florida,” she said. (Her encouragement was actually what led Brennan Suen to call up his grandmother.)

But the experts I interviewed weren’t familiar with any large-scale effort on the part of campaigns to train loyalists in the art of persuasion so they can go to work specifically on their loved ones. When I asked Krosnick about an operation like this, he said it was possible that political strategists might be overlooking a potentially effective but as-yet-untested strategy. He seemed more interested, though, in how an organized educational effort might improve voters’ well-being: If they’re going to initiate these sometimes rancorous conversations, they would benefit from learning how to have them more harmoniously.

For those who seek to change a loved one’s mind, the first step is to figure out who is worth trying to persuade. If someone has a strong opinion on a candidate or issue, Krosnick told me, “don’t bother, because it’s going to be extremely difficult to change that [opinion], and it’s going to be interpersonally costly.” Having the conversation could damage the relationship, with very little hope of minds being changed.

A better goal is to seek out people who aren’t particularly attached to a candidate and are open to hearing information that might help them choose. There may be fewer undecided voters this election compared with 2016, but they are out there. Hillygus, the Duke political scientist, pointed out that while this year’s presidential candidates tend to elicit strong opinions, “in down-the-ballot races, or in open-seat races, when there are new people being presented, you really see this instability [of preferences among] a subset of the electorate.”

The second ripe category is people who might lean toward your preferred candidate but don’t always vote. Experts told me that it isn’t, as a rule, easier or harder to convince someone to vote than it is to convince them to vote a certain way. If someone is strongly opposed to voting, they may not budge on that either.

Whoever you pick, and whatever you’re trying to persuade them to do, it’s vital that they’re receptive to at least having a conversation. A situation in which you plead with someone to listen to you is, in Krosnick’s view, “hopeless.”

Once you’ve settled on a target, the second step is to calibrate your expectations. Hillygus advises adopting a long-term outlook. “It’s not just one conversation” that necessarily will change someone’s mind, she said. “It is a repeated discussion of the things that you value.” She stressed that persuasion is something that can happen, gradually, over time. “Just because someone moves their grandmother from a .25 probablity of supporting Biden to a .35 probability of supporting Biden, they may have not pushed her to cast a ballot for Biden, but that doesn’t mean persuasion hasn’t happened,” she said. The payoff might come later.

Steve, a 38-year-old engineer in Sioux Falls, South Dakota, told me that he worked on his dad, a retired Iowa dairy farmer in his 70s, for years. After the 2016 election, Steve, who asked me to omit his last name to protect his privacy, had been trying to convince his dad not to vote for a third-party candidate again. “The process” of persuasion, Steve said, “was definitely a drip-drip-drip and not an avalanche.” (Or maybe not: “I probably didn’t need it pointed out a dozen times,” Steve’s dad told me.)

Hillygus’s other suggestion was to talk in person, or at least on the phone. “Sending a text message or sending an email,” she said, “is going to be less effective than a conversation, where you are better able to communicate emotion, [and] emotion is very much a part of persuasive appeals.”

Also, approaching the conversation in a measured, noncoercive way is highly important. “The minute somebody starts to threaten your feeling of freedom, independence, and self-control, our natural inclination is to try and reestablish that sense of freedom” by resisting, Krosnick said. This means that your overall message should be less “Here’s how I want you to vote” and more “Here’s some information that you might want to know.”

The experience of Peg, a retired nurse in her late 60s in Ohio, illustrates the dividends of this posture. Peg, who asked to not have her last name published in order to protect her privacy, voted for Trump in 2016 primarily because she considers the Clintons to be “very corrupt,” and then was pleasantly surprised that Trump, in her eyes, delivered on his campaign promises. For the past four years, Peg says she has been periodically talking with her sister, a Democrat who voted for the Green Party candidate Jill Stein in 2016, about the possibility of supporting Trump. Initially, she remembers her sister’s responses being something like, “I’ve always been a Democrat—stop trying to change me.” To which Peg would say something like, “I’m not trying to change you. I just want you to not say, ‘I’ve always been a Democrat, and this is the way I’ve always voted,’ and to look at what is being presented on both sides. Then, you can see if you come up with the same conclusion.”

Peg kept up those messages, and, in time, this approach seems to have made room for her sister to do the “careful thinking” that Krosnick said is essential for producing shifts in opinion. Peg told me, “Before this pandemic hit and knocked everybody down, I said to her, ‘So you and your husband, how has life been for you since Trump has been in office? How has your retirement fund done? What is your quality of life?’” Peg says that her sister talked these questions over with her husband and, a few weeks ago, informed Peg over lunch that they both were voting for Trump.

Peg’s sister’s personal finances seem to have played a role in her decision, but people vote for lots of different reasons, which introduces the question of what the substance of your message to any given family member should be. Hillygus’s first piece of advice in this regard is to find an issue on which you and the listener have common ground. “If you’re trying to convince your Republican aunt to vote for Biden, and you know that she’s pro-life and pro-gun and anti-regulation, but she’s really turned off by the way that Trump talks about immigrants, then you would focus on [that],” she said.

Finding any common ground may seem daunting in such a polarized political landscape, but, Hillygus insists, “the national conversation is typically more divided than people’s actual views.” According to her research, many people who identify with one party or another actually agree with people on the other side about elements of a bunch of issues, including health care, taxes, the minimum wage, affirmative action, abortion, climate change, and gun control.

Forrest Brungardt, a 17-year-old high-school senior in Topeka, Kansas, recently was able to find common ground on the issue of abortion with a “super conservative” relative of his. “I told her how I am pro-choice but that I think abortions should be a last resort,” Brungardt said. He made the point that “abortions are going to happen whether they are legal or not, and only one party supports widespread access to contraceptives, health care, and child care, which would make having a baby more affordable.” After hearing this and other arguments, she told him she’d be voting for Democrats up and down the ballot this November.

This is also a tactic that helped Taylor Locke, a 29-year-old in Southern California, convince his 58-year-old mom, Deena, to vote for the first time in her life. She’d never done so in the past because she was distrustful of politicians and didn’t think her vote mattered. “I explained to her how I used to be in the same boat, where I didn’t vote because I felt the system was extremely rigged,” Taylor told me. That gave him an entry point into talking with her about California’s system of ballot propositions, which give voters a say over changes to state laws—an idea that was appealing to her. “You know, on the sticker, it says, ‘I voted,’” Deena told me. “I’m going to write on there, in little marker, I finally voted.” (Her change of mind even had a cascading effect: She managed to persuade her husband, a Trump voter in 2016, to vote for Biden this year.)

Hillygus’s second recommendation is to pick a subject “on which you have credibility.” That doesn’t mean you have to have deep, research-based expertise, she said, but it helps to center your argument around something relevant to your life or interests: It could just be that “you are a hiker and you can talk about how the environment is particularly important to you.”

One reliable pathway to credibility is to invoke a personal narrative. Lee, a 17-year-old who lives in small-town Texas, wanted to convey the stakes he saw in this year’s election to his mom, who voted for Trump in 2016. “I am tube-fed and have an IV port, and I need constant access to medical care and supplies,” Lee, who asked to be identified by only his first name to protect his privacy, told me. “Without insurance there is simply no way I would be able to afford that.” The risk that he’d lose coverage at age 18, if the Affordable Care Act were repealed, is what persuaded his mom to vote for Biden, according to Lee.

Those techniques are grounded in the first person, but Krosnick explained a formula for lasting opinion change that takes the emphasis off whoever is delivering the message. You start by saying something like, “Maybe you’ll find this information helpful when you’re deciding who to vote for—I did.” Let’s say, for example, that you’re trying to persuade someone that the Trump administration has been corrupt. You could cue up a video of Trump on the campaign trail in 2016 promising to “drain the swamp,” and then say, “What’s been established during his presidency is that he hasn’t ‘drained the swamp’—he’s created more of one.” Here, you could bring up a series of articles from a few different sources that are perceived to have “contradictory biases”—say, The New York Times and Fox News—covering various convictions of or charges made against Trump affiliates. “You’re not saying, ‘President Trump is corrupt.’ You’re just saying, ‘This event occurred,’ ‘This person was convicted of this,’” Krosnick said. “The key is to have multiple pieces of information from multiple sources, all of which converge on the same conclusion, leaving room for the recipient of the message to draw the conclusion him- or herself.”

Of course, if you aren’t able to change someone’s mind, sometimes you can at least settle for changing their behavior. Natasha Suri, the Texan who extracted photo evidence of her dad’s mail-in ballot, told me that she’s been “trying to point out Trump’s failings every chance I get” since her dad voted for the president in 2016. That didn’t persuade him, though, and neither did her argument that four more years of Trump, through his climate policies and general instability, would make the world inhospitable to her and, eventually, her own children. After everything else fell through, Suri and her dad made a deal: He’d vote for Biden, and she’d get married and give him grandkids.

This did not strike Krosnick as an example of lasting opinion change. But, at any rate, “she got the good end of the bargain,” he said, “since it doesn’t sound like she committed to a specific wedding date.”

As I first heard tales of Americans whose fears appeared to change a family member’s mind, I was touched: Here was evidence that people are capable of revising their beliefs when they learn about the concerns of those they love.

But when I revisited these accounts days later, I came to see something more discouraging in the triumph of the personal. Does it really take having a gay grandson to see the importance of LGBTQ rights?

The experts I spoke with didn’t share this concern. Krosnick noted that it’s “very costly time-wise” for people to do their own in-depth research on issues they aren’t informed about, and he thinks of first-person accounts as a “lovely” way of making the stakes of something abstract “vivid” and “personal.”

Hillygus, for her part, was not surprised that this is how many minds change: She mentioned research indicating that male members of Congress who have a daughter are more likely to favor legislation that promotes reproductive rights. Under a political system that “forces a choice between Democrat and Republican” and a media environment that “incentivizes focusing on disagreement,” Hillygus said, these individual conversations can convey insights that people might not receive through their usual channels.

If Americans, as Ocasio-Cortez urges, “use their relationships” in service of their political ideals, that could, in the collective, affect electoral outcomes in some small way. But more likely is that these attempts at persuasion reorder not the country’s political terrain but its social one, driving people further apart, or—with hope—bringing them closer to a shared understanding.