Nixon, Kennedy, and the victory of style over substance

TL;DR: A meme with very little supporting evidence

[Note: A version of this was originally published online on 2025/01/26]

Kennedy was considered the victor by those who watched on TV, while Nixon was thought to be the winner by radio listeners, driving home the importance of visual media in shaping our perceptions.

Even as a non-American, I’ve encountered this claim countless times about the first Kennedy–Nixon presidential debate of 1960. This time, the quote comes from Dress Code: Unlocking Fashion from the New Look to Millennial Pink by Véronique Hyland, a collection of essays on contemporary fashion—including the role of clothing in politics. And Hyland doesn’t even push the claim as far as others have. Just to give one example: Max Frankel, then executive editor of the New York Times, wrote in 1994 that “Nixon lost a TV debate, and the presidency, to John F. Kennedy in 1960 because of a sweaty upper lip.”

The meme goes like this: The first Kennedy–Nixon debate—also the first U.S. presidential debate ever televised—proved that style had overtaken substance in politics. TV viewers thought Kennedy won; radio listeners thought Nixon won. The implication? Looks now mattered more than words. And this, supposedly, helped tip the election.

But reading Hyland’s book, I couldn’t help thinking: Selection effects are the most powerful force in the universe. Did they actually control for anything when comparing radio and TV impressions?

Little did I know, that was the least of the problems.

Did Nixon even win on the radio?

In short: maybe.

According to a 2017 paper by Bruschke and Divine, while many contemporaneous sources tried to assess who won the debate overall (probably Kennedy), only three claimed Nixon won among radio listeners.

Of these three, two were anecdotal: one set of person-on-the-street interviews conducted in Atlanta, and one reporter’s impressions of attendees at the Southern Governors’ Conference. The third was a survey conducted by Sindlinger and Company—the only source that even vaguely resembles hard data. But even this account wasn’t published by the company itself; it appeared anonymously in the trade journal Broadcasting, and wasn’t included in Sindlinger’s official election reports.

According to that account, about 63% of radio listeners were willing to declare a winner, and 68% of those picked Nixon. Meanwhile, 46% of television viewers picked a winner, and 60% of those said Kennedy.

Even setting aside the large fraction of people who called the debate a draw, pinning the whole “radio listeners thought Nixon won” narrative on this one survey is shaky. The total sample size was 2,138 people—but only 282 of them were radio listeners. That’s not a lot of people to hang a sweeping and persistent theory of political optics on. It’s also below the minimum normally required for a robust national sample in the US.

More importantly, this dataset is missing key covariates. It doesn’t include partisanship, voting intention, or even voting eligibility: the survey may have included respondents as young as 12. It’s also plausible that radio listeners skewed more Republican than average—for instance, they may have been more likely to live in rural areas without TV access, or on the West Coast where the debate aired during commuting hours. But because the survey didn’t collect that data, we don’t know.

What is well-supported—and extremely plausible—is that partisanship strongly shaped debate impressions. A separate Sindlinger and Company survey found that just 4.4% of people who wanted Nixon to win believed Kennedy had bested him, and the same percentage of Kennedy supporters believed Nixon had won. In other words, debate impressions largely reflected prior loyalties—including perhaps at that Southern Governors' Conference, where we know 10 of the 11 governors present weren't Kennedy supporters.

And by the time of the first debate, most voters had already chosen sides. A Gallup poll just before the event found 46% support for Kennedy, 47% for Nixon, and only 7% undecided. After the debate, those numbers barely budged. Since Democrats outnumbered Republicans nationally (roughly five to three), it’s likely that even many of the “undecided” were leaning Kennedy. In a 1987 paper, Vancil and Pendell note that for Nixon to have won radio listeners by the margin Sindlinger reported, a huge number of Democrats would have had to cross party lines—an unlikely scenario given the 4.4% figure.

And the problems don’t end there.

The difficulty of untangling style and substance

Even if we grant the premise of the meme—that Nixon won on radio and Kennedy on television—interpreting its meaning isn’t straightforward. To conclude that the debate marked a turning point where style triumphed over substance, we’d need several more assumptions. For example: that Nixon’s radio success reflected stronger substance; that Kennedy’s TV success reflected superior style; that the first debate mattered more than the next three; and that the television audience’s perception directly cost Nixon the presidency.

Many of those assumptions are shaky.

Let’s start with the idea that Nixon had more substance. That’s hard to support. Several opinion polls at the time found Kennedy ahead not only on broad questions like “Who won the issues?” but also on specifics that mattered to voters, such as the perception that American prestige was declining. These were retrospective self-reports, but they suggest that Kennedy’s edge wasn’t attributed solely to style¹.

There’s also evidence that style and substance aren’t necessarily in conflict. In a 2005 study, college students were randomized to either listen to or watch the first Nixon-Kennedy debate. Viewers were significantly more likely to say Kennedy had won—but they also performed significantly better on factual questions about the debate’s content². In other words, watching may have made Kennedy seem more effective and helped viewers retain substance.

Debate performance itself often blurs the line between style and substance. The first debate focused on domestic issues, a Democrat strength. That seems like substance—but Nixon later wrote that agreeing to this topic was a tactical error. He also chose a strategy centered on defensively refuting Kennedy’s points rather than making his own case, a choice he and his staff later regretted. Nixon may have been rationalizing in hindsight, but it’s plausible that a more assertive approach would have helped. Yet is persuasive strategy style, or substance? Public speaking is, after all, part of a president’s job.

Now consider the flip side: was Kennedy better on style? This assumption is well-supported. Nixon was exhausted and recovering from a knee infection during the first debate; he still ran a fever and had already made a campaign appearance that day. Kennedy, by contrast, spent the entire day preparing with help from advisors, and looked calm, composed, and energetic. Nixon, who trained alone, refused makeup (a decision he later called “unwise”) and visibly paid for it. Even his supporters noted that he looked sick and gaunt on camera.

But style isn’t static. Nixon improved in later debates. While the second debate was more evenly matched, polling suggests he solidly won the third. That debate, notably, was on foreign policy—a Republican strength—and Nixon was no longer ill. The first debate was the most watched (about 60% of the adult population tuned in), so one could argue it had the most influence—but polls before the final debate still found the election too close to call.

Contemporaneous accounts do note Kennedy appearing more tan and vigorous than Nixon, especially in the first debate. But they don’t focus on Kennedy being more attractive per se; they focus on Nixon looking ill. That distinction again muddies the line between style and substance. As Bruschke and Divine point out—and as the 2024 election has confirmed—a presidential candidate’s health is a matter of substance, not just style.

At this point, I hope I’ve convinced you there’s little evidence for the tidy narrative that Nixon won on substance while Kennedy won on style. So if the evidence is this weak—why has the meme persisted?

The making of a meme

So it's not true—at least not according to any evidence we have—that Nixon made persuasive arguments that audiences on TV just weren’t receptive to because of his shabby looks. But could this be a figurative truth, if not a literal one?

Or maybe just a partial truth. Nixon probably did do worse than Kennedy in that first 1960 debate, including because he looked frazzled. The first debate probably was more impactful than the others—more viewed, sure, but also the first televised presidential debate in US history.

On top of that, it's plausible that style, or at least visual style, became more important in the TV era of politics. As an example, in the ten days leading up to the election, Nixon's campaign had a surge in popularity. This surge seems driven by TV, demonstrating its importance as a medium: for twelve days before the election, the Nixon campaign released a televised event each day.

But I don’t think that’s the strongest evidence. Early on when I looked into this, I encountered papers that referred to this debate as the “first televised US presidential debate,” or words to that effect. Then I tried to track down older debates from before 1960 to compare, and kept failing—because there weren’t any. In spite of the ubiquity of radio in the decades leading up to 1960, presidential debates weren’t held. There wasn’t thought to be demand before TV.

The first Nixon-Kennedy debate wasn’t just the first televised presidential debate. It was the first presidential debate. No wonder people were excited.

What it wasn’t, however, was the first time Nixon and Kennedy debated each other. In 1947 as congressmen, the two were invited to debate a bill on labor reform at the McKeesport Junto in Pennsylvania, a small event that wasn't televised, and was only barely remarked upon. The main account we have from the time was published in the McKeesport Daily News, and it didn't linger on the polticians' demeanor, even though both the audience, and presumably the journalist, saw the two in person. Neither Nixon nor Kennedy commented on their own or the other's appearance, either—both agreed that Nixon probably won, though Nixon thought this was because the audience was on his side to begin with.

So far, so good. But when the debate got rediscovered due to Nixon and Kennedy's later import as national political figures, descriptions of the debate became more and more focused on appearance—mentioning Nixon's five o’clock shadow, visible from the only photograph of him taken that night, or contrasting Nixon's “grit” with Kennedy's “charm”, or dwelling on Kennedy's youth.

By this time, of course, people had television. They also had a readily available popular narrative about how important appearances were in politics. By drawing on that narrative, they made it true—anachronistically centering appearances in a political event where that wasn’t the focus.

While the narrative that Nixon won on substance and Kennedy on style began to form immediately after that first presidential debate, it likewise crystalized later—following Kennedy’s assassination and the Watergate scandal.

In conclusion

I hope I've managed to convince you that Nixon lost on more than mere style in 1960—or at least that there's not a lot of good evidence that medium of exposure to debates had an effect on perceived winners.

But I think this meme got picked up for a reason, as evidenced by how often it gets repeated. After looking at the evidence, it's striking what a big part of the narrative is the narrative itself—how hard it is to untangle whether the first thing that changed was the politics, or whether it was what everyone knew that everyone knew about the politics. It seems that style mattering over substance, while pronounced cynically, is partly manifested by people believing it.

What this reminds me of is the COVID-19 pandemic, and more specifically the sentiment that kept getting thrown around that we were living through unprecedented times. To stick to the theme of weak evidence, I'll present an anecdote.

During lockdown, I conveyed my frustration to a friend that people seemed to think that pandemics were entirely new, rather than something humans have had to bear for millennia. He argued that it may not have been the first or deadliest pandemic, but it was the first global one, to which I argued that the Spanish flu was probably global, too. And then he argued that it was the first pandemic where such a large portion of the world's workforce were employed in service work, where they were vulnerable due to their jobs requiring so much human contact. And I granted this—but at some point, aren't we adding so many qualifiers that anything can be considered unprecedented?

The presence of television in politics was unprecedented in 1960, and it probably did change the game. People had a very strong sense that appearance was now more important, and there's evidence that television does change people's political views. But it's also very psychologically satisfying to know that your era is on the forefront of the tide of history, that it matters. And symbolism is psychologically satisfying, as is irony.

The increasing importance of style over substance, if not its decided victory, did happen. Just not right away—not in the very first televised debate. As poorly as Nixon did, many partisans were on his side regardless, and he wasn't predicted to lose the election on that basis.

No, the real trouble happened later: once everyone knew that everyone knew that appearance mattered, it mattered very much indeed, not just by deciding who won debates, but by deciding which candidates got to debate at all; who the masses got to be partisans of. By the time Nixon and Kennedy faced off in 1960, that was a settled matter.

And how could the shift not get pinned on the very first debate, a symbolically perfect linchpin? They were living in unprecedented times; it sure felt like a historic moment.

How, then, could it fail to matter much?

1. Or, I suppose, that people are extremely bad at untangling the two. But this isn’t the simplest explanation.

2. “Significance” in this sentence denotes statistical significance.