The Campaign That Performed

According to PsyPost’s October 5 account of new research, more than 80 percent of the TikTok videos posted by Donald Trump and Kamala Harris during the study’s 2024 campaign window featured campaign stops, motorcades or rallies. Both campaigns understood the usefulness of showing momentum. An itinerary can do a surprising amount of ideological work.

The study offers a way to examine that performance. It also exposes a temptation for anyone reporting on it: to make the numbers say more than they know.

What the counter can see

PsyPost reports that Jenny Bronstein and Tal Eitan, researchers at Bar-Ilan University, examined 276 videos from the candidates’ official accounts: 57 from Trump and 219 from Harris. The period ran from each account’s launch in summer 2024 through the day after the election. Researchers coded video formats and themes, then compared them with public counts of views, likes, comments and shares.

In that account, Trump’s videos averaged 18 seconds and Harris’s 37 seconds. Trump’s shorter clips were associated with higher average engagement per post. His humor-focused posts were also linked to significantly higher engagement across the measured metrics, though fewer than one in ten videos from either candidate primarily used jokes, memes or parody.

That is evidence about circulation within a particular collection of posts. It is not a controlled test of shortening a speech, adding a joke or changing a voter’s mind. The candidates differed in more than editing style.

Nor did every familiar campaign tactic produce a measurable advantage. PsyPost says the researchers found no statistically significant engagement differences for videos featuring informal, relatable moments or attacks on opponents. The tidy lesson—be casual, be nasty, win the internet—does not survive even this summary intact.

The primary paper, published in the Journal of Political Marketing, was inaccessible during this article’s source check; the findings here are attributed to PsyPost’s accessible report. That report also states the central limit: engagement measures do not establish how viewers felt or whether their votes changed. Two candidates, one election and one platform cannot carry a general theory of political persuasion.

The audience inside the performance

The cultural question is still worth asking. What happens when the most readily available account of a political message is a row of interaction totals?

A share can carry admiration, ridicule or a warning to a friend. Those are different relationships to a candidate. A dashboard can place them beside one another as evidence that the post traveled. The trouble begins when we supply the missing meaning ourselves, usually in the direction we already preferred.

There is no need to despise entertainment to notice this. A joke can expose a contradiction. A short explanation can respect someone’s time. Pleasure is part of public life, not a civic misdemeanor. A demand that politics become uniformly solemn would mostly reward people practiced at sounding important.

But a visible reaction is a thin form of feedback. Coherenceism asks whether an arrangement includes the responses of those it affects, especially the least-heard. Applied here, that means asking what the public can make a campaign answer—not simply how often a campaign can make the public react.

The study does not prove that TikTok’s algorithm caused the observed differences. It gives us a narrower opening: to inspect the formats through which candidates present themselves, and refuse to promote engagement into understanding without evidence.

The next time the crowd fills the screen, let the image register. Then ask what you have learned about the proposed government. The applause has answered a different question.