Prebunking vs debunking: what the evidence says

There are two broad ways to fight misinformation with communication. Debunking corrects a false claim after people have seen it. Prebunking warns people in advance, so they are less likely to fall for false claims in the first place. Both work, but they work differently and suit different situations.

Debunking: correcting after the fact

Debunking is the familiar fact-check. A claim spreads, someone shows it is false, and the correction is published.

What the evidence says:

• Corrections generally do reduce belief in false claims. The once-feared “backfire effect”, where corrections strengthen false beliefs, turned out to be rare.
• But misinformation keeps influencing reasoning even after people accept a correction. This is the continued influence effect. It is strongest when the correction leaves a gap in people’s explanation of events.
• Corrections reach fewer people, later, than the original claim.

How to debunk well, following the Debunking Handbook 2020:

1. Fact: lead with the truth in simple, memorable terms.
2. Warning: signal that a myth follows, and mention it only once.
3. Fallacy: explain why it is wrong, and what technique makes it misleading.
4. Fact: repeat the truth, ideally with an alternative explanation that fills the gap.

Prebunking: building resistance in advance

Prebunking is based on inoculation theory, developed by the social psychologist William McGuire in the 1960s. A vaccine exposes the body to a weakened virus. In the same way, prebunking exposes people to a weakened form of manipulation, together with an explanation of how it works.

There are two main types:

• Issue-based prebunking warns about a specific false claim that is likely to appear, for example before an election or a vaccine rollout.
• Technique-based prebunking teaches people to recognise manipulation techniques: emotional language, false dichotomies, scapegoating, fake experts, conspiratorial reasoning. These techniques recur across topics, so one lesson protects against many future claims.

What the evidence says:

• Online games such as Bad News, where players take the role of a fake-news producer, improved players’ ability to spot manipulative content in several studies.
• In a 2022 study, Jon Roozenbeek, Sander van der Linden and colleagues tested 90-second videos explaining manipulation techniques. In the lab, the videos improved people’s ability to identify manipulation. In a field test, the videos ran as YouTube ads, and viewers then recognised the techniques better than a control group.
• Prebunking campaigns have since been run at scale in Europe, notably by Google’s Jigsaw unit ahead of elections.
• The effects fade over time, so “booster” reminders help.

Which should you use?

Situation	Better choice	Why
A specific false claim is already spreading widely	Debunking	People need a direct answer to what they have seen
You expect predictable false claims (election, health campaign, crisis)	Issue-based prebunking	Getting there first removes the novelty that drives sharing
You want long-term resilience across many topics	Technique-based prebunking	One lesson transfers to many claims
A fringe claim with little reach	Often neither	Responding may spread it further
An audience that distrusts your institution	Either, via trusted messengers	Who speaks matters as much as what is said

The two approaches complement each other. A good strategy uses prebunking to raise general resilience and keeps debunking capacity ready for claims that break through.

Limits and criticisms

• Discernment, not just scepticism. An intervention that makes people distrust everything is not a success. Good studies measure whether people get better at telling reliable from unreliable content, not just whether they reject more.
• Effect sizes are modest. Neither approach transforms belief on its own. They are tools in a larger system that includes platform design, journalism and education.
• Political sensitivity. Prebunking can be seen as telling people what to think. Technique-based approaches reduce this risk, because they explain manipulation methods rather than endorsing particular conclusions.

The bottom line

Debunk when a false claim is already causing harm, and do it with a clear fact-first structure. Prebunk when you can predict what is coming, and especially when you can teach the techniques behind manipulation. Measure whether people become better judges of information, not merely more doubtful of it.