In recent years, misinformation about vaccines has discouraged parents from having their children vaccinated, which puts their own children – as well as their neighbors’ children – at risk.

In response, many doctors, scientists and journalists have worked together in an attempt to correct false beliefs about vaccination. For example, earlier this month, Jimmy Kimmel invited real doctors to school his audience on the dangers of non-vaccination.

Yet research has shown that common correction strategies frequently backfire. Not only do these strategies increase the acceptance of misinformation, but they can also spread it to new audiences. While Kimmel’s entertaining approach avoided most of these pitfalls, other media outlets aren’t often that lucky.

To assess whether something is likely to be true, people often rely on how often they’ve heard an opinion. If many people believe it, the thinking goes that there’s probably some truth to it.

Unfortunately, we’re bad at tracking what we hear and where we hear it from, and instead rely on a message’s familiarity. The more we hear something, the more familiar it feels, but we quickly forget the context (for example, whether or not we were assured something was true or false).

As time passes, all that remains is the feeling that we “heard something like this before,” which we then interpret as evidence that it’s probably true.

When it comes to false information, every repetition increases its familiarity; for this reason, it’s a bad idea to repeat inaccurate information in the act of correcting it. Yet repeating falsehoods is a key element of articles like “10 Deadly Myths about Childhood Vaccines” or “6 Dangerous Anti-Vaccine Arguments Analyzed, Explained and Shutdown.”

Indeed, controlled experiments show that even serious news analyses that confront myths with facts backfire as time passes. In some studies, we showed participants a series of statements and told them that some were true (“aspirin can damage tooth enamel”) and some were false (“corn chips contain twice as much fat as potato chips”).

When we tested people immediately after, they made few errors. But when we tested them three days later, many had forgotten which statements were true and which were false. Instead, they relied on whether the statement felt familiar (“seems as if I heard something like this before”). In fact, after three days, the more times we had told participants a statement was false, the more likely they were to believe it was true. Because repetition increases familiarity, repeated warnings can actually morph into recommendations.

One (repetitive) voice a chorus

Just as social consensus can make any idea seem true, a lack of consensus can call truth into question. People consider disagreement an indicator that the truth is unknown – that either side may be right. It doesn’t matter if there’s only a vocal minority; what matters is the sheer fact that there’s any disagreement at all. A single, repetitive voice can act as a chorus.

In some experiments, hearing an argument three times from the same person was almost as effective in persuading people that there is broad support for an opinion as hearing the message from three different people.

An opinion, then, can be perceived as widely shared, even if it’s only heard, over and over, from the same person. This gives repetitive voices a big advantage, and as the vaccination discussions illustrate, a few persistent voices can shed doubt on the overwhelming consensus of 99% of medical professionals. Correction attempts should not provide an additional platform for further repetition.

Repetitions are particularly problematic when the disagreement is presented in overly polite terms. We often hear that the infamous vaccination study by Andrew Wakefield was “discredited.” But this word is often interpreted as merely meaning “there are also contradictory findings.” It fails to convey that the original report was found to be fraudulent and based on fake data, and its author was stripped of his professional license. The credibility of a source is important, and readers can’t properly evaluate a source when the media withholds factual information about fraud, or softens it to “discredited” or “controversial.”

Media can make misinformation memorable

Misinformation is often presented with anecdotes, which are easy to understand, evoke vivid pictures in our minds, and are particularly likely to stick. This gives anecdotes the power to override hard facts, like statistics and scientific evidence. A vivid example about a single case of a vaccination’s apparent side-effects – complete with quotes from a worried parent – can then make it difficult for readers to reconcile the anecdote with statistical evidence intended to combat it. Worse, as time passes, the anecdote will remain memorable, while the statistics will be forgotten.

Misinformation is also more likely to stick when it’s accompanied by photographs. Even when the pictures have no evidence-based value and are merely intended to draw attention to an article, they can influence people’s beliefs. For example, in one study participants were more likely to believe a false claim (such as “Magnesium is the liquid metal inside a thermometer”) when it appeared beside a photo of a thermometer.

Photos make claims feel true – not because they provide factual evidence, but because they make it easier for readers to imagine the claim. In a recent CBS report about vaccines, a photo of a boy behind a glass window may make it easier to imagine unfounded claims about autism.

In short, media reports that intend to correct false information can have the unfortunate effect of increasing its acceptance. Using anecdotes and images makes false information easier to imagine – and by highlighting disagreement, they distort the amount of consensus that actually exists.

It’s better to be a repetitive voice that highlights the facts, rather than one that draws attention to falsehoods. Ultimately, readers, viewers and listeners are better served by receiving correct information about the benefits of vaccination and the risks of non-vaccination in an easy-to-process format that makes accurate information feel familiar, credible and true.

This article was originally published on The Conversation. Read the original article.