Want to Change Someone's Mind? Just Show Them a Random Chart

Want to Change Someone's Mind? Just Show Them a Random Chart

Charts are great storytelling tools. They're incredibly useful for visualizing or summarizing often complicated topics, from macroeconomic data to social and political trends. As a result, they're a favorite tool for journalists who want to convey important (but often boring) information in an attention-grabbing way.

The problem? Charts can be used to make a claim more believable, even when the underlying data is rocky. That's according to new research in the journal Public Understanding of Science by Cornell researchers Aner Tal and Brian Wansink. A graph's persuasiveness has nothing to do with its ease of communicating, the researchers claim. Instead, a graph signals to readers that the information has a scientific basis — even when the information is totally, utterly wrong.

The study: Wansink and Tal recruited 61 participants online to participate in a study to test their hypothesis. Participants were presented information about a new medication which allegedly enhanced immune function and reduced the occurrence of the common cold. Half the participants were shown a chart related to the medication that presented no new information related to the medication, while the other half were not.

The researchers found that when a claim about a new drug's effectiveness was presented in text form, 67% of research participants said they believed it. But when the text was accompanied by a simple graph making exactly the same claim, 97% of participants believed it.

For shits and giggles, here's that breakdown in chart form:

Why it matters: The researchers see a troubling problem brewing for both scientists and journalists tasked with articulating complex topics to the public. "If a claim 'looks and smells' scientific, a person may be inclined to believe," write Wansink and Tal. "In other words, communications may be made more convincing without any alteration in content, simply by virtue of being presented with elements associated with science."

This is great if you're trying to convey the facts, but it also means that the average news consumer doesn't actually spend time thinking critically about the data they're being presented. People do not care what the data actually says: If it's presented in chart form, it must be true. 

This principle is best demonstrated by Tyler Vigen's Spurious Correlations, a website that lets users chart their own poorly thought-out correlations.

Yes, charts and graphs are great, especially when deployed effectively by scientists and journalists, but they also have the potential to obfuscate and misinform. Here's a pro-tip: Always look at the data — and use those critical thinking skills first.