“I’ll believe it when I see it,” is a hackneyed idiom, but it sticks because it happens to be mostly true.
It’s one of the reasons why data visualization is emerging as a powerful information force in the crowded, intensely contested world of social media. It’s why, during last year’s State of the Union, the president live-tweeted charts to illustrate the points he was making in his speech. It’s why Ezra Klein publishes stories that deign to suggest they can explain something as vast as the Roman empire with a few dozen maps.
When we want to be persuasive, we are more effective if we buffer our case with visuals like charts and graphs. For example, I can tell you that the United States incarcerates five times as many people as most nations despite having similar crime rates. I can cite sources and link to essays and research. Or, I can show you this:
The immediate, visceral reaction we have to charts like this one, which was created by the Prison Policy Initiative, is no accident. Visual perception research has established the fact that visual information is powerfully and inescapably persuasive in a way that text and speech aren’t.
The user’s attitude matters. Research from Ansul Pandey and colleagues at New York University indicates that the persuasive power of dataviz may not be perfectly universal. The success of a visualization seems to be dependent on the initial attitude of the person assessing it. When participants in their study didn’t have strong opinions about the message being conveyed, visuals persuaded effectively. But they were less effective when participants held strong opinions in opposition to the message being conveyed.
This makes sense. It takes more to convince someone who doesn’t believe you than someone who simply doesn’t know or doesn’t care. But there’s more. Those with stronger opposing views were more likely to be swayed when a disagreeable message was presented in the form of a table. Even now, good old-fashioned rows and columns sometimes outperform charts.
Charts may even backfire. An unrelated study led by Brendan Nyhan at Dartmouth focused on political beliefs and found that charts could correct some factual misperceptions. However, people who were most likely to hold misperceptions tended to reject factual corrections no matter what. The researchers speculate that that’s because the facts presented (about hot-button topics like climate change) felt threatening to their world-view and their identity. In other words, a chart may cause someone with a polar opposite view to dig in their heels because it’s too psychologically painful to confront compelling evidence that conflicts with their beliefs.
Cognitive state and trait matter. Lane Harrison from Tufts University’s Visual Analytics Lab (home of some of the most interesting new research on visualization) has shown that the persuasive power of a visual–and even our ability to understand one–isn’t a fixed thing. The effectiveness of visualization changes with mood, gender, personality type and other factors. What’s more, priming a subject can affect how they perform when reading a graphic and what they take away from it.
Execution matters. It also appears that the ability for people to get information from a chart affects their judgment of the data itself. That is, credibility suffers if the visual is difficult to make sense of. The most anti-persuasive thing you can do is to make a bad chart that frustrates people.
What are the ways charts cause us to look away? Complexity is one. If there’s too much information and no clear, salient point that we can intuit, we tend to shut down, unless we have the time, space, and inclination to work hard at finding the narrative in the visual. Certainly in a presentation setting, that’s unlikely.
Another way charts go off the rails is if there are too many salient points fighting for our attention. This is often the result of poor design. When trying to make a persuasive chart, keep in mind the design aphorism, “if everything is bold, nothing is.”
Because the tools to create visualizations have become relatively cheap and easy to use, visual communication is exploding. One of the assumptions behind this new era of dataviz is that charts and graphs provide a necessary, compelling, and better way to persuade people. Indeed, in highly charged debates littered with polarized attitudes, our impulse is to just throw more and more charts at each other. “X Charts that Explain Y Topic” articles are legion, and demonstrate just how deeply ingrained the idea has become that if we make it a chart, people will be persuaded it’s true.