Soon after the attack, Twitter users including US white nationalist Richard Spence posted the photo above with captions criticizing the anonymous woman for appearing unconcerned, earning thousands of retweets. Some defended the woman, but others complained that ”[she] can’t be bothered to put down her phone,” and “Holy crap this is infuriating.

Photographer Jamie Lorriman spoke out on March 24saying that critics of the young woman were taking her image out of context and noting that officials had asked passersby to leave the scene. The woman herself was quoted on the blog Tell Mama, “I’m shocked and totally dismayed at how a picture of me is being circulated on social media,” she said.

To those individuals who have interpreted and commented on what my thoughts were in that horrific and distressful moment, I would like to say not only have I been devastated by witnessing the aftermath of a shocking and numbing terror attack, I’ve also had to deal with the shock of finding my picture plastered all over social media by those who could not look beyond my attire, who draw conclusions based on hate and xenophobia.

In other photos taken at the same moment, the woman looks clearly distressed. But the outrage sparked by this photo is not just about her expression in that split second; it’s a product of pre-existing prejudice and cognitive bias.

Confirmation bias: Believing what we already believe

In other photos of the attack, several people who are not wearing headscarves pass by with neutral facial expressions. The fact that only the hijabi woman went viral suggests that its popularity has to do with the fact that it fits anti-Muslim prejudices. In psychology, this is called confirmation bias.

Confirmation bias causes us to place more importance on evidence that echoes what we already believe. The belief that Muslim people rejoice at violence is a common myth, disseminated even by US president Donald Trump, who falsely claimed while campaigning last year that Muslim Americans celebrated the World Trade Center bombing in September 2011.  To someone who already believes that most Muslims are indifferent, or even happy about terrorist attacks against the West, Lorriman’s photo from the Westminster Bridge has extra value because it validates their beliefs.

Selection bias: Cherrypicking evidence

But let’s unpack this even further. Another trick at work here is selection bias, a problem that often arises when statisticians or scientists select sample subjects from a population, group, or data set in a non-random way, skewing their conclusions. In this case, the selection of one image has been interpreted to mean that its subject behaved with nonchalance—even if she didn’t really—because of who she is. But with a broader look, you can find plenty of other photos showing a diversity of Caucasians, Asians, men and women who also walked past victims lying on the ground, with apparent indifference. So while people who have a bias against Muslims chose this photo as their “sample” of the event, it’s hardly representative.

Consider this image that could have been framed as evidence of nonchalance among New Yorkers during 9/11:

On September 11, 2001, Magnum photographer Thomas Hoepkers captured a group of five New Yorkers who seem to be casually chatting and enjoying a sunny day in a Brooklyn park, while a cloud of smoke rose from Lower Manhattan in the background. The reason this image did not cause much reaction is because Hoepkers anticipated the misinterpretation that could arise, and did not publish the photo until four years later. In 2006, he explained: “Publishing it might distort the reality as we had felt it on that historic day.

Picture superiority effect: Worth a thousand words

Photography has a powerful ability to enforce pre-existing biases, which is why it is particularly damaging to use a photo as evidence about any group of people. Psychologists refer to the ”picture superiority effect,” which states that concepts are better understood and remembered when they are presented as images, not words. Advertisers and propagandists use this to a great effect.

The headscarf the woman is wearing is a strong visual symbol: We immediately know that she is Muslim because of the hijab, and the visual reminder resonates because the hijab is such a common subject of political debate. And in this politically-charged context of Islamophobia, terrorism, and fears over globalization, anti-Muslim advocates and alt-right leaders in the US managed to manipulate social media users’ own cognitive biases, simply by selecting the right photo to elevate at the worst possible time.

📬 Sign up for the Daily Brief

Our free, fast, and fun briefing on the global economy, delivered every weekday morning.