What Makes Us Care? Visualizations and Empathy

Can data visualization make us more empathetic?

Photo by youssef naddam on Unsplash

Sometimes it feels like the world is falling apart. War, natural disasters, and political conflicts all seem to pose an unrelenting threat. And in the midst of it all are individuals and families who are suffering. We feel for them, we worry about them, but at some point, we sort of… stop caring?

Psychologists and human rights experts call this compassion fatigue or compassion collapse. Learning that a percentage of people in a far away place are affected by something may not always cause an emotional response. Scientists believe this is because using abstract data or terminology to describe human rights issues creates too much distance between a viewer and the suffering.

But empathy for individuals experiencing a humanitarian crisis may lead people to raise money for relief or generate political will to find solutions. Due to privacy concerns, it’s not always possible to use an individual personal story or photographs to humanize the information about a crisis.

With this in mind, Jeremy Boy, et al., decided to develop anthropomorphic visualizations (which they refer to as anthropographics) that might achieve the goal of humanizing data to elicit empathy, while respecting the privacy of the people experiencing the crisis.

The authors designed anthropographics based on issues in Syria stemming from its civil war, with a focus on how these issues affect children.

In each experiment, the issues were presented using slideshows that followed the same structure:

Slide 1 introduced the problem using text only. Slide 2 gave a geographic overview of the situation using a map and some text. Slide 3 provided a demographic statistic from before the crisis using a chart or graphic and descriptive text. Slide 4 provided that same statistic in the current situation using an updated chart and descriptive text. Slide 5 gave a conclusion statement on how the children are suffering due to this issue using only text. Simulated slides images created by M. Suri.

Slides three and four were varied so that some participants saw anthropographics, while others saw pie charts. The participants were all recruited and tested through Amazon Mechanical Turk, so they saw these slides on their own computers.

The authors defined empathy in two ways: personal distress, which is how moved or emotional one feels, and prosocial behavior, which is how much action one takes to solve a problem. The researchers measured personal distress using a scale and free-form text boxes. They measured prosocial behavior by asking participants to make a fictional donation to an organization working on the issues.

In addition to the user experiments, the authors also developed a design space to create anthropographics to work through the design and creative options for creating this type of visualization. The authors identified four main elements that are involved in designing anthropographics: whether the icon represented unit or aggregate data, how human-like the graphic is, the information included in the labels, and the arrangement of the icons on a slide or page.

The four elements of designing anthropographics identified by the authors of the study. Images from Boy et al., 2017.

The authors decided to also test the effect of varying the accompanying text between experiments. In different trials, the authors removed all text, they made the text more boring, and removed labels that provided context about who the children were.

After running almost a dozen experiments with different combinations of anthropographics and standard charts, the authors found that using anthropographics had no significant effect on reported empathy compared to a standard chart. Overall, the participants didn’t report more or less personal distress based on which type of visualization they viewed, and they were not moved nor deterred from making a donation.

However, participants reported a lower level of empathy when the text and labels were totally removed or altered to be drier and less compelling. This led the authors to conclude that the story of a human rights issue is what moves an audience. As the authors write:

A tragic human rights narrative simply is a tragic human rights narrative.

Using anthropographics versus a standard chart is not as important in eliciting empathy as the story of the crisis. The authors note that future work could be done on whether anthropographics may play a role in attracting attention and drawing viewers into a story. The authors also suggest that adding elements like animations or movement may increase the effect of anthropographics on empathy. Notably, these visual elements have an inherent story-like quality.

This robust study may highlight a limitation of visualizations in communicating information. At some point, viewers seem to need context and a story — something human — to more deeply internalize information.

This article is inspired by:

• Boy, J., Pandey, A. V., Emerson, J., Satterthwaite, M., Nov, O., & Bertini, E. (2017). Showing people behind data: Does anthropomorphizing visualizations elicit more empathy for human rights data?. In Proceedings of the 2017 CHI Conference on Human Factors in Computing Systems (pp. 5462–5474). ACM. [PDF]. [Medium article]