We need more research on how different audiences engage with data stories through visuals, but we also need to be mindful of how we generalize about our learning.
I always find research on data visualization (particularly on what works for different audiences) interesting, but also get frustrated when articles lead with sweeping conclusions that may be the result of a flawed study design (and a tiny sample of 49 people).
Internews recently conducted a study on data and visual literacy in Kenya. The big headline conclusion:
"Bubble charts have little to no traction with Kenyan audiences of print news. Bar charts, on the other hand, are generally considered more credible and “scientific.” But if you want interest and inferences drawn from the news you’re conveying, then the best ones to use are pictorial infographics."
Or so says the leading paragraph on their Medium article. It's important to note that the study tested understanding of the three different kinds of graphics (bubbles, bars, and pictoral infographics), but each graphic was focused on a different data story and quality of the design varied. The sample size of 49 is unlikely to be representative of the entire population of Kenya either.
The pictoral infographic that scored best for recall of information was also focused on a data story about Kenya, where the other two examples (the bar charts and the bubble) were not as locally relatable, telling global data stories. A colleague also called out the potential for issues of cultural insensitivity using crosses on the gravestones, given the large Muslim population in Kenya.
The bar chart, while at East African context, is also not terribly well designed as far as bar charts go, so it seems an unfair comparison to the pictoral infographic.
Despite these limitations, the study results are framed as recommendations for how we think about communicating data in print media across an entire country. While I love and find great value in well-designed infographics for the right audiences, I worry that some will read the study post and conclude that infographics are the only and best way to present information, rather than encouraging budding data viz designers to be thoughtful about their audiences and their needs.
I would love to see this study replicated with the same data story represented in three different well-designed visual graphics (of different types). By using a consistent data story (or comparable data stories at least) relevant to the local audience, and testing the graphics with a larger, representative sample, there's a great opportunity to learn about what makes information in a print medium sticky in a particular country.
Does anyone know of studies (from sub-saharan Africa in particular) with that kind of data? Would love to hear your thoughts!