A new approach to evaluating data visualizations could predict why people misread charts before they even take a test. By simulating the cognitive processes behind chart reading, researchers hope to make visual design more reliable and less prone to user error.
The Research
Researchers at the National Renewable Energy Laboratory and Northeastern University, led by Harrison Goldwyn and including Graham Johnson, Christopher Ibarra, Lace Padilla, and Kenny Gruchalla, developed a computational model that mimics how people interpret visual information. They drew on Active Inference, a framework from neuroscience that describes how the brain constantly updates its beliefs and chooses actions to reduce uncertainty. Their study, posted on arXiv in July 2026, models chart reading as a dynamic search: an agent looks at a bar chart and decides where to focus attention to estimate an average value.
They implemented two types of agents: a Fast, heuristic (Type 1) agent that relies on quick, intuitive judgments, and a Slow, analytic (Type 2) agent that is more deliberate and memory intensive. The Fast agent was prone to tick-salience bias—it paid too much attention to prominent tick marks on the axis—while the Slow agent suffered from working-memory decay, forgetting earlier values as it processed the chart. Both agents produced “cognitive traces,” sequences of beliefs and eye fixations that reveal how errors arise.
This simulation offers a mechanistic account of why certain chart designs fail. Instead of just measuring how many people get a chart wrong after the fact, this model can predict where and why errors occur, providing a testable hypothesis for future experiments.
Why It Matters
For anyone who reads charts—whether in news, business, or education—this research points to the need for designs that account for human cognitive limits. Our attention is drawn to visually dominant elements, and our working memory has a limited capacity. By understanding these biases, we can design charts that minimize misinterpretation, reducing the risk of errors in important decisions.
What You Can Do
When reading charts, be aware of potential pitfalls: visually prominent axis ticks may not represent the data’s true scale, and comparing many values can tax your memory. Practice by discussing chart interpretations with others, and consider training your working memory and attention through targeted cognitive exercises.
Source: arXiv q-bio.NC
Curious about your own brain? Take our free adaptive IQ test or try 306 brain training levels.
