Measuring Real-World Understanding of Patterns in Data Graphics

Abstract

Presenting data visually is a cornerstone of effective science communication. While prior studies have investigated humans' ability to effectively perceive values in charts, fewer have focused on the translation of perceived values to real-world conclusions. Those that do focus on real-world understanding often utilize convenience samples or focus on very simple graphic formats, resulting in an incomplete understanding of how viewers translate data graphics into meaningful conclusions. We utilize a probability-based sample of over 3,000 participants in the U.S. to test user understanding of three chart types and find that both educational attainment and age play a role in ability to interpret data graphics. Our work demonstrates a need for further study on how chart comprehension and comfort with drawing real-world conclusions differs across demographic groups and commonly-used chart types. Additionally, this work highlights that complex charts can be inaccessible to viewers who lack confidence in reading a chart.