Do Visualization Literacy Assessment Tests Account for High-Level Visualization Comprehension?

Authors

Tapendra Pandey (University of Oklahoma), Aaryani Chowdary Ambati (University of Oklahoma), Arran Zeyu Wang (University of North Carolina-Chapel Hill), Ghulam Jilani Quadri (University of Oklahoma)

Abstract

Visualization literacy assessment tests such as VLAT and Mini- VLAT measure people’s ability to perform isolated, low-level visual analytics tasks, such as estimating statistical quantities. However, it remains unclear whether these assessments also account for the overall understanding viewers intuitively gain about the data with- out explicit cuing or guidance, often referred to as high-level com- prehension. In this work, we examine whether Mini-VLAT scores predict high-level comprehension, characterized as alignment be- tween viewers’ uncued interpretations and designer intent. We con- ducted a study with 80 participants, in which each participant com- pleted the Mini-VLAT test and interpreted 16 charts selected from a set of 40 real-world visualizations. Our qualitative and quantita- tive analyses show no significant relationship between participants’ Mini-VLAT scores and high-level comprehension. Our findings suggest while visualization literacy tests like Mini-VLAT capture viewers’ ability to perform directed, low-level tasks, they do not as- sess high-level comprehension in natural, uncued settings. Future work should develop VLAT that accounts for high-level compre- hension beyond estimating prespecified statistical quantities.