Investigation of Visualization Literacy: A Visualization Sensemaking Model, a Visualization Literacy Assessment Test, and the Effects of Cognitive Characteristics

Sukwon Lee, Purdue University


By using data visualizations, we can amplify the cognitive and analytical capabilities of people. It allows people to complete tasks and make logical decisions with large data sets. Even though data visualizations effectively and concisely represent complex data, it does not necessarily mean that data visualization users comprehend the visualizations well and appropriately read and interpret visually represented data. Nevertheless, various data visualization techniques and tools are being developed as the amount of available data increases. Given this situation, an individual’s ability to read and comprehend data visualizations, visualization literacy, is becoming as important as the ability to read and comprehend text because the individual’s comprehension and interpretation of data visualizations can strongly influence his/her tasks and communication. However, we still do not have clear answers about some critical questions concerning visualization literacy: “what is visualization literacy?” “how do people make sense of data visualizations?” “how can we measure the visualization literacy of an individual?” and “what is the nature of visualization literacy?” Thus, in this dissertation, I focus on investigating visualization literacy through a series of studies. In the first study, I qualitatively explore data visualization sensemaking activities as a first step toward visualization literacy. Visualization comprehension is one of the critical components of visualization literacy. Thus, I endeavor to understand how people make sense of data visualizations. I observe users when they try to make sense of unfamiliar data visualizations. I collect think-aloud data from the observation and analyze the data using a qualitative inquiry approach, the grounded theory method. As a result, I identify five salient cognitive activities (i.e., encountering visualization, constructing a frame, exploring visualization, questioning the frame, and floundering on visualization) and devise a grounded model of novice’s information visualization sensemaking (NOVIS model). In the second study, I develop an assessment test for measuring the visualization literacy of users. In order to gain deeper understanding of visualization literacy, we should be able to measure and evaluate the ability in a practical manner. However, we still lack instruments for measuring visualization literacy. In order to address this gap, I systematically develop a visualization literacy assessment test (VLAT) by following the established procedure of test development in Psychological and Educational Measurement. The VLAT consists of 53 multiple-choice items and covers 12 data visualization types and 8 essential data visualization tasks. I also share various evidence for validity of the VLAT, which is gathered from the test development procedure, based on the test content, reliability, and a criterion variable. In the last study, I examine the effects of cognitive characteristics on visualization literacy to understand the constructs of visualization literacy. In particular, I focus on three cognitive characteristics: numeracy as a cognitive ability, need for cognition as cognitive motivation, and visualizer-verbalizer style as a cognitive style. I measure users’ visualization literacy and the three cognitive characteristics using the visualization literacy assessment test (VLAT), the decision research numeracy test (DRNT), the need for cognition scale (NCS), and the verbalizer-visualizer questionnaire (VVQ), and analyze the measurement scores. From the analysis, I confirm that numeracy and need for cognition are influential cognitive characteristics in visualization literacy. However, I do not find evidence for the effects of visualizer-verbalizer cognitive style on visualization literacy.




Yi, Purdue University.

Subject Area

Industrial engineering|Cognitive psychology

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server