Visual representations of data are increasingly prevalent, but we lack a detailed theoretical framework to explain what factors make easy or difficult to read and understand; nor do we know how such factors can impact data visualizations’ efficiency as learning material. In this Master’s Thesis, I start to address this gap by exploring the validity and applicability of Cognitive Load Theory, an educational and cognitive science theoretical framework, for designing and evaluating data visualizations. To that end, I conducted an online randomized study in which each participant (N=34) performed learning tasks on two different data visualizations from the 6th IPCC Summary for Policy Makers. One was presented in three successive parts, according to the “segmenting” principle in Cognitive Theory of Multimedia Learning; and the other was presented as a single image. Although most learners preferred the segmented style, this treatment did not significantly affect the overall mental effort they reported. I found significant negative correlations between readability and extraneous cognitive load, indicating that learners felt they needed to put less effort into learning from a visualization when they found it more readable. In addition to a qualitative analysis of learners’ preferences, in this work I also contribute an interdisciplinary perspective on visualization cognitive processing and a discussion of implications for evaluating readability in visualization studies. Highlights: • Visualization readability was strongly and negatively correlated with extraneous cognitive load in learners; • Segmenting data visualizations did not significantly reduce overall mental effort in learners; • A majority of participants preferred the segmented style as they felt guided in reading and understanding the content, but the need to scroll between parts of the visualization could also feel cumbersome at times.