Abstract:
Visual representations such as graphs and charts are important tools to make data and models more understandable. In our daily life, we are confronted with graphs that visualize information, such as election results or the development of stock prices. Across disciplines, graphs and charts are also used in research and teaching to explain domain concepts, for example when the relationship between price, demand and supply is modeled in economics or when biologists analyze how prey and predator populations influence each other over time. The ability to understand these visual representations is therefore not only necessary in daily life but also part of domain expertise. Although graphs and charts are omnipresent in the 21st century, prior research has demonstrated that all students cannot be assumed to intuitively understand these visualizations. In contrast, large-scale and in-depth studies have identified various difficulties of learners, for example when they are unable to read and interpret data graphs or cannot connect visual representations to the underlying domain principle. Different research communities have thus modeled and analyzed the ability of learners to work with visual representations (mostly in the science domains) and investigated the effect of instructional support, which helps learners to understand graphical representations and relate them to their context. Bringing together these different research disciplines, this dissertation investigates, in three separate studies, how learners can read data graphs; how visual representations are used in secondary economic education; and lastly, how learners can be supported in integrating graphs and text.
In the first study, eighth graders’ ability to read graphs was investigated. Focusing on graphs related to sustainable development, which students could encounter in their everyday life, the study measured how well they could read single data points and trends or perform small extrapolations. The instrument was used with 198 students from four different schools, all with the highest school track (Gymnasium). To test whether tasks with increasing complexity (from points to extrapolations) would also be more difficult for learners, the data was analyzed with item response theory. Furthermore, the relationship between graph reading and learner characteristics such as academic performance, motivation, interest and domain knowledge was examined. The results revealed that eighth graders were able to read data graphs rather well, and no systematic relationship was found between what an item asked for and the item difficulty. The ability to read graphs correlated with academic language performance, academic math performance, as well as content knowledge and prior engagement with sustainable development. In the second study, a textbook analysis was combined with teacher interviews to investigate the use of graphical representations in learning material and instruction. To gain an overview of the graphical representations in textbooks, they were categorized according to their form (graph/chart) and the extent to which they visualize a domain principle. In 10 semi-structured interviews, teachers were asked how and why they use graphical representations, what they expect of their students and what typical mistakes their students make when they work with visual representations. The teacher interviews revealed that graphs and charts are used regularly in teaching, not only to visualize economic models but also to display data related to economic variables (e.g., development of growth domestic product [GDP]) and to train students in critically analyzing graphical representations. The following are among the challenges for learners: math- and data-related issues (e.g., when learners are unable to differentiate between absolute and relative numbers) and a lack of integration of representation and domain (e.g., when learners cannot identify the relevant information for a domain question or are unable to connect graphical information to other external representations such as texts).
Finally, through a quasi-experimental design, the third study tested how learners can be supported in learning with text and graphs. For this purpose, students received learning material from two domains in one of three conditions: Either the correspondences between text and graph were already highlighted or they were asked to highlight the relevant connections themselves while studying the material (active signals) or, lastly, without alterations to design or learner-task (control). After the study phase, the learning outcome was tested with recall and comprehension questions. Overall, students who studied already signaled material performed equally well compared to the control group. On average, students in the active-signal group achieved significantly fewer points in the biology posttest compared to the control group. When learners had high prior knowledge, however, they could profit from actively integrating both representations. Furthermore, in economics, the relationship between prior knowledge and learning outcome was partially mediated via the quality of learner-generated signals; that is, learners with high prior knowledge were better in connecting graphs and text, which in turn was associated with higher learning outcomes.
In this dissertation, the findings of these three studies are summarized and discussed against the background of the research context in different disciplines (economic education, science education and educational psychology). At the end, implications for future research and educational policy and practice are derived.
logical pictures
Print-on-Demand