Translating Educational Theory Into Educational Software: A Case Study of the Adaptive Map Project

In this paper, the authors describe the development of an instructional software, where developers engaged in the process of translating educational theory into a cyber-learning tool, and the challenges encountered in evaluating its usability and effectiveness of the tool as a learning aid. Specifically, the authors reflect on their experience in creating the “Adaptive Map” – an instructional software designed to help students gain conceptual understanding of large stores of content information. This concept map -based system explicitly shows how discrete concepts are linked to the whole of the course with a large node-link diagram. This explicit mapping of expert knowledge structures has been shown to promote conceptual understanding in students. Because concept maps become visually cluttered and unusable when they get too large though, an interactive visualization tool was developed to maintain the advantages of concept maps as learning tools while managing the visual clutter in maps that cover entire courses or even an entire curriculum. In this paper, the authors discuss the process they took in integrating the educational literature with the information visualization literature to understand how to best make an information visualization that addresses educational goals. Results from a heuristic analysis using Munzner’s four level validation framework for an information visualization are also presented. 1. Background and Motivation Technology has the potential to aid instruction, but the simple act of using technology to deliver instruction does not improve the instruction being delivered [1] . In order to have a positive impact on student learning, instructional technology developers must draw on what is known about how people learn and then use technology to improve the quality of the instructional materials. This often involves collaboration between researchers with backgrounds in education and those with backgrounds in software, or other technologies. This paper serves as a case study of one such instructional software development process, the development of the Adaptive Map digital textbook, so that future collaborations might gain insight into developing instructional technologies. 1.1 The Theory Behind the Adaptive Map The Adaptive Map is an attempt to use the flexibility of modern software to improve the design digital textbooks in ways that was not previously possible. The Adaptive Map tool is a collection of content pages that collectively serve as a textbook for a course. In its current iteration, the Adaptive Map (available at www.adaptivemap.me.vt.edu ) has been developed with content in engineering statics. Statics was chosen for an initial content area because it is a P ge 23270.2 foundational course for a variety of engineering majors, thus maximizing the potential impact of the tool. The central design feature of the Adaptive Map tool is a concept map based navigation system. Research studies have previously found that by using expert generated concept maps to show how different topics in lesson are related, students are better able to understand and retain the content information presented [2–6] . Expert-generated concept maps serve as advance organizers [7] and improve understanding by mobilizing relevant prior knowledge the new content can be cognitively connected too [8] . However, concept maps have problems with scalability. When concept maps become too large and complex, users encounter “map shock” [9] . Map shock is a cognitive and affective reaction to large scale concept maps where learners become overwhelmed and disengage from processing the concept map. This diminishes the effectiveness of large expert generated concept maps as learning tools. Concept maps are therefore effective for organizing small amounts of content information in a way that students can better understand it, but they cannot currently be used to organize large quantities of information. This limits the effectiveness of concept maps as a way to enhance the effectiveness of course textbooks. The symptoms of map shock match the more broadly defined phenomenon of cognitive overload, as described in cognitive load theory [10, 11] . In order to avoid map shock in learners, the visuals presented must be managed in a way that helps learners manage their cognitive load. In order to better manage the cognitive load of users, the authors used information visualization techniques to create the Adaptive Map’s interactive navigation system for large scale concept maps. This instructional tool was built to test the hypothesis that information visualization techniques could be used to increase the effectiveness of large scale concept maps as advance organizers by managing a user’s cognitive load and eliminating map shock. 1.2 Overview of the Adaptive Map Tool The Adaptive Map software features semantic zooming techniques to enable the user to explore the large-scale concept map. With this technique the scope of the material being covered determines the level of detail presented in the visualization. By having the software limit the amount of information presented at any one time, the software is managing the cognitive load imposed on the learner. The tool opens by presenting users with an overview of all the information covered at a high level of abstraction. At this overview level the topics are grouped into clusters of highly related ideas, similar to chapters in a traditional book. Each cluster is represented as a node in a graph, and the links between the nodes represent direct relationships between the topics in each of the clusters. These links are directed and generally flow from more basic prerequisite clusters at the top of the screen to more advanced post-requisite concepts at the bottom of the screen. The P ge 23270.3 links’ line thickness is directly related to the number of direct connections between topics within those two clusters. An image of the overview can be seen in Figure 1. Figure 1: Adaptive Map Overview The user at any level can pan by either clicking on cluster nodes to center them on the screen, or by clicking, holding and dragging the background. The user can also zoom in or out using scroll wheel, or by using the + and – buttons on the screen. If the user zooms in to (or clicks on) any of the clusters in the overview, the cluster node will break apart into topic nodes and give the user more details on that cluster. Each cluster node contains several topics, where a topic was defined as the smallest independently teachable lesson. Information on how the topics and clusters were identified can be found in previous literature [12] . At this level, the topics are represented by individual nodes in the concept map and the relationships between the topics are represented by links. A sample screenshot of the “Static Equilibrium” cluster is shown in Figure 2. Figure 2: Adaptive Map Cluster View P ge 23270.4 In Figure 2, the user is focused on the static equilibrium topic, as indicated by the yellow border around the node and the description of the topic in the node. The background color and all the topics within the static equilibrium cluster are red to match the color of the static equilibrium cluster node in the overview. Any topic that is directly related to the focus topic from other clusters is also drawn in. In this case, the two force member topic is directly related to the static equilibrium topic, though the two force member topic is part of another cluster. These crosscluster relationships are amalgamated into the links seen in the overview. More details on the topics or the relationships can be found by hovering over the nodes or link in these views. If one zooms in further to a topic nodes, one will view the content page associated with that topic. These topic pages resemble the content in more traditional textbooks, with explanations, images and worked example problems. A screenshot of part of the “Static Equilibrium” Topic Page is shown in Figure 3. Figure 3: Adaptive Map Topic Page Through these three levels of zoom, learners can explore the topics and the relationship between topics contained in the adaptive map. The interface helps users maintain a sense of context within the overall concept map, and helps match the level of detail displayed to learner’s current level of focus (from a course wide overview, to a chapter overview, to a single topic focus). 2. Software Development Process 2.1 Developing the Software Goals The software development process began with the formation of software design goals. These design goals were based on an extensive review of the educational literature relating to concept P ge 23270.5 maps and cognitive load theory. In addition computer science literature related to information visualization, particularly the information visualization literature pertaining to graphs, was consulted. The software design goals and citations of the related literature are listed below. 1. The tool will act as a digital content repository with a concept map based navigation scheme. The tool must be able to display detailed information as well as concept maps of the embedded information at several levels of detail and abstraction, and it must allow the user to easily navigate horizontally (from one topic to another) and vertically (from one level of detail to another). 2. Usability of the visualization as both a learning and navigation tool is the primary goal of the software. The information should be displayed in a way that does not cognitively overload [10, 11] the user. 3. The tool will automatically generate the visualizations based on metadata from the content developer. The content developer will develop content pages with metadata to inform the software what other topics a given idea is directly related to, what the nature of those relationships are, and to what groups the topic belongs. The software should interpret this information and determine how to best visualize the