ABOUT WCER NEWS Events Cover Stories Research News International Research Press WHAT'S THE RESEARCH ON...? PROJECTS All Active Projects All Completed Projects PUBLICATIONS LECTURE SERIES PEOPLE Staff Directory Project Leaders ERG - EVALUATION RESOURCES GROUP RESOURCES Conference Rooms Equipment GRANT SERVICES GRADUATE TRAINING SERVICE UNITS Director's Office Business Office Technical Services Printing & Mail EMPLOYMENT CONTACT INFO MyWCER WORKSPACE LOGIN

Sadhana Puntambekar

August 2006

Hypertext learning environments hold great promise, but they also present challenges for learners and designers. As part of her research into designing optimal learning environments, UW-Madison education professor Sadhana Puntambekar developed the project she calls CoMPASS (Concept Mapped Project-based Activity Scaffolding System).

CoMPASS consists of a hypertext system and curriculum modules based on the pedagogical framework of Learning By Design,¹ which has students learn science in the context of design-and-build challenges. The CoMPASS system includes two tightly integrated representations of content, one textual and the other visual (in the form of concept maps). Each CoMPASS screen (see illustration) represents a concept such as mass or gravity and provides both a concept map (left half of the screen) and a textual description (right half of the screen).

Concept maps display a fish-eye view—that is, the focus concept appears at the center, with the most closely related concepts displayed in the first ring and the less closely related concepts in the outer ring. The maps show connections between science phenomena and give students alternative paths to pursue so they can see how different phenomena relate to each other.

In contrast to a hierarchical index, which orients the learner but does not show all the relations between concepts, concept maps present ideas in the form of nodes linked by a word. In its simplest form, a concept map is just two words connected by a linking word to form a semantic unit (for example, friction and force connected by the link type of). Concept maps help students visualize the interconnections between concepts.

Acquiring advanced knowledge requires students to revisit the same material at different times, in different contexts, for different purposes, and from different perspectives. The alternative views in CoMPASS make this possible, helping students study science concepts and phenomena in depth by encountering them in multiple contexts. For example, a student might want to learn about the phenomenon of force in the context of objects falling in air. In CoMPASS, the student could change views to study the same concept (force) in other contexts, such as linear motion or simple machines.

To take another example, in one physics unit parachutes serve as an example of falling objects. If a student selects the concept motion of falling objects, CoMPASS provides an explanation of that phenomenon, along with a map of related concepts and principles. The student may then choose any concept from this map, and CoMPASS creates the corresponding fish-eye view. The maps reflect the semantic relatedness of the concepts within a topic and show students what related topics they can go to.

Evaluating CoMPASS’ Effectiveness

Puntambekar recently evaluated the effectiveness of CoMPASS by studying two classes of eighth graders who used the system to learn about force and motion. One class, the maps class, used CoMPASS; the second, the index class, used an index version of the system that presented lists of concepts instead of navigation maps. Both classes used the same text, teacher, and curriculum.

In the study, students were asked to design roller coasters. As the students raised questions about how roller coasters work, they encountered the concepts of force and motion. After finishing their initial designs, students built models, exploring concepts about the physics involved. During each class, students used CoMPASS to answer their questions.

Puntambekar found that, compared to students in the index class, students in the maps class

• visited more goal-related concepts and spent more time on them;
• improved more on an essay question; and
• performed better when tested on their depth of knowledge.

The two classes did not differ significantly in their knowledge of facts as demonstrated on the pretest and posttest. But students in the maps class improved more on the essay question than did the students in the index class. In fact, students in the index class performed worse on the essay question in the posttest than in the pretest. Puntambekar speculates that the index presentation, which was linear, may have been detrimental to learning.

The study suggests that the type of navigational aid may have affected students’ navigational decisions and thereby their learning. The index class largely followed the structural aids provided in the index, no matter what the goal. Although the goals were increasingly open-ended over the course of the three days, the structure of the index consistently drove the navigation. In contrast, the navigation used by the maps class clearly reflected the particular goal on each of the three days.

For more information visit the CoMPASS web site. Puntambekar’s research is supported by an NSF CAREER grant.

¹ Kolodner, J. L., Crismond, D., Fasse, B., Gray, J., Holbrook, J., & Puntambekar, S. (2003). Putting a student-centered learning by design™ curriculum into practice: Lessons learned. Journal of the Learning Sciences, 12(4), 485–547.