Diagram Translation With Concrete Models: Why Are Helpful Tools Not Used?

• In Event: Spatial Thinking in Chemistry

Abstract

Manipulations of concrete models are often assumed to aid in complex cognitive tasks by enabling internal thought processes to be externalized (Cary & Carlson, 2001; Gilbert & Boulter, 1998). In representation rich domains, such as organic chemistry, professionals and students must be both facile and accurate when translating between different representations of molecules, such as models and 2D diagrams (Keig & Rubba, 1993). We hypothesized that translating between diagrams would be more accurate when concrete 3D models were used to offload the students’ cognitive processes onto manipulations of the models.

In five studies, participants translated between different structural diagrams of molecules with or without the availability of concrete models. They were videotaped while drawing the diagrams and their use of the models was coded from the videotapes. In Study 1, spontaneous use of models was analyzed. In Study 2, participants were either encouraged to use models or received no models. In Study 3, participants either received into their hands models aligned with the given diagram or received no models. In Study 4, participants were either explicitly instructed on how to relate the structural diagrams to the models, received aligned models with basic instruction about the diagrams, or received no models. In Study 5, students were interviewed during initial diagram translation tasks (pretest), allowed to use models to check the accuracy of their drawings (intervention), and given additional diagram translation tasks after the intervention (post-test).

Results of these studies showed large individual differences in how much students used the models, with some students never using the models. Use of models was positively correlated with performance. This performance gain could be traced to offloading necessary translation processes onto manipulation of the models and not simply using models as a visual reference. Although model use was effective, models were used on only 45% of the trials (Study 2). Physically aligning the models with the given diagram (Study 3), and enhanced on the structural relations between the diagram and model (Study 4) did not appreciatively increase the use of models over basic instruction. Importantly, during the think-aloud protocol (Study 5) students’ performance improved in the post-test when they discovered their errors by comparing the concrete model with their solutions.

In summary, our results suggest that cognitive offloading of internal manipulations onto external representations can enhance problem solving in organic chemistry. However, the general lack of model use by some participants suggests that this cognitive offloading strategy is not necessarily readily discovered or easy. Cognitive offloading is likely affected by the perceived complexity of the external representations, individual differences in spatial abilities, and experience with the representations to be manipulated. Lacking the necessary ability or experience with models, students are less likely to rely on such external aids and more likely to utilize error-prone internal cognitive strategies or less reliable heuristics. When guided to use the models in a productive manner that illuminates their errors, participants discover the utility of the models. Merely providing models, without appropriate training, may not be helpful for most students.

Authors

• Andrew T. Stull, University of California - Santa Barbara

• Shamin Padalkar, University of California - Santa Barbara