Leveraging on Assessment of Representational Competence to Improve Instruction with External Representations

Abstract

External representations are often used to explain complex scientific concepts. These representations can augment learning outcomes to include the ability to engage in highly cognitive tasks, such as creating novel solutions to scientific problems. Nevertheless, interpreting these representations can be cognitively demanding for learners with low levels of representational competence especially those unfamiliar with the learning context. Moreover, the representational competence of an individual learner can change depending on the difficulty of a given task. Therefore, assessment practices that (1) differentiate among learners of different levels of representational competence as well as (2) gauge its change along tasks with hierarchical difficulty can be quite informative for designing instruction with external representations. In this chapter, we provide an example for how assessment can be designed to meet the two objectives. Given that representational competence is context specific, we demonstrate how the first objective can be met through developing a valid and reliable assessment instrument that can discriminate between learners of related but different majors. To satisfy the second objective, we suggest utilizing a continuum of problems based on the hierarchical cognitive orders in revised Bloom’s taxonomy. Potential benefits of such practices are discussed based on a case study that included 111 college students from three different institutes. Holistic and Item psychometric analyses are also detailed to further elucidate how failure in instructional intervention can result from lack of representational competence that, if not accounted for, efforts of redesign may prove fruitless.