Cognitive Task Analysis for Implicit Knowledge About Visual Representations With Similarity Learning Methods.

摘要: Visual representations are prevalent in STEM instruction. To benefit from visuals, students need representational competencies that enable them to see meaningful information. Most research has focused on explicit conceptual competencies, but implicit perceptual might also allow efficiently information visuals. common methods assess students' rely verbal explanations or assume attention. However, because and not necessarily verbally accessible, these ill-equipped them. We address shortcomings with a method draws similarity learning, machine learning technique detects visual features account for participants' responses triplet comparisons of In Experiment 1, 614 chemistry judged the Lewis structures 2, 489 ball-and-stick models. Our results showed our can detect drive perception suggested knowledge about molecules informed through top-down processes. Furthermore, 2 tested whether we improve efficiency active sampling. Results random sampling yielded higher accuracy than small sample sizes. Together, experiments provide first implicitly, without requiring verbalization assuming These findings have implications design instructional interventions help acquire competencies.