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

Area of Science:

• STEM Education
• Cognitive Science
• Chemistry Education

Background:

• Visual representations are crucial in STEM learning, requiring students to possess representational competencies.
• Existing methods for assessing these competencies often overlook implicit perceptual skills and rely on verbalization, which may not capture non-explicit knowledge.
• Implicit perceptual competencies, distinct from explicit conceptual ones, are vital for efficiently interpreting scientific visuals.

Purpose of the Study:

• To develop and validate a novel method for assessing implicit perceptual representational competencies in students using scientific visuals.
• To investigate the role of conceptual knowledge in shaping perceptual competencies through top-down processing.
• To compare the efficiency of active versus random sampling in a similarity learning approach for assessing perceptual skills.

Main Methods:

• Employed similarity learning, a machine learning technique, to analyze students' judgments of visual similarity in molecular representations (Lewis structures and ball-and-stick models).
• Collected data from 614 chemistry students in Experiment 1 (Lewis structures) and 489 students in Experiment 2 (ball-and-stick models).
• Utilized triplet comparisons to identify visual features influencing students' perceptions, bypassing the need for verbal explanations or explicit attention.

Main Results:

• The developed method successfully detected specific visual features that guided students' perception of molecular structures.
• Findings indicated that students' conceptual understanding of molecules influenced their perceptual competencies via top-down mechanisms.
• In Experiment 2, random sampling demonstrated higher accuracy than active sampling for small sample sizes when assessing perceptual competencies.

Conclusions:

• This research presents the first implicit method to assess students' perceptual representational competencies with scientific visuals, without requiring verbalization or assuming explicit attention.
• The findings highlight the interplay between conceptual knowledge and perceptual skills in STEM learning.
• The study provides a foundation for designing targeted instructional interventions to enhance students' ability to interpret scientific visuals effectively.