Relationships between Facial Expressions, Prior Knowledge, and Multiple Representations: a Case of Conceptual Change for Kinematics Instruction

Abstract

Kinematics is an important but challenging area in physics. In previously published works of the current research project, it was revealed that there is a significant relationship between facial microexpression states (FMES) changes and conceptual conflictinduced conceptual change. Consequently, the current study integrated FMES into a kinematics multiple representation instructional scenario to investigate if FMES could be used to help construct students' conceptual paths, and help predict students' learning outcome. Analysis revealed that types of students' FMES (neutral, surprised, positive, and negative) were important in helping instructors predict students' learning outcomes. Findings showed that exhibiting negative FMES through all three major representation segments of the instructional process (i.e., scientific demonstration, textual instruction, and animated instruction) suggests a higher probability of conceptual change among students with sufficient background knowledge on the topic. For students with insufficient prior knowledge, the result was the opposite. Moreover, animated representation was found to be critical to the prediction of student conceptual change. In sum, the results showed FMES as a viable indicator for conceptual change in kinematics, and also reaffirmed the importance of prior knowledge and representations of scientific concepts.

Practices and Theories: How Can Machine Learning Assist in Innovative Assessment Practices in Science Education

“…This is because multiple-choice items are difficult to elicit higherorder thinking that is associated with sophisticated cognitions and performance. The articles in this special issue extended the approaches to collecting evidence in ways such as virtual reality (e.g., Sung et al, 2020), representations (e.g., Zhai et al, 2020c), and facial expression identification (e.g., Liaw et al, 2020). In their study, Sung et al (2020) employed the augmented reality technology with a thermal camera attached to a smartphone to elicit students' understanding and asked students to write constructed responses; students' responses were analyzed using a deep learning approach.…”

“…Zhai et al (2020c) demonstrated how to apply deep learning to automatically evaluate students' modeling competency by scoring their drawing and writing explanations. Liaw et al (2020) also employed a deep learning-based app to automatically identify students' emotion and examined how their emotion was associated with their learning outcomes. While these measures are only likely to be achieved with novel technology, it should also be noted that these measures have significantly extended humans' ability to infer students' competence and thinking by examining complex products developed by students or their problem-solving procedures.…”

“…I realized that the approaches to collecting evidence were also improved because of ML. For instance, in their study, Liaw et al (2020) employed a facial expression identification approach to capturing students' emotional information during learning and applied ML algorithms to predict student learning outcomes. It is also delightful to find that the timely feedback assisting students in meeting challenges in scientific practices was improved by using ML (e.g., Lee et al, 2021).…”

“…It is evidenced that the articles in this special issue present examples of how technologies (e.g., ML, simulation, facial expression identification) are integrated with other components of assessment practices. These technologies support the development of authentic assessment tasks (Sung et al, 2020), collecting data (Liaw et al, 2020), automatically scoring student responses (Rosenberg & Krist, 2020), etc. However, it should be noted that technology can be useful only if it is situated within authentic task scenarios and integrated with appropriate learning goals (Neumann & Waight, 2020).…”

On the Validity of Machine Learning-based Next Generation Science Assessments: A Validity Inferential Network

“…Compared with traditional statistics, ML can handle data with an enormous number of variables using methods beyond linear regression approaches. Also, ML is appropriate for analyzing process data such as log files to provide insightful diagnostics of student learning (Liaw et al 2020;Shin and Shim 2020). Given such great potential, as found in a recent review study (Zhai et al 2020a), three major fashions of ML, supervised, unsupervised, and semi-supervised machine learning have been applied to automatically score constructed responses, essays, educational games, simulations, and interdisciplinary assessments, showing great potential to advance science assessment practices.…”