DescriptionEye movement data include rich information on temporal and localization aspects of text reading. However, most analyzes simplify eye-movement data so that much of its original information is lost. This study explores the potential of using smart learning analytics technology for the assessment of reading comprehension to retain and utilize much of its rich information. In specific, eye movement data collected from 31 participants during reading given text is analyzed and imported in a learning analytics software tool, namely Smart CAT (Smart Configurable Assessment Tool), along with an adapted statistical model for reading from prior research (D’Mello, Southwell & Gregg, 2020). Smart CAT is a valid and reliable assessment software tool which allows the automatic application of two different machine learning (ML) algorithms, that is a Gaussian Naïve Bayes Network and a C4.5 Decision Tree, to provide inferences regarding their competence in reading at three different levels of mastery (i.e., low, medium, and high performance). The training and testing samples of the algorithms come from a within subject analysis of the eye tracking data collected from the reading of a given text. As this is ongoing research, only preliminary results exist. Finally, the data used in this study was anonymously collected and with the informed consent of the participants.
|Period||31 Aug 2022|
|Event title||EARLI SIG 27 Conference: Online Measures of Learning Processes|
|Location||Southampton, United Kingdom|