In-the-Moment Digital Reading Behaviors and Links to Comprehension for Young Adolescent Learners

• In Event: Text(s) Comprehension: Learner Factors, Processes, and Supports

Abstract

Purpose: Considering multimodal in-the-moment digital reading behaviors are central to understanding reading comprehension. The current study explores trends and uses machine learning techniques to investigate the role of in-the moment digital reading behaviors including eye-gaze, emotional response, and digital reading behaviors in supporting reading comprehension.

Methods: Recurrent neural networks (RNNs) using data from 191 5th-8th grade students were used to show how reading behaviors relate to reading comprehension beyond demographic and pretest data using multimodal time-series data.

Results: Analyses suggest including process data results in vastly improved predictions of students' reading comprehension compared to similar models with only non-process data.

Conclusions: Results suggest the need to consider in-the-moment reading behaviors. Implications for theory, research, and practice are shared.

Authors

• Amanda P. Goodwin, Vanderbilt University

• Sun Joo Cho

• Matthew Naveiras, Vanderbilt University

• Jorge Salas, Vanderbilt University