American Educational Research Association (AERA) Annual Meeting
Theme: Leveraging Educational Research in a “Post-Truth” Era: Multimodal Narratives to Democratize Evidence
April 5-9, 2019
Title: Opening the Black Box: Log Analyses of E-Book Engagement and Vocabulary Growth (Paper)
Event: Improving Literacy Through Digital Scaffolding
Authors: Osman Umarji, Ying Xu, Elham Zargar, Renzhe Yu, Moshe Yang, Stephanie Day, Carol M. Connor
Abstract: Scaffolding can help young readers develop effective comprehension strategies. Interactive ebooks have opened new opportunities to promote strategic reading by embedding questions, which prompt readers to monitor their comprehension. Previous studies have suggested ebooks with embedded questions support children's reading development, yet little is known about students e-reading behaviors -- the mechanisms through which they learn vocabulary and reading comprehension skills. Taking advantage of user log data generated from a choose-your-adventure ebook, this study examines specific e-reading behaviors, e-reading outcomes, and their associations with reading comprehension.
Children’s reading behaviors predict their reading comprehension. Research conducted in print-based reading contexts has identified behaviors that predict comprehension, including number of words read (Topping, Samuels, & Paul, 2007), reading speed, and comprehension monitoring (Wasik & Bond, 2001). While interactive ebooks can embed questions to support strategic reading, technology alone cannot replace the need for teachers to scaffold the development of reading strategies (Warschauer, 2006). Therefore, this study included an RCT to test whether a teacher-led book club influenced reading behaviors and outcomes.
Two variables of reading behaviors were created from the user logs for average time on reading pages and average time on question pages. Two variables of e-reading outcomes were derived from the logs for the percentage of comprehension questions answered correctly and the number of traps readers fell into.(e.g., when given a choice of what to do and only one choice made sense). Structural equation modeling was used to model these log variables, along with pretest vocabulary scores, grade level, and book club condition to predict vocabulary gains.
The participants in the study were 581 third, fourth, and fifth grade students (49% female, mean age = 9.5 years) from two low-income schools in South Central Arizona. These participants read the ebook in class over three weeks. User logs and students test scores on vocabulary were used in the analysis (for details, see Connor et al., 2018).
The model fit the data extremely well (χ2(4) = 8.46, p = .08; RMSEA = .04; CFI = .99; TLI=.97). Posttest vocabulary scores were predicted by pretest vocabulary (β=.57, p<.001), percent of questions answered correctly from the ebook (β=.16, p<.01), and participation in the book club (β=.05, p<.05). Percent of questions answered correctly were predicted by pretest vocabulary (β=.35, p<.001), time spent on questions (β=.25, p<.001), and students’ perceived calibration ability (β=.09, p<.05). Third graders did worse overall in both percentage of answering questions correctly and in vocabulary scores than fourth and fifth graders, although fifth graders did not differ from fourth graders.
This study sheds light on how young readers engage with interactive ebooks and investigates how specific behaviors relate to gains in vocabulary. This study also reveals how teachers can further scaffold ebook engagement through book clubs that promote the usage of comprehension monitoring strategies, highlighting the importance of teacher scaffolding in the usage of ebooks.