Poster #88 - Variability in Log-in Time Predicts Progress on Personalized Learning Software

• In Event: 1-169 - Poster Session 04
  In Poster Session: PS 04 Section - Education, Schooling

Integrative Statement

Introduction: Prior research shows that the time of day in which individuals learn or perform other cognitively demanding tasks can affect performance. Both children and adults show better academic learning when they learn at the peak time in their circadian cycle (e.g., “morning people” learn better in the morning; Randler et al., 2009; Cruz et al., 2016). Unfortunately, traditional classrooms do not allow children much control over when certain subjects are taught. Newer educational philosophies like personalized learning, which assumes that the strongest learning arises from teaching that is individualized to the student, emphasize the importance of flexible scheduling in supporting personalization (Pane et al., 2017). Despite growing interest in personalized learning and flexible scheduling, little research has looked at how growth is affected by variable learning schedules.
Methods: In this study, we explore how the time of day in which children use an online personalized learning platform designed to promote literacy affects their progress in the platform, as well as generalization to a standardized measure of English language ability. This online platform is a component of a blended learning program that combines personalized online instruction with teacher-led lessons. Our sample consisted of 25211 logins from 481 children in grades Pre-K - 5 in a single school in California during the 2017-2018 school year. Data was collected as part of a larger study where children used the online platform during some combination of English Language Arts blocks and computer lab time, as well as miscellaneous times in and out of school. We used linear mixed-effects models with random effects for classrooms nested within grades to investigate the effects of time of day variables on two program outcomes: 1) total number of units completed and 2) whether students met a grade-level benchmark.
Results: For both of these outcome variables, the variability (standard deviation) of login time within students predicted significant variance over and above multiple control variables. These models show that children with more variable login times were more likely to have completed more units by the end of the school year than those who logged into the software at more consistent times of day. They were also more likely to have met grade-level benchmarks. Although meeting benchmark in the online platform predicted standardized English Language test scores, variability in login times did not.
Discussion: This study suggests that flexible scheduling of personalized learning can lead to better performance and perhaps more learning. Multiple explanations can describe these findings. First, children with more variable login times may be electing to use the program at times they prefer, which could fluctuate across days. In addition, distributing learning opportunities across multiple times of day (on- and off-peak) could allow children to sample times of day that are likely to prioritize explicit and implicit learning strategies (May et al., 1993, 2005). Follow-up research should focus on how to transfer knowledge acquired in the online platform to everyday reading situations in a way that is measured by standardized tests.

Authors

• Kayleigh Ryherd, University of Connecticut
  Presenting Author

• Rachel L Schechter, Lexia Learning
  Non-Presenting Author

• Nicole Landi, University of Connecticut
  Non-Presenting Author

• Emily B Myers, University of Connecticut
  Non-Presenting Author