Learning Science & Insights

Learning Research and
data-driven insights

We rigorously assess and synthesize education research and cognitive science into principles that guide every design decision

ID Principle Analytics Type Confidence Citations
2 Analytics naturally leverage or can be implemented in support of a variety of learning theories. Examples include metacognition, self-regulated learning, project-based learning, collaborative learning, and constructivism. Specific processes supported include reflection, self-monitoring, self-judgement, and study strategies. Both HIGH 18
4 Select data that directly align to the learning design. For instance, tracked data should support intended goals and behaviors. Dashboards HIGH 14
5 Initial data displays should be simple and present an overview of the data. At the user's request, more detailed data points or original source artifacts may be explored. Dashboards MED 11
7 Analytics (including dashboards and prompts) and interventions can be used in combination to identify at-risk learners, then improve their performance over time. Both MED 10
8 Just-in-time prompts can support the reflection, goal setting / evaluation, and behavior adjustment / intervention processes. Prompts MED 9
9 Prompts may be provided in real time. For instance, an automated notification may be sent to a learner the moment an achievement is earned. Prompts MED 8
12 Align the reporting of analytics to the timescale of events in the learning design. For example, achievements may be made in real time, discussions may occur weekly, or exams may occur quarterly. Data should be reported in ways that is consistent with the real-world activity it is based upon. Dashboards MED 9
13 Prompting self-regulation strategies (e.g. organize, monitor, plan) and remedial strategies can lead to improved performance and strategy use. Prompts MED 5
15 Ensure learners are aware of and understand the pedagogical intentions behind the design and use of analytics. Both MED 8
16 Prompts may be scheduled to align with the design of the learning experience. For instance, an automated activity summary may be sent to learners each week in a course that discusses a new topic every week. Prompts MED 7
17 People prefer visual representations as compared to text-based representations in dashboards. For example, graphs and charts are prefered over tables and text. Dashboards MED 7
20 When using analytics, note that there is a risk that stakeholders will perform only according to what is measured and conveyed, which is unlikely to fully represent ideal behavior in the context. Both MED 9

We use the same research to construct a learning model that provides the blueprint for a product design

The design principles and models are critiqued by experts

Dr. Christopher Dede

Timothy E. Wirth Professor in Learning Technologies, Harvard

- new types of educational systems for the 21st Century, large-scale educational improvement initiatives

Dr. Mark McDaniel

Professor of Psychology and Co-Director of the Center for Integrative Research on Cognition, Learning, and Education (CIRCLE), Washington University in St. Louis.

- human learning and memory, with an emphasis on prospective memory, encoding and retrieval process

Dr. Robert Atkinson

Associate Professor in the School of Computing, Informatics, and Decision Systems Engineering in the Ira A. Schools of Engineering and the Division of Educational Leadership and Innovation in the Mary Lou Fulton Teacher’s College, Arizona State University

- personalized learning, social media, learner analytics, mobile learning, cognitive science, usability testing, human-computer interaction

Learn More About Our Advisory Councils →

We use education research to identify qualities critical to a student’s success...

...and ethnographic research to build personas of real-life students who we design to help and then test against

We use extensive data mining to reveal empirical insights into how students learn, study, behave, and are motivated

Feature Use A

Detailed analysis of students using a prototype eBook reveals how they review content (the sequence, time on task, frequency of revisits, and subsequent actions), the tools they use (search, bookmark, note taking), and how both vary during the semester. Analysis of where students get stuck and the impact of different interventions guide us with where to place links to the eBook and other learning resources optimally.

We use data forensics to examine and improve the efficacy of each learning and assessment item

We analyze anonymous student performance and engagement on every learning and assessment item nationally, by course, major, and other segments. This allows us to iteratively, continuously, and materially improve the efficacy of content we offer.

We also use data mining to explore national trends in course design and success

Network Graph
Subtopic Mastery

We analyze coverage of learning objectives and student mastery to identify course trends and successful designs