The Role of Learning Analytics in Education
In October of 2016 at NYU, the Office of the Provost, the Center for the Advancement of Teaching, and NYU Information Technology hosted “Understanding the Role of Learning Analytics in Technology-Enhanced Learning,” a discussion and panel presentation that focused on the changing role of learning analytics in higher education and assessment.
Various thought leaders from universities and organizations around the country came together at the event to talk about the emerging role of learning analytics in enhancing educational outcomes, as well as some early insights being collected.
The conference featured two panel discussions along with afternoon workshops. The morning panel featured speakers Andreas Paepcke (Stanford University), Christopher Brooks (University of Michigan), Alfred Essa (McGraw-Hill Education), and was moderated by Ted Madger (NYU Steinhardt Vice Dean for Academic Affairs). Ted Magder gave opening remarks about the importance of how learning analytics may lead to the fundamental change in the practice of teaching and real disruption.
"Learning analytics as an idea may very well lead to absolutely fundamental changes in the practice of teaching and real disruption. We have to proceed with caution, and we have to proceed collaboratively. We have to make sure as we’re making these decisions we’re not making decisions solely on the basis of what the data is telling us to do, but on the basis of conversations we’re having with faculty, and we have to think too about how to involve students in the process," said Madgers.
Andreas Paepcke provided a historical look at the evolution of the university to putting courses online particularly in the format of MOOCs. Digitizing lectures did not improve or radically change the pedagogy but did provide ample amounts of data and insight into what students were doing in these spaces.
Christopher Brooks provided valuable insights from the efforts at University of Michigan, and he explored various examples of the use of learning analytics from predictive models to academic reporting tool kits and eCoaching.
Alfred Essa from McGraw Hill spoke about the value of learning analytics to address retention and graduation issues at the national level. Learning analytics can help identify at risk students and provide recommendations in terms of courses to address knowledge and skill gaps. The end game is really providing educational opportunity and access.
During the afternoon panel discussion with Ryan Baker (University of Pennsylvania), Alyssa Wise (NYU Steinhardt), Vincent Aleven (Carnegie Mellon University), and moderated by Yoav Bergner (NYU Steinhardt), the panel discussed ways that learning analytics may be applicable and ways they may not. All panelists provided valuable insights to the questions posed by Yoav Bergner.
Of note, Alyssa Wise spoke about how we are never going to have the perfect algorithm. We are going to get better ones but they will not be perfect. She cautioned about how we make use of data and analytics that are inherently subject to interpretation. In other words, we are not looking for the data to tell us what to do, but rather, we are looking for data to inform decision making. Moving to this paradigm makes the data much more manageable, and also requires data literacy on the part of faculty, researchers, and students.
Later in the day, Marina Marin and Martin Pusic from the NYU School of Medicine, Institute for Innovations in Medical Education gave a presentation on their work using learning analytics and the insights they have generated from that.
Participants learned how these tools are increasing in popularity among students, as well as how they are driving learning insights. For example, the two following images from NYU Stream demonstrate how learners are interacting with Stream media, and also provide faculty with detailed engagement analytics to see who is watching and help determine if learning objectives are being met.
Instructors and class leaders were shown how the NYU Survey Service (powered by Qualtrics) is another tool they can leverage to record and assess student learning. The survey-building tool incorporates easy learning assessments into teaching and generates metrics of how many students understood a particular concept, along with clean and simple visualizations, charts, and histograms. This combination allows instructors a powerful glimpse into how well students have learned a piece of material or progressed through a module, and can also be used by the instructor to quickly adjust the timing or content of a lesson or class sequence to ensure better uptake of information.
NYU Web Publishing specialists were on hand at the conference discussing and demonstrating the utility of delivering online content through the service. In 2015, NYU communicty members spent an average of 12 minutes per day engaged with a Web Publishing site. While in 2016, usage increased by more than one hundred fifty percent with community members spending on average 31 minutes per day.