A View from the Front Lines: Building Adaptive Learning Courses for Large Section, Introductory Classes

In 2016, we—Drs. Chris Brown, Jeannie Grussendorf, and Mike Evans—became part of Georgia State University’s Association of Public and Land-grant Universities (APLU)/Gates Foundation grant seeking to explore the use of adaptive learning courseware in Georgia State classrooms, particularly in large section, introductory courses. Brown and Grussendorf are both highly involved with GSU’s “Global Issues” course (POLS 2401); Evans is highly involved in GSU’s “Introduction to American Government” course (POLS 1101). Both courses have many large sections (face-to-face 100+ per section; online 200++ per section), over 2,500 students (Global Issues) or 4,500 students (American Government) annually, and many different annual instructors, including a lot of Graduate Research Assistants (GRAs) teaching for the first time. Each of us had previously been developing tools and strategies to address common problems, such as different levels of prior student knowledge, students coming to class unprepared, professors needing to rely on basic lectures/pedagogy, and high “drop, withdraw, fail” (DFW) rates. Other courses involved in the grant, in Economics and Psychology, are similarly challenged. After much review through the grant’s internal processes, we chose to work with the company Realizeit on the two Political Science courses, as well as engage in academic studies, with IRB approval, assessing the impact of the adaptive learning courseware.

The adaptive Global Issues course was piloted full-scale in the Fall 2017, with more than 1,200 students across 13 sections (face-to-face and online; six taught by GRAs). In the Spring, we conducted a control study (half the sections adaptive, half “traditional”/non-adaptive). The non-adaptive, traditional course was also created with Realizeit, holding factors such as cost and interface constant, while mirroring pedagogically how the course had traditionally been taught (i.e., we specifically trying to test the courseware’s “adaptivity”).

Fall 2017 DFW rates in Global Issues fell nearly 20%; in the Spring, DFW rates were lower in the adaptive (19.1%) versus traditional classes (21.4%). Across 36 common test questions in the Spring, adaptive sections did significantly better than traditional sections (i.e., showing strong statistical significance). A 29-question student survey (roughly 800 student respondents total) indicated consistent, strongly positive and statistically significant impressions of the adaptive courseware over the traditional course across a wide-variety of indicators. Brief snapshots include: 74% agreed that the adaptive courseware helped them learn the material better; 69% said the instructor used the courseware to personalize the content or give personal advice; 87% said it helped them track their progress. Faculty reported shifting to more active, higher-level pedagogic strategies. There is now discussion of exporting the course outside the university, as well as expanding the use of adaptive-learning courseware at Georgia State.

The American Government course was built during 2017-2018 and is being piloted in Fall 2018 with one large section course. Full implementation is planned for Spring 2019. An academic study is underway in conjunction with this rollout. However, our proposal does not center on the results of our shift to adaptive learning courseware, but rather on what it takes to build a course using adaptive learning courseware. We propose a “view from the front lines” on the challenges one might expect if they take on this exciting, relatively new, digital pathway. Both courses required real teamwork between an outside vendor (Realizeit), faculty members, and many others to pull together full course offerings. Multiple diverse resources needed to be identified and integrated, such as outside vendor texts/readings, significant open educational resources (OER; videos, PowerPoints, articles, etc.), original content, and a unique data-literacy tool developed at Georgia State (http://rcii.gsu.edu/). Hundreds of diverse questions had to be written and assignments developed, all aligned to designated learning outcomes. Extensive copy editing and beta testing were essential. The courses had to be amenable to specific instructor needs, including instructors being able to add additional materials and sequence the modules to their liking. Faculty and graduate students needed to be trained (almost all instructors were unfamiliar with adaptive learning). Student support systems needed to be developed. The Realizeit courseware had to be integrated into GSU’s learning management system (LMS), with course enrollment immediately opening the courseware within the student’s LMS interface. Online and face-to-face versions were created, each with their own unique challenges. As the pilots and full-scale implementation progressed, numerous unforeseen issues have had to be tackled.

We propose an educate and reflect session where we present our experiences, thoughts, and results tied to building adaptive learning courses. In addition to the three professors presenting, we would be joined by Ashley O’Conner from Realizeit, who can offer the vendor’s perspective. For the 10-minute Q and A, we would like to have four different subgroups, where attendees can connect with one of us individually. Ten minutes is a short period of time, so a more personalized interaction may have deeper impact. Of course, all of us can be available for further, post-session discussion. We can also allow for temporary access to our courses so attendees can take them for “test drives” post-session.