Scaling Up With Adaptive Learning

Two universities share how they are using adaptive courseware to increase retention and graduation rates — and the challenges they have run into along the way.

Last year, eight universities across the country embarked on a bold experiment to see if they can scale up the use of adaptive courseware to increase retention and graduation rates. With support from the Association of Public Land-grant Universities (APLU), these schools have set a target of using adaptive courseware for 15 to 20 percent of general education course enrollments between spring 2017 and fall 2019. APLU's Personalized Learning Consortium is overseeing the grant program, which is funded by the Bill & Melinda Gates Foundation.

The universities — Arizona State University, Colorado State University, Georgia State University, Northern Arizona University, Oregon State University, Portland State University, University of Louisville and the University of Mississippi — certainly have their work cut out for them. Tailoring a course to adaptive learning can be difficult, as we learned when Campus Technology interviewed faculty members involved in pilot implementations of the technology. Despite being encouraged by the results, those pioneers reported being exhausted by how much work is involved in retrofitting their courses to the adaptive platforms. So how will the APLU grantees tackle the task? We spoke with two executives participating in the grant program about their goals and the early challenges they see.

Focusing on High-Enrollment Courses in Oregon

"The reason we are interested in adaptive courseware is to improve student learning and student success. The adaptive courseware is just one approach to help us increase retention and graduation rates," said Julie Greenwood, associate dean for undergraduate studies at Oregon State University.

OSU is focusing its initial efforts on eight high-enrollment courses that also have high attrition rates, in fields such as college algebra, psychology and chemistry. Greenwood said some OSU faculty members have already experimented with adaptive courseware, but not in an extensive fashion. "It is being used for homework or extra credit," she said, "but the technology is challenging enough that they are not integrating it into the course and not taking full advantage of the analytics."

Greenwood pointed out that the adaptive courseware could help faculty members shift content delivery from lectures to a blended format, where students view course materials online before coming into class — and instructors tap into the data on their progress. "A faculty member can come into the classroom with 200 students and know that a majority struggled with three particular problems and decide to focus on those with their class time," she said. "That is the value of the analytics."

APLU has given its grantees a list of 20 approved adaptive courseware vendors to choose from, and OSU is seeking to develop a model and structure that will allow the university to support faculty members in implementing the courseware that fits the curriculum they teach. The math department, for instance, is planning to do an in-depth evaluation process, starting with development of the metrics and rubrics it will use to make that decision. "We would like to see flexibility from the vendors for the instructors to be able to provide their own content," Greenwood added.

Because of the additional time and effort involved in implementing adaptive learning, one of the challenges is how to motivate faculty members to participate. Schools can offer stipends or other faculty incentives, but timing can be an issue, Greenwood said. "We had one faculty member who had just spent five years redesigning her curriculum. This was not a good time to ask her to throw all that away and start over again."

A positive faculty experience is going to be critical to OSU's ability to scale up. "We have to provide structure and support to help the faculty deal with the technology," Greenwood said. "We want them to focus on curriculum and instruction, not to have to figure out how to get this to work in Canvas, how to access the analytics or worry about students not being able to log on. We have to provide a structure here at the university that eliminates those barriers, so this can run smoothly. We also need a positive student experience. We need to measure the student experience as it goes along, respecting their feedback and monitoring their success."

Going University-wide in Arizona

Northern Arizona University has an even broader mission than the APLU grant entails. "Instead of approaching just general education, high-enrollment courses, I think it is time for a university-wide culture of adaptive learning," said Don Carter, NAU's director of e-learning. Through a combination of blending and flipping classes, the university has redesigned 35 courses over the past five years, and Carter said data shows DFW rates (D-grades, F-grades or withdrawals) have decreased from 23 percent to 19 percent.

In addition to the high-enrollment courses, NAU is looking for courses in areas such as the graduate nursing school or higher-level chemistry that may benefit from adaptive courseware. Of the 20 APLU-approved vendors, Carter has identified five that he thinks are the most promising, and NAU instructional designers and faculty are reviewing their offerings.

"We are going to approach things in three ways: One is a full course redesign with us adopting the offering from a vendor as a full textbook replacement," he said. (NAU's pre-calculus course is already based on Pearson's MyLabsPlus.) The second approach involves finding smaller building blocks of pre-built modules about particular topics or troubling content. "The third approach is adopting a platform in which my instructional design team and faculty can build their own modules," Carter said. One discipline that may try that option is foreign languages, because the vendors have almost no material in that area, he added.

NAU has engaged faculty members in a course redesign effort called "intentional curricular design." But Carter said designing for adaptive technology takes even more work. "Once you add technology, you have to design down to the concept level. You have to sequence and connect things. We are talking about hundreds of minutiae if they want to build their own." Do faculty members realize what a time commitment that is? "Those who haven't been involved before may want to try it, but I don't think it will be very long before they realize how hard it is," Carter said. "We don't have enough money on the grant to pay very many people to build their own [modules]. The approach of the grant program is to buy off-the-shelf things. We are not allowed to develop our own technology system, but we can put our own content in and use someone else's platform. But it is going to be problematic to do very much of that."

Both Carter and Greenwood said the degree of difficulty is also raised by the fact that the vendors themselves are all new to adaptive learning. The best offerings are created in close collaboration with faculty members, they noted.

Carter said the eight universities participating in the APLU grant program are sharing procedures, protocols and results and will try to do some cross-institutional studies. They also plan to sustain their efforts beyond the grant program. "We were a strong candidate for the grant program because we were already doing this," he added. "We had no intention of taking the money and running. We are using it to give us the next-level quantum leap into doing this and we are going to continue."

Adaptive Learning in the Humanities

Most software vendors developing adaptive courseware have so far focused on math and science curricula first before turning to any humanities courses. But Introductory English Composition is another general education course universities want to address, so the Association of Public Land-grant Universities is supporting four universities in a collaborative project to develop English composition adaptive courseware. Georgia State University, Montclair State University (NJ), the University of Georgia and the University of Mississippi are using the Smart Author platform from Acrobatiq to co-develop courseware and test it out at all four campuses.

Robert Cummings, associate professor of English and director of the Department for Writing and Rhetoric at the University of Mississippi, said the term adaptive has sometimes been construed as checking student answers for how they are right or wrong. "In content courses, the ways that things can go wrong are limited or predictable. But of course, the ways things can go wrong with writing are perhaps unlimited. But that doesn't mean that adaptive doesn't work in the composition classroom."

In earlier work with the Open Learning Initiative at Carnegie Mellon University, Cummings said, electronic courseware developers at Ole Miss learned that they had to tighten up the English composition course space and narrow down the tasks students were asked to do in a writing context. "We have to allow students to focus on one task at a time, and then try to assemble the writing they had already done into more coherence. We always have taught writing as a process. This gave us a much more structured approach to that."

The campuses broke up an English composition course into four units and the universities developed one unit each. Some had faculty members teach with the new curriculum in the summer and others started this fall.

"We are always trying to think of ways to help our weakest writers scale up quickly and efficiently, including technologies to accelerate and personalize the learning experience for student writers," said Emily Isaacs, associate dean of the College of Humanities and Social Sciences at Montclair State. "Each of the four schools in the grant project had been picking at this difficult challenge in our own ways on our own campuses for a while," she explained. "I like the idea intellectually that you can do this with any discipline, but it is very hard. The theory is very easy; the practice is very, very difficult. Sometimes we can overestimate the capacities of the technology." A big piece of the puzzle is that the adaptive platform itself is still in development at the same time that the faculty members are developing the ideas, she said.

Most of writing is doing, not knowledge attainment, she stressed. "The a-ha moment does not come from reading about what an introduction should have in it." But she said the adaptive software does help instructors see which students are misunderstanding a concept such as counter-argument. "You either know you did a bad job in creating the quiz or you know that nobody gets it, and you have to spend more time on it. It can clue teachers in."

Both Cummings and Isaacs said the English composition courseware available on the market still leaves a lot to be desired. "I haven't seen everything out there in the adaptive space in English composition courseware," Cummings said, "but I haven't seen anything that I would consider to be truly adaptive so far."

Featured