Published on 2014/09/18

How Adaptive Learning Can Make Higher Ed More Customized and Effective (Part 2)

Co-Written with Karen Vignare | Associate Provost, University of Maryland University College

Adaptive learning can help students learn at their own pace, by targeting areas they need improvement in and quickly moving beyond areas they’ve mastered.

This is the conclusion of Beth Mulherin and Karen Vignare’s two-part series on adaptive learning. In the first part, Mulherin and Vignare explored the theory supporting adaptive learning and explained how the technology could be of value to today’s highly-discerning student. In this conclusion, they will share a few examples of how adaptive learning has been practiced to great effect at institutions across the United States.

New adaptive learning models and supporting software are emerging that offer a variety of approaches for learner support, and they’re showing positive results. Online learning resources such as the Pearson MyLabs, powered by Knewton adaptive technology, provide students with a customized pathway for learning content based on an entering diagnostic and provide faculty with dashboards of student progress. A Knewton and Arizona State University partnership increased math pass rates by 18 percent and withdrawals dropped by more than 47 percent. Rio Salado College is also partnering with Pearson and Knewton to develop a personalized learning pilot based on Pearson’s MyWritingLab, and several community colleges have integrated MyWritingLab into their developmental writing courses and seen improved success rates.

Carnegie Mellon University’s Open Learning Initiative (OLI) is pioneering cognitive tutoring models in online courses based on learning science research. Research such as Bloom’s 2 Sigma findings (1984), which showed a significant difference with student outcomes in an instructional environment that combined tutoring, formative assessment and feedback, and more recent research on “intelligent tutors” indicates they’re nearly as effective as one-to-one human tutors (Van Lehn, 2011).

At University of Maryland University College (UMUC), we successfully piloted the OLI in several undergraduate courses. In an undergraduate statistics course with one instructor teaching two sections of the same course, the percentage of A’s and B’s for the OLI section was 59.38 compared to 41.67 for the control group. In a biology course, there was a significant positive correlation to the percentage of students who achieved an A or B grade and the percentage of activity completed on low-stakes assessments. End-of-course evaluations for all sections using the OLI platform are consistently higher than the average evaluation scores for all sections of the courses.

UMUC is actively investigating and evaluating the incorporation of adaptive learning tools into our academic model. Our new Center for Innovation in Learning and Student Success unit is at the forefront of developing and evaluating innovative learning and academic support models. We’re currently reviewing adaptive learning tools for piloting in early 2015, and the hoped-for outcomes will be to decrease the drop, withdrawal and failure rates of students in undergraduate first courses. We will be piloting some of the most promising adaptive learning software, basing our tool selection on features that meet the needs of our students and curriculum.

An adaptive learning platform provides a more focused approach for students to learn without forcing students to relearn concepts they’ve already mastered.  It also provides a learning science approach to course design; over time, the data collected by the adaptive learning software can be used to improve course design. And it allows the university and students themselves to identify the most common concepts for which remediation and reinforcement are needed, information that will be used to improve instruction and support models.

– – – –

References

Bloom, B. (1984). The 2 Sigma problem: The search for methods of group instruction as effective as one-to-one tutoring. Educational Researcher(13)6, 4-16.  Retrieved from http://web.mit.edu/5.95/readings/bloom-two-sigma.pdf

Smith, P., ed. (2014). MyLab and Mastering humanities and Social Sciences: Efficacy and implementation and results. Retrieved from http://www.pearsonmylabandmastering.com/northamerica/results/files/HSSLE_EfficacyResults_2014.pdf

Thompson, J. (2013). Types of adaptive learning. Retrieved from http://www.cogbooks.com/white-papers-adaptive.html

Print Friendly
Non-traditional-eBook-V

Readers Comments

R. Jenkins 2014/09/18 at 3:29 pm

What really struck me is that research on the value of tutoring has been widely shared and available since the early 1980s – 30 YEARS AGO! Why are these concepts not common-practice? People complain about the price of implementing these kinds of changes, but look where we are now. Maybe if we had paid more attention to completion in the past, we wouldn’t be staring down the barrel of performance-based funding today.

    Brandon Emerson 2014/09/18 at 4:50 pm

    Or people not even thinking higher ed is worthwhile anymore.

Peter Farris 2014/09/19 at 10:17 am

I’ll be interested to find out how successful this model is a year after launching. It would be interesting to see what kinds of kinks had to get rolled out in the early stages

Wallace Kenyeres 2014/09/19 at 10:22 am

I’m interested to see the way UMUC and other institutions continue to integrate adaptive learning tools and processes into their programming, particularly for non-traditional students. I think this effort will lead to a necessary discussion on the future role of course instructors. We can’t leave it to software alone to identify struggling students and potential interventions. However, I’m not sure what exactly the link would be between adaptive learning software and an instructor; at which point in a student’s journey is software still appropriate and where’s the line where an instructor should step in?

Gen Zhao 2014/09/19 at 10:32 am

For me, the most exciting part about adaptive learning software is its ability to capture data that can be used down the line to improve programming. For instance, if students consistently struggle in one topic of the course, perhaps future iterations would have the instructor devote more time to the topic or introduce additional learning aids/materials. The data could also tell us about how to scale a program or course, leading to more effective program design. I hope more institutions take advantage of the possibilities of adaptive learning tools, not just for today’s student but for future ones.

Leave a Reply

Your email address will not be published. Required fields are marked *

[if lte IE 8]
[if lte IE 8]