Improved Analytics Critical to the Personalization of Online LearningMichael Horn | Co-Founder and Distinguished Fellow, Clayton Christensen Institute for Disruptive Innovation
The evolution of technology and technological tools over recent years has positively impacted the effectiveness of online learning, which has transformed into a highly engaging, highly integrated platform for students to pursue postsecondary credentials with maximum flexibility. Of course, as with any technology, there is still room for improvement and growth. Online learning has the space to become even more personalized. In this interview, Michael Horn discusses the current state of personalization in the online learning space and shares his thoughts on what the future might hold for online education.
The EvoLLLution (Evo): How truly personalized is online programming today?
Michael Horn (MH): Online learning today is personalized in the sense that it starts to give students control over the pace of their learning and the time when it occurs. It can offer much more flexibility given the asynchronous technologies.
Where there is still a lack of personalization is in the different pathways that students take towards mastery. Certain programs are certainly addressing this and we’re seeing adaptive learning engines like Knewton appear to do some exciting things to better target and personalize for different students. It still feels like we’re really in the early beginnings of the dramatic revolution that we’ve seen in a lot of other technology sectors where really smart recommendation engines come in and assist the student in picking and choosing their unique path.
Evo: What are the most significant limits on the amount of personalization and adaptability that can be introduced into an online course?
MH: In order to really go towards adaptive learning, you need huge numbers of students on your platform and there aren’t a lot of platforms that have that. If you think about it, the ability of Google to personalize advertisements for you, or Amazon to personalize shopping recommendations or Netflix to personalize movies, those are still relatively rudimentary themselves. There’s a limit to these engines and when you talk about learning. What’s exciting is there are potentially a lot more data points available. Every few minutes you can be having interactions that help you understand what a student does and doesn’t understand.
It’s a more complex problem to collect all that data and there are many more variables affecting it. There are also a lot of policies and regulations in place that potentially prohibit the data that we can use to improve this problem quickly. These regulations can inhibit what data we can collect and how we can use it to create the best learning experience at the right time for students.
Evo: Why is personalization of programming so important in terms of student success and outcomes?
MH: It’s related to the value tutoring has to the learning experience. There’s a great deal of evidence that tutoring is actually the best learning opportunity. A tutor can constantly see where a student lacks a certain understanding, or doesn’t quite have the background knowledge about something, and then tweak and tailor the approach and try different things to personalize it for that student. The fundamental insight is that learning is really based on a few things. One is that people are motivated by and passionate about different things. Secondly, we all have different amounts of knowledge that we can manipulate in active memory. Additionally, we all have different levels of background knowledge when we enter a learning experience.
Personalization along those dimensions is critical to unlocking student success.
Evo: To your mind, what might personalization of online courses look like in 20 years?
MH: What we’re going to learn over time is to get much more specific about what sorts of differences there are between learners and which ones have the most impact on learning and learning outcomes.
Right now people are unsure how much adaptive capacity we want in our learning compared with student control and agency. In the future—much as with Google or Amazon where the user has a lot of control but the engine is also automating and making suggestions to enhance that control—you’re going to see similar marriages in learning. You’re going to see a range of approaches for students, where some students go through game-based learning where they’re going through some really exciting simulations or games to master something, and other students they’ll just read a text because for them, they have the background knowledge to access it and it will be a more efficient way to learn.
Evo: What are the biggest roadblocks to realizing this vision of the future?
MH: We need platforms that can collect the data we need and can make better use of data so that we can figure out different ways to serve different learners. We also need to pay a lot of attention to the learning models themselves. If we use an adaptive platform like Knewton in a traditional classroom, it actually won’t be that useful because the teacher and students are not going at different paces. We need to really shift the learning models themselves, put students at the center and change the way they interact with educators. Competency-based learning is a really important ingredient to this and together, these will be the things that need to fall into place for this to really have the impact that it could.
Evo: As personalization and online courses evolves and grows, do you see the possibility of a higher ed environment emerging where there is no “traditional higher education” left but instead a range of hybrid models?
MH: That’s exactly right. Online learning serves certain people well but if you imagine online learning more as a platform that helps students and teachers find the right path forward in any given subject—whether that’s offline or online—then you can imagine this pervading every single learning experience in the future.
The only places where it might not take root are in those truly specialized subjects for which there are only a few students and teachers capable or interested in studying.
Evo: Is there anything else you’d like to add about the transformation and evolution of personalized online learning and what it will take to create the sort of learning landscape that we’ve been talking about today?
MH: Certain people have cast big doubts in the last couple years on the wisdom of personalized learning. One of the reasons you see pushback on that notion is because there are actually many different definitions for what personalized learning even is.
I have a relatively simple one: it’s the right approach for the right student at the right time. Nothing more nothing less. If people step back from it a little bit and see it a little more simply, it’s easier to understand the power that personalized learning can and should have for all students and realize that it’s something that we would all want in our learning.
This interview has been edited for length.
– – – –
- Personalized learning should be defined as the right educational approach for the right student at the right time.
- The growth of personalized learning is highly dependent on the capacity of institutions to improve the collection and analysis of learning data.
- True personalization of online learning will create an educational process similar to what tutoring provides now.
Author Perspective: Analyst