In my last post, I talked about the need for educators in general and faculty in particular to develop literacy around data and analytics. But it’s really broader than that. Back when college was intended for a relatively small percentage of the population, the idea of “weeding out” students who couldn’t make it without help was not obviously out of alignment with its mission. Now that the mission of higher education is more about educating everyone to the benefit of everyone, having skills to help students achieve their potential should be a core competency of every teaching faculty member. That includes learning some very old skills that K12 teachers have known for ages. It also includes learning new skills that will support the growth and application of the nascent learning sciences. When I made an analogy to 19th-Century medicine in the last post, that wasn’t casual. Where we are in 21st-Century learning sciences bears a striking resemblance. We are discovering some important basic science but are barely beginning the process of figuring out how those discoveries should influence our practice. Advancing more quickly on that front will require more than a handful of data scientists working in labs. It will require the proliferation of skilled educator/clinicians who can help advance our understanding of the ways in which theory interacts with the real world of practice.
This transformation of academia, at both the individual and institutional levels, must be driven by educators, not by vendors. But vendors can serve critical support and enabling functions. In our weird role as paid and unpaid marriage counselors between universities and vendors, Phil and I get the privilege of seeing how both sides of this potential partnership are grappling with these changes both separately and in partnership. I believe that many of the most successful ed tech companies of the next decade will be the ones that figure out how to support educators in making this professional transformation. There will be many different ways to do this. I’m going to write about one example of a company grappling with this challenge today and will write about more in the future.
Like all of the major textbook publishers, McGraw-Hill Education (MHE) is grappling with the existential crisis that their historic core business—selling content in the form of paper books—is it is in the process of becoming commoditized.(Full disclosure: The company is also currently a client of ours.) The publishers need a new gig to survive. I wrote about this general problem, and how publishers are going to need to find a way to be more loosely coupled to their content going forward, in a blog post a while back. Since then, MHE’s announcements at EDUCAUSE included a project in which Georgia Tech is using MHE’s adaptive platform to deliver the university’s computer science content and another project with the Cleveland Clinic using MHE’s same platform to deliver a mix of MHE and Cleveland Clinic content to nurse practitioner and phsysican’s assistant students. These are indicators of the company’s long-term direction. In the future, MHE will not be a content company per se. Content will be important but not always the primary differentiator. Instead, as I’ve heard CEO David Levin and other senior MHE executives say on multiple occasions, they want to be a “learning science” company.
What does that mean?
I don’t think we know yet. But we get some clues from an interview I did with some of their senior digital and learning science folks last spring as part of our ongoing e-Literate TV work exploring different aspects of “personalized learning” and “courseware.” You can find the whole series here. (Reminder: In the intervening time since these videos were recorded, MHE has become a consulting client of ours.) We wanted to capture some of the sausage making behind the design of these products so that educators could get a better sense of the thought that goes into them. We also wanted to capture the vendor’s thinking about their relationship to the educators who adopt their products and to the learning sciences.
Here they are talking about they design adaptive learning products:
- There are theories about learning, but they need to be tested in educational practice.
- (Those theories can be explained in terms that a layperson can understand, as Al Essa does here with memory decay.)
- The machine tutor can make recommendations that are wrong. Both theories and applications have to be constantly tested to make the implementation more effective.
- MHE uses human content experts to review how well the machines are doing at tutoring and to tune and improve them over time.
- In MHE’s view, an integral part of developing an effective adaptive learning system is human curation.
Notice that the company isn’t starting with some story about “big data” and how the machine is going to mysteriously extract everything we need to know about learning from the clickstream. They are starting with learning theories—theories about humans rather than about machines—and then using the machines to test those theories.
Next, we talk a little more about what “learning science” means to them:
Some details of their position:
- They take seriously Benjamin Bloom’s “two sigma” research, in which he showed students who are taught via mastering learning techniques to improve by “one standard deviation” and those who receive one-on-one tutoring improved by “two standard deviations.” (Think of a “standard deviation” as one course grade in this case, e.g., going from a D to a C or a C to B course grade.)
- Preliminary peer-reviewed students show a one sigma improvements under certain circumstances. More research is needed to validate these findings.
- A one sigma result requires both the software and a change in pedagogy to take advantage of the software’s benefits.
That last point is interesting because it highlights the vendors’ interdependency on educator/clinicians. The educational product “efficacy” frame that Pearson has popularized is based on a pharmaceutical analogy. But it is becoming increasingly clear to the industry that the drugs aren’t all that helpful without skilled educator/clinicians who can perform a proper diagnosis and develop a comprehensive treatment plan. That requires skills that many instructors in higher ed haven’t learned. Getting the impact on student outcomes that these products purport to offer will require some kind of formal or informal professional development. Some of that can be accomplished through cues and tutorials in the software itself. But some of it can’t.
In the last segment, I ask what it would take for educators to actually become authoring of adaptive learning content:
The main points:
- Faculty can learn how to author adaptive learning content.
- But they need to be trained. If they don’t have a background in learning theory—as most faculty don’t—then they will need to learn some or get help from somebody who has the training.
- MHE provides support services to help faculty learn how to author the content.
- MHE is also banking on the fact that more and more colleges and universities have instructional design support staff who can help the faculty with the course design work.
I got a demo of MHE’s authoring platform, which I captured on video but which didn’t make it into this particular series. It is thoughtfully designed but necessarily complex. The learning curve is non-trivial and would require signficant professional development for faculty in many cases, particularly if they are not able to work with trained course designers that they trust. It’s manageable. But it would require commitment. Especially if you want the faculty to understand how the system works well enough to know what it’s good for, when to trust it, and how to tune it. It’s not just software training. It’s a literacy problem.
Stepping back and looking at the whole interview series (as well as other signals MHE is sending in the market), it’s clear that the company is moving away from thinking about content as the final product and toward the idea that content is constantly being tuned as part of a feedback loop that good educators and good scientists both use to learn and improve. Bringing to market an authoring platform and wrapping coaching and design services around it suggest that MHE understands that it has to help foster the proliferation of educator/clinicians in order to get out of the dead tree business. The fact that the authoring product and support services still appear to be in the soft launch phase indicate that they are in the early stages of figuring out what that even means.