Predicting the Future

I’m a person that generally doesn’t spend much time predicting the future. I’m more comfortable trying to imagine the possible futures I find desirable, and that’s mostly what I do on this blog, talk about the futures we should strive for.

But two and a half years ago, at the encouragement of the folks at e-Literate, and with the world just coming out of its xMOOC binge, I made some predictions about the future of edtech for e-Literate. I decided to put aside my 10 year visions of the desirable, and just straight up predict what would actually happen.

I spent about a week thinking through all the stuff I talk about and trying to be brutally honest with myself about the future of each item. I literally had a pad where I crossed out most of my most beloved futures. Most things I loved were revealed to be untenable in the short term, due to the structure of the market, the required skills, cultural mismatches, or the lack of a business model.

It was immensely painful. Still, when I was done, a few things survived. They weren’t like most people’s predictions of the time, and in fact ran against most of the major narratives in play as of December 2013.

Here were the predictions. I made three firm predictions under the title “Good Opportunities That Will Be Taken Seriously by the Powers That Be”. I’ll put aside one of these “Local Online”, as I noted even at the time it was a bit of a cheat: local online was a transition that had already happened; it was just no one had noticed.

I’ll deal first with my two other major predictions, which ran counter to the narratives of the time.

  • At a time when asynchronous learning was king, I predicted the rebirth of synchronous learning.
  • At a time when Big Data was the rage, I predicted the rise of Small Data.

How’d I do?

Synchronous Online

In a time when the focus was on asynchronous and self-paced learning, I predicted a renaissance of synchronous learning:

Synchronous online is largely dismissed — the sexy stuff is all in programmed, individuated learning these days, and individuated is culturally identified with asynchronous. That’s a mistake.

I went on to describe how the emergence of new tools based on APIs like WebRTC would make possible the merging of traditional synchronous learning sessions with active learning pedagogies, and how this would result in a fast-growing market, as it would address the needs of a huge existing population of students currently underserved. I compared the market for videoconferencing products to where the market was for the LMS on the eve of Canvas entering it: people believed the LMS wars were over, but in fact they had just begun, because Blackboard had treated the LMS as a corporate tool rather than an educational one:

Adobe Connect and Blackboard Collaborate are, I think, in a similar place. They are perfect tools for sales presentations, but they remain education-illiterate products. They don’t help structure interaction in helpful ways. I sincerely doubt that either product has ever asked for the input of experts on classroom discussion on how net affordances might be used to produce better educational interaction, and I doubt there’s all that much more teacher input into the products either. The first people to bother to talk to experts like Stephen Brookfield on what makes good discussion work *pedagogically* and implement net-based features based on that input are going to have a different pitch, a redefined market, and the potential to make a lot of money. For this reason, I suspect we’ll see increasing entrants into this space and increasing prominence of their offerings.

Suggested tag line: “We built research-driven video conferencing built for educators, and that is sadly revolutionary.”

I don’t know if you can remember how unpopular synchronous was in January 2014, but contemporary takes on it ranked it somewhere between Nickelback and Creed as far as coolness.

So where are we today? Well, WebRTC is propelling a billion dollar industry. Blackboard Collaborate got its first refresh in a decade in 2015 (based on a WebRTC purchase they made in November 2014). Minerva, the alt-education darling, released its platform later that year, which was based on synchronous video learning.

And today, we find an extended article in the Chronicle about the surprising new trend in online education: the rebirth of synchronous education, the hottest trend in learning right now. The reasons for it?

What’s giving rise to the renewed interest in more-formalized synchronous courses is that the technology for “high-touch experiences” in real time is getting more sophisticated, says Karen L. Pedersen, chief knowledge officer at the Online Learning Consortium, a nonprofit training and education group. Institutions are catching up to their professors, and tools are now widely available that let professors share whiteboards simultaneously or collect comments and on-the-spot poll results in real time.

The article goes on to explain that the recent ability of tools has paired traditional synchronous classes with active learning, which makes the difference.

I have some ambivalence on where this will go, as mentioned in the intro to this post, these were predictions, not my top desired futures. Opportunities. And opportunities can be perverted. But this was surprisingly on target.

Small Data

At the height of Big Data madness, I predicted the rise of small data products:

Big Data is data so big you can’t fit it in a commercial database. Small Data is the amount of data so small you can fit it in a spreadsheet. Big Data may indeed be the future. But Small Data is the revolution in progress.

Why? Because the the two people most able to affect education in any given scenario are the student and the teacher. And the information they need to make decisions has to be grokable to them, and fit with their understanding of the universe.

Small Data was a relatively new term at the time the prediction was made. The Wikipedia page for the term was actually birthed on January 2, 2014, about the same time I was writing the post, and looking back now I only see a smattering of uses of the term in 2013. I was at the time reading the wonderful critiques of Big Data by writers like Michael Feldstein and Audrey Watters and thinking through the question, if not “Big Data” then what?

Then in Spring of 2013 I saw a presentation by the local school district on their use of data. The head of their operation said the most useful data for them had been the “One-F” test. They would just compile the grades of the students in all their classes and look for students that had an F in one subject but A’s and B’s in others. Then they’d go to the student and say — look, you obviously can do the work in other classes, what’s happening here? And they’d go to the teacher and say hey, did you know this student is an A student in their other classes — what is going wrong in this class?

And the reason why it worked, they said, was you could talk about standard deviations or point-biserial correlations all day, but it would never make sense to the people whose actions had to change. But people could understand the “One-F” metric. It wasn’t a p-valued research finding: instead it was a clue, understandable by both teacher and student, that something needed investigating and a bit of guidance on where the problem might be, and how to address it. And that — not research level precision on average behavior — was where the value was.

And so it was really Lisa Greseth, the IT head of Vancouver Public Schools at the time, who showed me the way on this. “Small Data” seemed to encompass this idea — it was theory-informed data collection. It was data as a clue for action. And most importantly, it was data that is meant to be understood, in its raw form, by the students and teachers involved.

How’d this prediction go? Pretty well. In the two and a half years since there’s been an explosion of interest in small data. Here’s the first eight results from a Google search on “small data education”:

The Washington Post. May 9, 2016: ‘Big data’ was supposed to fix education. It didn’t. It’s time for ‘small data’

EdWeek. May 16, 2016: Can Small Data Improve K-12 Education? –

InformationWeek. Nov 24, 2015 – McGraw-Hill Education’s chief digital officer has driven the company’s effort to leverage small data to improve student outcomes.

Helix Education. Oct 22, 2015: Big and Small Data are Key to Retention

Portland Community College. Mar 9, 2015: Distance Education: Using small data

Pando Daily. March 9, 2014: The power of small data in education

Center for Digital Education. Jul 1, 2015: 7 Benefits of Using Small Data In K-12 Schools

Times Higher Education Journal. Jul 1, 2015: The Power of Small Data

The prediction, of course, was about the introduction of “small data products”, and there’s been growth there too. McGraw-Hill, for example, is pushing a small-data focus in its Connect Insight series. In many ways, this is a return to a data focus that existed before Big Data madness, a focus on small, teacher-grokable data points collected for a specific purpose. And though McGraw-Hill calls it “Small Data” explicitly, it is the direction that most products seem to be re-exploring after the crash of Big Data hype.

By the way, I still believe Big Data has a place, applied to the right problems. It just wasn’t the place people were predicting two and a half years ago. Maybe I’ll save thoughts about that for a future prediction post.

Other Predictions

I had a category for things that I thought would develop but mostly remain under the radar, and not see broad institutional adoption. I put the return of Education 2.0 (blogs, wikis, etc) in there as well as “privacy products”. I think I was more or less right on those issues. In Education 2.0 we’ve seen real growth, particularly with Reclaim Hosting’s efforts, but it’s still off the institutional mainstream for the moment. On privacy products there has been less development than even I thought there would be, though the recent development of the Brave browser and increasing use of personal VPN provide some useful data points.

I did make the brave, and completely wrong, prediction that Facebook had peaked, thinking that many of its features could be supplanted by OS-level notification systems. Looking back on this prediction I learned something about making predictions: don’t make predictions about things you don’t use, at least not without observing closely how other people use them. My use of Facebook at that time was limited to a quasi-monthly visit.

So lesson learned there? In the time since that I’ve worked on Wikity and Federated Wiki, I’ve come to a greater understanding of what Facebook provides people, almost invisibly. And I have to say, paired with my prediction from 2014, it has really demonstrated to me that what a lot of people build to “replace Facebook” (including things I build) don’t really replace what Facebook provides people. If you look at Facebook and the rise of Slack you start to realize that maybe centralized control of these platforms is key to the sorts of experiences people crave. It may be that you can’t make a federated Facebook anymore than you can make an alcohol-free bar.

I’m not saying that many things can’t be federated. But I have a new appreciation for why they aren’t. (And, as expected, it’s probably this failure of prediction that is most useful to me at this point).

Anti-Predictions

Finally I made some anti-predictions about hyped trends of the time that I believed would go nowhere. Here I predicted that Gamification and Mobile Learning would crash and burn.

I turned out to be largely correct. Gamification seems to be entering its death throes, as it is really just rehashed behaviorism, with the dubious distinction of being even less self-reflective than behaviorism. (The “good” parts of “gamification” are really just learning theory — scaffolding, feedback, and spiral designs come from Vygotsky, Bruner, and others, not Atari).

More interestingly, my prediction about mobile came out more correct than I imagined. As predicted, we’ve gone through the iPad optimism of 2013 and 2014 to find that, unsurprisingly, learning and creating are not really mobile endeavors. Deep learning, it turns out, tends to be an activity that benefits from focused time and rich input systems. (We tried to tell you). So as we watch the iPad one-to-one programs crash and burn, let me revise my previous claim that Education Analysts Have Predicted Seven of the Last Zero Mobile Revolutions.

They’ve now predicted eight of them.

Conclusion

I don’t know. I feel like this is a pretty good record. The Facebook prediction was arrogant and misplaced. I am seriously contemplating that error at this point, hoping for some insight there.

Most of the rest of the predictions were arrogant as well, but came true anyway.

What was behind the right and wrong predictions? There’s no overall trend, but the Facebook failure is instructive when put next to the other predictions.

The key in all these things is to try to truly understand where the value in the current system is, as well as what the true pain points are. And the key is to imagine technological solutions that that address the true pain points without taking away the existing value of the system.

  • Synchronous Online manages to preserve valuable elements of synchronous learning while addressing its main problem: feelings of isolation and disengagement.
  • Small Data builds on the strengths of a system built on the intuitions of the teacher, instead of the data analyst, and works backwards from their needs as a consumer of data.

Things that don’t take off tend to misunderstand central features as flaws. The iPad misunderstood the rich input systems of the laptop as a hindrance rather than a benefit. And its “benefit” of being a “personal” device didn’t map to a classroom where devices weren’t personal, but constantly swapped between students and classes.

Likewise, the centralization of Facebook turns out to be one of its great features: people are actually craving more filters, not less, for the information they consume, and they’d prefer to stay in a standard environment rather than venture out onto the web for most things. Plus, in the two and a half years since I wrote this we’ve seen what has happened to the notifications panel on phones: it’s a Tragedy of the Commons. With every app now pushing messages into the notifications panel, I can’t go to it without finding it littered with thirty or forty ridiculously mundane “updates” from 18 different apps, all cloying for my attention. Facebook’s centralized, privatized ownership of its newsfeed allows it to reduce noise in a way that federated systems have trouble doing.

The biggest blindspot tends to be our own experience. I was able to see the mobile mismatch, because it matches my own experience as a learner. I couldn’t see the strength of Facebook because I don’t *want* the world to like the things about Facebook that it so obviously likes, and I never should have predicted anything about it until I understood its present value to people.

On a personal note, going back through this reminds me that I should probably try to predict more. My tendency is towards futurism, unfettered by reality, and I remember how painful the process of trying to truly predict things was. But truth is, if you can dredge up some ruthless honesty, you can see what the likely routes forward are. That’s not quite as fun as advocating what should be, but it’s probably a useful skill to develop.

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s