What if the adventure chooses you?

I was reading those e-Literate guys, Michael Feldstein and Phil Hill, write about personalized learning at edSurge the other day. Oddly enough, I was writing my own little account of the same phenomenon for the book I’ve been working on at the same time. While what follows obviously owes a lot to Audrey Watters, I still thought I’d drop in the first draft of this section of Education Is Not an App for the sake of diversity of opinion and so that this blog doesn’t go completely dead before the book comes out. [I’ve hyperlinked the material that originally came from the web and omitted footnotes.]

ELIZA was a computer program written by Joseph Weizenbaum of MIT between 1964 and 1966 that simulated a psychotherapy session. Type a question into the program, and ELIZA would create another question based upon the language in the original question. For example, if you wrote something along the lines of “My mother is making me angry,” the computer might respond with a question like “Tell me more about your mother.” Despite ELIZA’s simplicity, people using it tended to quickly get entranced by the opportunity to talk about themselves. Weizenbaum himself was deeply concerned that people were being fooled into thinking that the machine actually cared about them.

Computers have gotten a lot more sophisticated since the 1960s. As a result, the kinds of conversations that they can carry out have gotten more sophisticated too. “The holy grail of learning is personalized or adaptive learning,” explained Anant Agarwal of edX in April 2015. “This form of learning is what you might experience from an excellent personal tutor who is able to tailor your individual experience. In many ways, adaptive learning can be compared with those old ‘Choose Your Own Adventure’ books. At each step in the learning process, the user is given multiple options that satisfy his or her level of comprehension, style or direction. They may all lead to the same place (mastery of the material). but the path can be very different and structured for a particular learner.” While Agarwal mentioned this in the context of edX’s MOOCs, there’s no reason that this technology has to be scaled to thousands of people at once. It would, at least in theory, work just as well with twenty students as it would with twenty thousand.

Of course, advocates of this kind of technology in the classroom will tell you that it is not intended to replace teachers. In an Inside Higher Education article about work being done at the University of Wisconsin – Madison on what researchers there have dubbed “machine teaching,” one of the principle investigators told the reporter “this will not minimize the teacher or faculty member role, but would help to optimize the teacher’s time, so he or she could spend the least amount of time necessary on a subject before every student fully understood it.” Unfortunately, the professors who develop this technology have no power over exactly how it gets employed – especially if it ever gets licensed by private companies. Arguing that automation will free up workers to concentrate more meaningful work is a common argument in Silicon Valley. As the author Nicholas Carr explains, “high-flown rhetoric about using technology to liberate workers of masks contempt for labor.”

Look at this situation from as administrator’s point of view. If they buy these expensive computer programs, where will they get the money to pay for them? At cash-strapped schools the inevitable justification will be because it can save labor costs. Computerized teachers, computerized scoring – these days computers will even tell students whether they’re on the most efficient path towards graduation, thereby eliminating the need for advisors. Sometimes it seems as if every aspect of modern universities that can be mechanized has been mechanized. Why would actual teaching be any different?

Making this kind of switch depends upon advocates of technology changing the definition of what education is. The classes where we attended college depended upon prolonged interaction between the instructor and the students. Even in the largest classes that we took, professors took questions before, during and after their lectures. If we were feeling shy, there was ample opportunity to visit professors or our teaching assistants in office hours to work through whatever problems we had in the material. Our papers were graded by human beings who explained why we earned the grades we did. It is through these kinds of exchanges that the most intense learning happens, the ones where habits of mind are set and where inclinations develop into skills that students can employ not just in other classes, but for the rest of their lives (long after they’ve forgotten what their undergraduate professors’ names happened to be).

Personalized learning, the pitch goes, allows professors to spend less time doing things that others can do better (like lecture), you can spend more time helping students learn. Unfortunately, like Lucy and Ethel in that chocolate factory back in the 1950s, it is easy for your employers to speed up your line by giving you more students – particularly if you work in an online setting where the size of the classroom is no longer a limiting factor. In this way, unbundling the professor’s job limits the contact between the student and their instructor. By limiting the contact between students and their instructor, unbundled classes have to focus on how much content a student remembers rather than the kinds of skills that they develop since testing for those skills requires more labor, not less. Anyone who really cares about teaching should consider that outcome a shame.

Posted by Jonathan Rees

Professor of History, Colorado State University - Pueblo.


[…] few weeks ago, Jonathan Rees wrote a post calling out that, no matter what potential of so-called “personalized learning” for […]

Fred M Beshears

When I worked at UC Berkeley (1987-2007), I tried to initiate a program that would see if it would be feasible to switch from the large lecture classroom to a small study group environment led by student tutors.

My concern was that the large lecture depersonalized education, whereas the small face-to-face study group made learning an integral part of the student’s social life. Using small study groups to improve learning outcomes has a long tradition at Berkeley that goes all the way back to Uri Treisman’s work in the early 1980s.

Improving the Performance of Minority Students in College-level Mathematics

Also, the idea of using videos of lectures in small study groups has been around since the early 1970s; it’s called Tutored Video Instruction. John Seely Brown praised the TVI approach in his famous book The Social Life of Information. Here’s an extended excerpt.

John Seely Brown on Tutored Video Instruction

Although I had received NSF funding for the project, the cooperation from a non-tenured professor who wanted to experiment with TVI and two study group programs on campus – the Minority Engineering Program and the Professional Development Program – the project was halted by a senior member of the Berkeley faculty.

It was not halted because he thought that TVI wouldn’t work. It was halted because he thought that TVI would work, and that it would undermine the tradition of the large lecture classroom at Berkeley. The history of this episode is described on my blog

The High Tech Small Study Group Saga

(BTW: by ‘high tech’ I simply mean supplying the existing small study group programs with computers that could run symbolic algebra software and, of course, video tapes of the lectures.)

The conservative professor who put an end to the program was OK with the idea of small study groups, but he was not OK with the idea of allowing the study group students to not attend live lectures. He wanted to preserve the practice of herding students into large lecture halls, which can seat hundreds of students.

One of the largest lecture halls at Berkeley can seat up to seven hundred students. We used to joke about it by asking: “How many rows back can a student sit before it becomes distance education.”

So, whenever you hear of faculty who warn that technology might be used to dehumanize higher education, remember the thwarted High Tech Small Study Project at UC Berkeley. And then, ask the individuals standing up for the status quo in Higher Education if they would be willing to institute a TVI experiment on their campus.


That’s a really interesting, complicated story. You can kind of see both sides if you look hard, but then by the time you get to the end you know doing nothing was probably the worst thing to do in the end.

While I know this is kind of evasive, the philosophy that I’ve worked out in these matters is that all students deserve a caring, well-qualified living breathing teacher who can monitor their progress and help them through the inevitable roadblocks on the way to useful learning. I’m willing to consider any technological set up that meets those criteria as something I’d endorse or even practice myself, but the Devil is always in the details.

Fred M Beshears

Professor Rees,

If your colleages in the professoriate agree with you that:

“all students deserve a caring, well-qualified living breathing teacher who can monitor their progress and help them through the inevitable roadblocks on the way to useful learning”

then why do many schools, especially research universities, continue to routinely herd students into large lecture courses?

At research universities, I suspect the answer is that research faculty would rather teach small graduate seminars that dovetail nicely with their research interests. That’s why community colleges with relatively modest budgets are able to afford smaller class sizes for basic introductory courses than is the case at the typical research university.

The point I wanted to make to the faculty at Berkeley was as follows: if you insist on dehumanizing undergraduate education with large lecture courses, then we should at least be willing look at more humanized alternatives such as TVI with small study groups.

Regardless of what we think students “deserve” a prioi, we should do controlled experiments to see how the learning outcomes of students in the TVI groups do relative to the students who attend a live lecture in one of our massive lecture halls.

Unfortunately, the faculty I had to deal with at the time were unwilling to look at those alternatives and were unwilling to conduct these experiments.

They were willing to use their authority as senior faculty to prevent the experiments from going forward. Even though they hold Enlightenment ideals when it comes to their own scientific research, when it comes to teaching methodology their management style is decidedly pre-Enlightenment.


[…] como más barato y “suficientemente bueno” reemplazos para el trabajo de la facultad. Jonathan Rees, profesor de historia en la Universidad del Estado de […]

Leave a Reply