Higher Ed Disruption: Not So New

Are you a faculty member or administrator who thinks that the latest technologies are finally going to enable us to teach our students well, or do you at least hope that’s the case? If so, you should reconsider, because the vaunted elements of the latest technologies have been around for some 100 years. It isn’t having the technology, but using the technology that is key to helping students learn well.

For at least the past decade there has been much talk about the advantages of highly sophisticated online courses and the use of online tools in traditional courses. One of the significant advantages of technology-enhanced courses, it is said, is that they can be tailored to individual students’ needs, and thus achieve desired learning outcomes for each student better and faster.

Consider for example, this quote from the website of the Apollo Group, the parent company of the University of Phoenix: "Based upon the belief that learning is not a one-size-fits-all experience, Apollo Technology developed the technology to deliver data-driven, personalized education tailored to the individual. Apollo Technology’s unique student data system collects and analyzes individual student data, and delivers automatic just-in-time guidance that can significantly improve student outcomes." In 2010, the University of Phoenix announced a new Learning Management System, the Learning Genome Project, that "gets to know each of its 400,000 students personally and adapts to accommodate the idiosyncrasies of their 'learning DNA.'" Similarly, a recent article in The New York Times stated: "Because of technological advances — among them, the greatly improved quality of online delivery platforms, the ability to personalize material … MOOCs [massive open online courses] are likely to be a game changer."

These statements are evidence of the general belief that now, using technology, we can achieve all sorts of personalized instruction, which constitutes a revolution in how we can help students learn.

But using technology to individualize student learning is not at all a new idea — it does not originate with online courses or with the technology developments of the past decade, or two, or even three. Using technology to individualize student learning is an idea going back at least 100 years. One of the original learning theorists of the modern era, Edward Thorndike, stated in his 1912 book: "If, by a miracle of mechanical ingenuity, a book could be so arranged that only to him who had done what was directed on page one would page two become visible, and so on, much that now requires personal instruction could be managed by print."

A couple of World Wars later, one of Thorndike’s intellectual descendants, B.F. Skinner, recognized as the most eminent psychologist of the 20th century, was developing and crystallizing the field of operant conditioning, the form of learning in which so-called voluntary behavior changes as a result of its consequences. In the third and final volume of his autobiography, Skinner relates that in 1953, in seeing how his daughters were being educated at the Shady Hill School, "I suddenly realized that something had to be done. Possibly through no fault of her own, the teacher was violating two fundamental principles: the students were not being told at once whether their work was right or wrong (a corrected paper seen 24 hours later could not act as a reinforcer), and they were all moving at the same pace regardless of preparation or ability. But how could a teacher reinforce the behavior of each of 20 or 30 students at the right time and on the material for which he or she was just then ready?.... A few days later I built a primitive teaching machine."

Skinner later developed more sophisticated versions of teaching machines, demonstrating one at the University of Pittsburgh in 1954. These machines presented math problems one at a time, with students having to solve each problem before being able to go on to the next.

In 1961 Skinner took a somewhat different approach to personalized instruction when he published, with Holland, the programmed textbook The Analysis of Behavior. This book focused on the principles of learning, more specifically, the principles of classical (Pavlovian) and operant conditioning, with an emphasis on the latter. The introductory pages of the book, echoing Thorndike in 1912, state that "the material was designed for use in a teaching machine…. Where machines are not available, a programmed textbook such as this may be used. The correct response to each item appears on the following page, along with the next item in the sequence."

Students wrote down their answers before turning the page, and repeated a section if more than 10 percent of the answers in that section were incorrect. I first encountered this book in the summer of 1968, as a 15-year-old student in a psychology course taught under the auspices of the National Science Foundation. Similar to other students in my group that summer, I finished this text within weeks and loved it. In 1964, in seventh grade, I had been the beneficiary of another programmed textbook, English 3200. This book was part of a very successful series that taught English grammar.

Another well-known figure in the origins of operant conditioning, Fred Keller, published his iconic article, "Good-bye Teacher…" in 1968. In this article he essentially advocates breaking down the entire teaching process to its elements, and conducting each of those elements more efficiently. The prime function of the teacher becomes, not to lecture, which is best left to automated means, but to engage in direct interaction with students in support of their individualized instruction. More specifically, Keller points out as important the following teaching elements:

1. Highly individualized instruction that allows students to progress at their own speed.

2. Clear specification of learning outcomes (the specific skills to be achieved).

3. Clear specification of the steps needed to achieve these learning outcomes.

4. A goal of perfection for each student and for each stage in the learning process.

5. Two types of teachers: Classroom teachers whose duties include "guiding, clarifying, demonstrating, testing, grading," and other teachers who deal with "course logistics, the interpretation of training manuals, the construction of lesson plans and guides, the evaluation of student progress, the selection of [classroom teachers], and the writing of reports for superiors."

6. Using lectures as little as possible — more as a way to motivate students, and using student participation as much as possible.

7. Lots of testing, all with immediate feedback to students, which helps to ensure student learning.

This breakdown of the learning process makes large parts of that process, parts that are ordinarily done in classrooms involving direct human interaction, well suited for being done by technology. However, humans are clearly still needed for specifying the learning outcomes and the steps required to reach them, as well as other tasks involving analysis and creativity and complex interactions with students.

Just a few years later, in the fall of 1972, I took an undergraduate course on learning at Harvard University, taught by William Baum, that followed the "Keller plan." The work was divided into 26 units, each requiring some reading, some questions to which answers had to be found and learned (50 to 80 such questions per unit, some of which would require an essay to really answer properly), and a written and an oral quiz. Students were not allowed to progress to the next unit until they had passed the written and oral quizzes for the preceding unit, and individual instruction with Baum or his graduate teaching assistant was always available. However, due to the large number of units in this 14-week course, and the difficulty of the quizzes, which students often did not pass, very few students finished the entire sequence and so very few students received an A. Thus using the Keller method does not automatically result in students doing well. The application of such teaching techniques is critical.

Lest anyone think that visions of improving learning by the use of technology are limited to psychologists, 1995 saw the publication of an outstanding work of science fiction by Neal Stephenson, The Diamond Age. A central theme in this work is an interactive book, owned by a small girl, that greatly facilitates her learning, development, and upbringing. We cannot yet achieve the degree of device interactivity that Stephenson describes, but we can achieve elements of that interactivity, and Stephenson gives us a vision of the possibilities.

In 1998, Frank Mayadas, then a program director at the Sloan Foundation, gave the keynote address at the City University of New York’s Baruch College’s first annual Teaching and Technology Conference. In this address he pointed out that all forms of college learning have three elements in common: an expert, who oversees the process; information sources; and colleagues, with whom a student learns. All three are important in the learning process, and all three may be instantiated in different ways depending on the modality of instruction. Although current technology cannot by itself design a new course, it can serve well as an information source, and it can assume some of the functions of colleagues. As technology continues to develop, the functions that it can serve will increasingly closely resemble those that have traditionally been served by humans.

The more recent past, 2010, saw the publication of DIY U by Anya Kamenetz. Consistent with Keller in 1968 and Mayadas in 1998, Kamenetz also would separate the components of the learning process, instead of concentrating them all in a course’s single professor as has been largely the case until now. In her vision of the future, individualized instruction is assumed, with technology playing a significant role, including by taking over those parts of teaching that can be automated.

Kamenetz’s vision is not far away given what is already happening on today’s campuses.  As stated in a 2012 report from the Ithaka organization, "Barriers to Adoption of Online Learning Systems in U.S. Higher Education": "Literally for the first time in centuries, faculty and administrators are questioning their basic approach to educating students. The traditional model of lectures coupled with smaller recitation sections (sometimes characterized as 'the sage on the stage') is yielding to a dizzying array of technology-enabled pedagogical innovations." One primary use of technology is to deliver lecture material outside of class, while class time is used for discussion and other active interactions involving the instructor and the students. This is known as the flipped classroom, which turns "traditional education on its head." But recall Keller’s 1968 suggestions about how teachers should be used for "guiding, clarifying, demonstrating, testing, grading," and that lectures should be "used as little as possible … and student participation as much as possible." It seems that the new invention of the flipped classroom is not so new at all.

What encourages these recent statements about the benefits of technology for learning is a worldwide recognition that what is important in higher education is the achievement of specific, agreed-upon learning outcomes. Although this emphasis was present at least from 1912 in the work of learning theorists such as Thorndike, who emphasize the end result — the behavioral goal — in their approach to changing behavior, it has only been in the past few decades that such recognition has become prominent in higher education.

One example is contained within what is known as the Spellings Report (the 2006 report of the commission that was appointed by then-Secretary of Education Margaret Spellings). A major point of this report was that "[a]ccreditation agencies should make performance outcomes, including completion rates and student learning, the core of their assessment as a priority over inputs or processes." It is this emphasis on learning outcomes that, in part, enables the use of technology in the learning process. Once the learning outcomes are specified, the process of helping students to achieve them can be programmed, using increasingly sophisticated technology.

Many of the elements of good teaching discussed here — for example, individualized instruction, frequent testing, focus on outcomes, immediate feedback — now have sound laboratory evidence to support their use (see a comprehensive survey here). We seem to have forgotten their behavioral psychology origins and history, yet it is their effectiveness that is important in the end. Perhaps there are additional lessons to be learned from behavioral scientists, however, in the use of technology to facilitate instruction. We have only to look at casino attendees, particularly the users of slot machines, to see evidence of what Skinner and Keller knew firsthand in the laboratory with rats, that animals (including humans) respond at a high, continuous, persistent rate on variable ratio schedules (situations in which each reward arrives after a variable number of responses). Using such knowledge, in addition to knowledge from cognitive psychology about how best to structure concepts, can result in online courses that not only make concepts easy to learn and remember but, similar to slot machines, are almost irresistibly attractive.

Keller in 1968 summed up his position on teaching with the following:

Twenty-odd years ago, when white rats were first used as laboratory subjects in the introductory course, a student would sometimes complain about his animal’s behavior. The beast couldn’t learn, he was asleep, he wasn’t hungry, he was sick, and so forth. With a little time and a handful of pellets, we could usually show that this was wrong. All that one needed to do was follow the rules. “The rat,” we used to say, “is always right.”

My days of teaching are over.  But … I learned one very important thing: the student is always right. He is not asleep, not unmotivated, not sick, and he can learn a great deal if we provide the right contingencies of reinforcement.

Although we can all agree that college students are certainly not the same as casino attendees or lab rats, we can also all agree that technology, designed and used correctly, can facilitate instruction through personalization as well as through motivation. (The popular appeal of many online role-playing games is one example of that.)

The teaching techniques and tools discussed here have been promoted by behavioral psychologists for the past century.  What lessons can we learn from this?  One is that it is possible to facilitate learning using the techniques discussed here, such as personalized instruction, without ever having to use the latest (very expensive) technology.  There are times when a relatively cheap programmed textbook will help someone learn, perhaps not as well as the best online programs, but very well.

A related lesson is that it is not the existence of the latest technology or its potential uses that will help us to maximize student learning, but using what we know and have. Faculty must be both aware of the techniques and tools at their disposal, and want to use them. This requires proper training during graduate school, professional development later on, and appropriate college and university incentive structures (all of which have been too often missing if the repeated rediscovery of these techniques and tools during the past century is any indication).

The sorts of tools that we have needed to help students learn have been around for 100 years, albeit continuously improved. It is our job to — finally — use those tools.