Skip to content Skip to navigation

Research & Ideas

Search this site

Lessons in remote learning from the 1970s: A Q&A with James Gibbons

The former dean of Stanford Engineering looks to experiments he did more than 45 years ago to help answer the question that’s on everyone’s mind: How will online learning work out?

An early image of distributed tutored video instruction

Look familiar? An early image of distributed tutored video instruction, circa 1994/95 | Courtesy Jim Gibbons

In these days of COVID-19, everyone from parents to teachers to school administrators, not to mention the students themselves, is worried how this nationwide experiment in online learning is going to work out.

But there is one Stanford professor who says there are lessons to be learned from work done more than 45 years ago that might ease the transition and the worry.

Back in 1971, long before he became dean of the Stanford School of Engineering, Jim Gibbons, a highly regarded professor of electrical engineering, was asked to join President Nixon’s Science Advisory Council, which was then studying the effectiveness of televised education. A notion Gibbons had during those early days soon led him to a partnership with Hewlett-Packard Company and the creation of what Gibbons then dubbed “Tutored Video Instruction,” or TVI.

TVI is based, in essence, on recognizing that students learn best when they are allowed to ask lots of questions during a lecture, many more than the lecturer has time to answer. “TVI solves this dilemma,” Gibbons said. His solution was to collect students in small groups who would watch videotaped recordings of a live lecture in the presence of a tutor/facilitator. “The tutor’s role is not to answer questions outright, but simply to stop the tape when there is a question and guide the students to work out the answers among themselves.”

This method proved to be very successful. The process was an alternative to the then-current use of closed-circuit TV for distance learning in which the students could interject questions in a live lecture over an intercom. A variation on Gibbons’ basic process was developed in the mid-1990s in which each of the students and the tutor were not in the same physical location but joined by an early form of the internet that looks remarkably similar to the videoconferencing environments of today.

We sat down with Gibbons (remotely, of course) to talk about the lessons he drew from that time and what educators might do in the age of COVID-19 to improve outcomes at schools across the country.

How did TVI come about?

A subset of the Science Advisory Council started by reviewing a very large study comparing televised classes with live classes. The study covered every subject matter from math to arts, from kindergarten to a baccalaureate degree. It was a huge study, 363 different experiments. The overall answer was: There is no significant difference in student learning between TV and live instruction. That of course means different things to different people. If you are a fan of live instruction, you’d say: “See, there’s no reason to do TV.” And, if you’re a proponent of televised learning, you’d say, “See, there’s no reason to build all these schools.”

I was flying back to California from Washington when the light went on, and I thought, “I know how to fix this!” So I wrote down the basic TVI idea in which I proposed to use pre-recorded lectures that could be viewed by small groups of students facilitated by a tutor, much like the lecture and recitation model familiar to most research universities. TVI was designed to allow the students to manage the pace of the lecture to suit their need, to ask lots of questions and get them answered, on the spot if possible.

This idea met with considerable enthusiasm among my colleagues on the advisory council; but it would still be buried in the bowels of the White House to this day except for the fact that Hewlett-Packard was moving one of its divisions from Palo Alto up to Santa Rosa and their CEO at the time, John Young, wanted us to figure out a way for their young engineers in the honors co-op program — Stanford students who worked for HP while finishing their degrees — to take classes remotely. This was perhaps the first experiment ever done in what is today called collaborative learning.

What was your guiding philosophy going in?

I always understood that students learn best when they have the opportunity to ask lots of questions in a psychologically supportive environment. Their questions should be answered when they arise, while fresh. Not doing that often leads to increasing confusion. As a corollary, I believed that answers are most meaningful when the students work them out themselves. These objectives are best met in a small group led by a tutor who’s helping the students find answers, not feeding answers to them. That’s the essence of TVI.

The Stanford honors co-op students working in Santa Rosa would gather locally at a given time each day, and the onsite tutor would play a lecture recorded the previous day by a Stanford EE faculty member: me or one of my colleagues. The facilitator’s job was to stop the tape anytime there was a question, and keep it stopped until the group could work out the answer.

We found that on average the tape would be stopped for 3-5 minutes per question. So it took between 75 and 90 minutes to watch a 50-minute lecture. But when it’s finished, you have a bunch of very satisfied students. We eventually expanded the program to other HP sites around the world, where bilingual HP employees served as tutors to offer TVI sessions in Japanese, German, Spanish and French.

Did you do any in-depth study of the result?

We did. We looked at the data every way you can imagine. We compared that first class in 1973-74 against students on campus. The TVI students averaged a 3.7 on the grade scale. They outperformed the vast majority of the on-campus class (3.73 vs. 3.37). Some people said, “Oh, well, they cheated on the exams,” but I had insisted that the TVI group do the same homework and take the same exams at the same time as the in-class students. If the exam was at 8:00 in the morning in Palo Alto, those students in Santa Rosa would get up at 6:00, drive down here, have a little coffee, go in and take the exam, hand it in, turn around and go back.

Then we also tracked results over time — an eight-year study. On the Stanford campus, there were 830 students who took the first-year graduate-level courses over that period. Their GPA was 3.43. There were 89 TVI-only students who averaged 3.68. There’s no way that they aren’t just flat learning better. We published all these results in scientific journals.

What happened when TVI became “distributed,” with students and instructors all in remote locations?

Well, eventually, as the technology evolved, we began working with Sun Microsystems to create what we called distributed tutored video instruction — DTVI. We imagined the students to be remote from each other. We provided each of them with a microphone and a video camera to support remote communication within the group. We did an experiment at two campuses of the California State University system where we had 700 students at the two universities. We ran a regular lecture, a TVI group and a DTVI group for every class. The DTVI students were in their own rooms, connected to each other through our early version of the internet.

Sound familiar? Well, it should. This is exactly what Zoom does, right? In fact, it looked just like Zoom in the gallery view, with everyone wearing headsets and so forth. The results showed about the same performance academically, between TVI and DTVI, with each of them being superior to the live lecture class over a range of subjects.

How did you select tutors?

That is a critical point. We did a lot of experiments on what it took to be a good tutor. In the end, we found that one criterion was important above all others: a tutor needed to be a good facilitator. Familiarity with the course being taught is certainly useful, but it doesn’t have to be right up to date.

The tutors’ principal role is to engage their students in discussion when a point just made in the lecture is unclear. Tutors need to make sure that every member of their group has a chance to contribute to the answer.  When the group cannot resolve its problem, tutors are encouraged to contact the person who gave the lecture and then get back to their group to explain what they all missed. Both the students and their tutors appreciate this “deep backup” greatly.  The faculty member is standing behind his or her course.

It is also very important to say that tutors do not grade their students. The faculty member and the course team do that. The grade each student gets in the class has to be independent of how he or she takes the class. This is critical for maintaining the standards of the home institution, and for the satisfaction of the students who seek approval from the home institution.

So how is this different from an on-campus Problem Session led by a teaching assistant?

There are several differences. The main one is that all the TVI students’ questions are focused on understanding the lecture as it was delivered. On-campus problem sessions with a TA [teaching assistant] are based primarily on the homework except when the TA decides the lecture was not clear. 

So, a more refined question is: “Is the performance of TVI students on their homework better than their on-campus counterparts?” Since we had all homework graded on campus by the same TAs, we could show that the TVI students performed better on the homework. Studying what the lecturer actually said made doing the homework easier!

What lessons can teachers today take away from your experience?

If you have to, or choose to, teach online, and you have one or more TAs, then you use them to tutor students via Zoom in small groups of 8 or 10 at a time. Of course, you may have 100 students taking the class collectively, but they’re going to be divided into groups for the facilitated instruction. The professor remains on-call in an office-hours-like arrangement for tough questions that require clarification for the tutor and the class.

With some basic organization of TAs, I think this model could have an important impact on the way we do remote instruction today. It’s not that different from the on-campus model of lecture and recitation at many schools today, with one big advantage: The TA can stop the lecture at any time and the students can work out answers to their questions immediately. The grade averages from our earlier experiments indicate that online education can be made to work at least as well as in-person education on a purely academic level.

During the pandemic, when online is the only option for many schools and school districts across the country, these lessons might prove really helpful in ensuring that the content gets learned. The model works.