January 13, 2009

M.I.T Drops Physics Lectures; NY Times Confused

The NY Times "reports" that M.I.T. has dropped its Introductory Physics lectures in favor of "smaller classes that emphasize hands-on, interactive, collaborative learning." Dropping the lecture is probably a good thing; but, the Times article makes little sense. Probably because the reporter, Sara Rimer, doesn't seem to understand the nature of science instruction and doesn't provide enough information for the reader to understand what's going on here.

Introductory Physics is a typically a four credit class. Actually, two semesters of four credits. This means that there's four hours of classroom instruction provided each semester. Typically, these courses are followed by (at least) a two credit lab course in which students take what they've learned and conduct highly scripted Physics experiments.

Here's how Introductory Physics is traditionally presented in a typical school week.

1. Student reads the next section in the textbook.
2. Professor explains the section, provides his insights, and works some problems from the section. (one hour)
3. Student attempts to solve the assigned problems from the section, either alone or in a study group.
4. Grad Students work with students in small groups reviewing the problems to ensure the student understands the material. (one hour)
5. Repeat steps 1-4 for next section. (two hours)

That's how it's supposed to work. In practice, it usually goes a little differently.

1. Student fails to read next section or doesn't understand next section.
2. Professor reviews exactly what is written in the textbook without providing any insight that the student could have gotten by reading the book on his own.
3. Student fails to do some, many, or all of the assigned problems either through lack of understanding or laziness.
4. Grad student reviews problems and student copies answers.
5. Repeat.

Note that the student is supposed to be spending at least eight hours a week studying and solving problems outside of the classroom. Some students front-load the work and do their work before the material was presented and the problems are reviewed like in the first example. Other students back-load the work and do their work after the material was initially presented and the problems solved like in my second example. Most students, however, fall out somewhere in the middle.

This is basic direct instruction using worked problem examples. Research shows that it is an effective way to teach novice students, and, by definition, students taking an introductory Physics class are novice students. These students have a long way to go before they are experts in Physics. Using worked problem examples is less effective with experts and non-novices with considerable domain knowledge. The difference is domain knowledge. The experts have it; the novices do not, at least not yet.

In fact, they won't have it the following semester either when they take Physics lab. Undergraduate Physics lab is closer to baking a cake from scratch than it is to real science. It is a highly scripted affair because the students do not yet know enough physics or how to conduct a real experiment on their own yet. They are novice scientists and the lab provides another opportunity to follow worked problems. In this case, the worked problems are the scripted experiments.

I think now we have enough background knowledge to make some sense out of the Times article.

For as long as anyone can remember, introductory physics at the Massachusetts Institute of Technology was taught in a vast windowless amphitheater known by its number, 26-100


The physics department has replaced the traditional large introductory lecture with smaller classes that emphasize hands-on, interactive, collaborative learning. Last fall, after years of experimentation and debate and resistance from students, who initially petitioned against it, the department made the change permanent. Already, attendance is up and the failure rate has dropped by more than 50 percent..

Right off the bat, I find it hard to believe that all four hours classroom time in Introductory Physics are present in a large ampitheater. Are there any science/engineering majors (especially those attending M.I.T) out there that that were taught like this. Undergraduate physics instruction is all about learning how to solve basic physics problems. BY necessity this will involve the student working hundreds of problems over the course of the semester on his own or with a study group. There's no getting around that fact. Even the most direct instruction of courses requires that the student work the problems on his own following an introduction by an expert (the professor) and concluding with a review of the problems with an expert (a grad student or the professor). This requires motivation on the part of the student. And that appears to be a problem at M.I.T, as you'll soon see.

Also note the petitioning of the students adn the failure of the Times to get to the bottom of that. We'll get to that later.

The traditional 50-minute lecture was geared more toward physics majors, said Eric Mazur, a physicist at Harvard who is a pioneer of the new approach, and whose work has influenced the change at M.I.T.

“The people who wanted to understand,” Professor Mazur said, “had the discipline, the urge, to sit down afterwards and say, ‘Let me figure this out.’ ” But for the majority, he said, a different approach is needed.

I think Professor Mazur is delusional. Physics forms a critical foundation for most of the students learning a hard science or engineering. Subsequent courses will build off of what is learned in introductory physics and the physics problems will be revisited and expanded upon often in subsequent years. So, a student who does not possess the urge to put in the hard necessary to learn physics is in for a rude awakening sophomore year. The years of coddling in high school are over; now is the time for real work.

“Just as you can’t become a marathon runner by watching marathons on TV,” Professor Mazur said, “likewise for science, you have to go through the thought processes of doing science and not just watch your instructor do it.”

That's stating the obvious now isn't it. And I find it hard to believe that M.I.T students were merely watching their instructor solve problems for four hours every week in a large amphitheater and not actually solving their own problems. I'm sure somewhere along the line students were being assigned problems to work out of class and that some time in-class was spent reviewing those problems and their solutions. Are we to believe that only the Physics majors were doing their homework?

Then we have this non-sequitur.

In an article in the education journal Change last year, Dr. Wieman noted that the human brain “can hold a maximum of about seven different items in its short-term working memory and can process no more than about four ideas at once.”

“But the number of new items that students are expected to remember and process in the typical hourlong science lecture is vastly greater,” he continued. “So we should not be surprised to find that students are able to take away only a small fraction of what is presented to them in that format.”

What does this have to do with anything related to this article. The magic number 7 is a problem under both the old way and the new way at M.I.T. Either way, the students are learning more than they can absorb. That's why they take notes and write stuff down. Students really bump up against the short term memory problem when they try to solve the problems until they have learned the underlying material. The new way of teaching doesn't fix that problem. The only thing that fixes that is lots of practice solving problems. Are the students getting more practice under the new system? Let's see.

At M.I.T., two introductory courses are still required — classical mechanics and electromagnetism — but today they meet in high-tech classrooms, where about 80 students sit at 13 round tables equipped with networked computers.

Instead of blackboards, the walls are covered with white boards and huge display screens.

Why are journalists such suckers for bright lights and fancy gizmos? I've yet to see any of this technology used in a way that is pedagogically superior to a blackboard and a slide projector.

Circulating with a team of teaching assistants, the professor makes brief presentations of general principles and engages the students as they work out related concepts in small groups.

Teachers and students conduct experiments together. The room buzzes. Conferring with tablemates, calling out questions and jumping up to write formulas on the white boards are all encouraged

This is the money graf. Here's where we find out that M.I.T hasn't really done away with the lecture, they've just shuffled the chairs. Instead of Lecture for an hour in a classroom and then solve problems for an hour in small groups with grad students, M.I.T. now has the professor lecture for a short period of time then the students solve problems for a short period of time with the help of the professor and grad students in the same room, repeat until the class is done. What's the difference?

I don't see the advantage, except maybe that the lazy students are being forced to do the work under the watchful eyes of the instructors instead of copying the problems they should have worked out before a later problem solving period. But, since they're working in groups now, there's no guarantee that they're not free-riding off of their neighbors instead of free-riding off of the grad student in the separate recitation period.

What the article describes is exactly what was going on in our problem solving classes with our grad student after our lecture. The only change is that we had blackboards.

M.I.T hasn't done away with the lecture; they've merely rearranged it in a way that is no more sound in a cognitive science sense than it was before. What M.I.T. is doing is providing another year of coddling. It's also still direct instruction. (Though I can't wait to see how Stephen Downes is going to try to spin it.)

And, the students aren't experimenting, they are solving problems. There is a big difference.

“There was a long tradition that what it meant to teach was to give a really well-prepared lecture,” said Peter Dourmashkin, a senior lecturer in physics at M.I.T. and a strong proponent of the new method. “It was the students’ job to figure it out.”

Our professor gave some really good lectures and then he ran one of the problem solving sessions. Is there a difference?

Apparently the problem is really an attendance problem.

John Belcher, a space physicist who arrived at M.I.T. 38 years ago and was instrumental in introducing the new teaching method nine years ago, was considered an outstanding lecturer. He won M.I.T.’s top teaching award and rave reviews from students. And yet, as each semester progressed, attendance in his introductory physics courses fell to 50 percent, as it did, he said, for nearly all of his colleagues.

“M.I.T. students are very busy,” Professor Belcher said. “They see the lecture as dispensable, that is that they can get it out of a book more efficiently than getting up, getting dressed and going to lecture.”

After three years, Professor Belcher had had enough. “I had poor attendance, and was failing 10 to 15 percent, and grading the tests and shaking my head in despair about how little was getting across,” he said. “And this is a subject — electromagnetism — that I love.”

Here's the thing. The problem solving sessions are critical to success. The lectures less so if the textbook presents the material well and the student reads it beforehand. This is especially so if the professor is a bad teacher. Under the new system the students are forced to endure the gas-bag "initial presentation,"i.e., mini-lectures, to get to the problem solving part.

Maybe that's why the students are petitioning. The silly lectures are no longer optional for the student. That's why attendance is up. Here's another reason why attendance is up:

Unlike in the lectures, attendance counts toward the final grade, and attendance is up to about 80 percent.

I suppose the clickers don't hurt.

“One of the newer professors, Gabriella Sciolla, who arrived in 2003, was teaching a TEAL class on circuits recently. She gauged the level of understanding in the room by throwing out a series of multiple-choice questions. The students “voted” with their wireless “personal response clickers” — the clickers are essential to TEAL — which transmitted the answers to a computer monitored by the professor and her assistants.

You know where they are,” Professor Sciolla said afterward. She can then adjust, slowing down or engaging students in guided discussions of their answers, as needed.

Lecturing in 26-100, she said, she could only look out at the sea of faces and hope the students were getting it.

Unless they had clickers because if they had clickers in the lecture hall, the professor would get the same feedback.

What I see here is a distinction without a difference. The learning is no more active then it was under the old system.

What is left unexplored by the Times is why there was protesting by the students. Students are no fans of boring lectures. And I'm sure under the old system there was plenty of in-class problem solving and opportunity for feedback. Under the old system students were supposedly left on their own to solve difficult physics problems. Now they get to do the same thing in a high-tech classroom with all their classmates and lots of teaching assistants milling around. So why the protests?

I suspect once we learn why, we'll get a better idea of the problems of the new system at M.I.T.


Anonymous said...

"I find it hard to believe that all four hours classroom time in Introductory Physics are present in a large ampitheater. Are there any science/engineering majors (especially those attending M.I.T) out there that that were taught like this."

When I was attending college -- 20 years ago -- a typical class week was three hours of lecture in a large ampitheater (MWF for one hour, or TR for 1½ hours) and one hour of TA time in a small classroom.

-Mark Roulo

Anonymous said...

"What the article describes is exactly what was going on in our problem solving classes with our grad student after our lecture. The only change is that we had blackboards."

Well, not quite. Consider the turn-time for the closed loop in the new scheme: minutes. If you get a wrong answer, you can know very quickly. Plus, you can then get help to correct things.

The old scheme might go like this:
Monday: lecture.
Tuesday: Do problems. Get them wrong.
Wednesday: more lecture. You still haven't understood the problems you had on Tuesday.
Thursday: More struggling.
Friday: Finally, section with the TA. You can get help (or maybe not depending on the TA).
Friday: Last lecture.

You can get *very* lost very fast with the traditional approach (especially with poor TAs). The new scheme tightens the feedback loop for the students who care.

-Mark Roulo

Mark Roulo said...

"Our professor gave some really good lectures and then he ran one of the problem solving sessions. Is there a difference?"

Two differences: tightness of feedback loop and the professor is running the problem solving sessions.

-Mark Roulo

KDeRosa said...

Hey Mark.

Good points.

I think I made some of them in my next post.

Unknown said...

As an MIT alum who majored in physics and was at MIT recently, during the transition to TEAL, I thought I'd chime in.

There's a lot more context needed here. The first thing you need to know is that first semester at MIT is pass/fail to take a little of the pressure off. You can think whatever you want about that, but it's a big part of the reason that there was such a difference in learning between physics majors and non-physics majors in these courses.

If you are majoring in physics, you need to know this stuff, and you need to get it at a really deep level. That means reading the book, going to lecture, doing the problems ahead of time and asking good questions of your TAs.

If you are not majoring in physics, you merely need to pass. You can skip class, avoid the book and copy answers off your friends and TAs.

I believe MIT was trying to solve this problem.

The result is that the new class is slightly less challenging, but much friendlier. There's more emphasis on making sure that students understand the problems during the class before they leave to do the problem sets.

You are right about the codling. MIT is making it a little easier to learn physics in these classes, and that does mean that students don't get as much exposure to what it really means to solve tough physics problems or do experiments (incidentally there is no physics lab requirement at MIT).

I think this codling is ok, because at the college level, doing physics problems is really different from solving engineering problems or learning biology or chemistry or math. You only need a deep understanding of how to solve physics problems if you are a physics major, and you will get plenty of that in later years.

Not to mention, there is still a more advanced version of the freshman requirement that many interested students elect to take. This version is still run as a traditional lecture/homework/TA class, and it does force you to begin learning the skills of investigating and solving physics problems on your own.

I personally prefer the old way, but I think that so long as MIT continues to offer the choice, it's a good move to have students learn in the TEAL environment by default.

Anonymous said...

"The result is that the new class is slightly less challenging, but much friendlier. "

When I was at UCSB 20 years ago, they "solved" this problem by having three physics tracks:

1) Physics for Poets (ie. w/o math)
Here you would learn that things fall down when you let go of them.

2) Physics for Engineers, Chemists and Physicists
This is/was the "real thing". 4 or 5 quarters long (depending on whether you took the quantum/relativity quarter). You needed to use calculus. Second quarter of this sequence caused me to become a chemist.

3) Physics for Biologists
This is/was a three quarter sequence that still had math, but wasn't as hard as (2).

My guess is that MIT doesn't have many students who care about Physics for Poets.

This suggests that MIT is weakening their physics for physicists class to make things easier for the biologists. I'd rather see two courses offered.

-Mark Roulo

Anonymous said...

Mark, your use of the letter R for Thursday made me think you went to MIT. But your last post about MIT needing two tracks for 8.01 confused me--they ALREADY have 3 tracks, at least, and have had at least 3 tracks for nearly 20 years now. (They have 8.01L, the long version, for those who can't handle 8.01. It adds one month to the course, and is smaller, more hands on. They have 8.012/8.022 for those with massive physics experience and with solid advanced calc already under their belts (it used the book by Kleppner & Kolenkow) and they used to have 8.01X, the experimental group not to mention the ones inside Concourse, ISG, etc.

MIT's 8.01 was taught with 3 hours of amphitheater, and 2 hours of recitations section in a classroom format with another professor running that--not a TA.

"If you are not majoring in physics, you merely need to pass. You can skip class, avoid the book and copy answers off your friends and TAs.

I believe MIT was trying to solve this problem."

No, not exactly. MIT tried to solve THAT problem by getting rid of Pass/No Record for the first year, now it's only the first term, maybe even less.

Recall that MIT shows EVERY SINGLE first year physics problem worked out by Walter Lewin on their internal television network. There is no need to copy answers from friends. You just copy Lewin.

Some teachers felt they needed a new method, didn't like the (typical) drop off in attendance, and that the current failure rate was too high. Of course, that could be attributable to a variety of factors, including less preparation for the students in high school. The attendance factor is huge of course. but attending to that as a issue rather than a symptom would be mindboggling for MIT's other cultural issues.

Anonymous said...

No, I think your takeaway is wrong. This is a fundamental change to the way 8.01 is taught. The comments on the NYT article are more telling.

Here is how MIT's old 8.01 worked:

after 3 hours of lecture, you had assigned problem sets, one a week. the problem sets took 10 hours a week, roughly. That was normal. Recitation section would usually show you how to work related problems. If all else failed, you could watch the problems be done by Walter Lewin himself, explaining each step, on the TV at MIT. It ran continuously all day, and all night, so you could watch it over and over again. Feedback was definitely not tight unless you made it so--you looked at the answer key, etc. But you had various opportunities to admit you didn't understand, if you were willing to do that--most young MITers aren't.

Apparently now, the feedback loop isn't tighter, because you work in "groups" where only one person has to do the quiz for all of you to get credit. The clickers are just given to the ones with the most prior knowledge, and you're seated to "Share the knowledge". So the strugglers learn less, but their attendance raises their grade.

This is inferior to the old way, where at least the kids with prior knowledge would generally group together, and leave the others all together, struggling. The strugglers would then have to face their own attempts to learn something while the more advanced would race ahead. This was healthier.

Anonymous said...

"Mark, your use of the letter R for Thursday made me think you went to MIT. But your last post about MIT needing two tracks for 8.01 confused me--they ALREADY have 3 tracks, at least, and have had at least 3 tracks for nearly 20 years now."

I didn't go to MIT for several reasons:

(1) I didn't know that I would get in, and didn't apply (I had a hard time seeing how yet one more white male with high SAT scores would be of interest to MIT).

(2) If I did get in, it would be very expensive. I wasn't real keen on the idea of graduating with up to $100K of debt. I didn't see much point in applying if I wasn't going to accept.

(3) I had pretty much *no* social life in high school. Another four years of this didn't seem very appealing, and I didn't think I would be competitive at MIT without spending most of my time studying.

(4) UCSB sent me a promotional package with a 8½"×11" glossy shot of their lagoon. It looked like this:


but from a slightly different angle.

But ... if MIT has already sorted physics by seriousness, then I have no idea what MIT thinks it is doing.


-Mark Roulo

Anonymous said...

Just to elaborate :-)

The 8 story building about ¼ of the way from the top of the picture pretty much centered left/right is a dorm (actually, two dorms). Two minute walk to the lagoon, and three minutes from there to the seashore :-)

Four years here or four years in the Northeast and ~$100K of debt. Hmmmm ... decisions, decisions ...

-Mark Roulo