Newton, One-on-One

from Harvard Journal, Summer 1995
by Kathleen Koman

The Levi Straus trademark shows two horses trying to pull apart a pair of pants. Suppose Levi had only one horse and attached the other side of the pants to a fencepost. Using only one horse would: (a) cut the tension on the pants by one-half, (b) not change the tension on the pants at all, (c) double the tension on the pants?

Physics professor Eric Mazur brings a new set of tools to the classroom - including hand-held computers, novel brain-teasers, and a lot of running around.

In 1989 McKay professor of applied physics Eric Mazur read an article in the American Journal of Physics that contained a diagnostic test to assess understanding of Newtonian mechanics. He gave the exam to his Harvard students, and the results shocked him: "The test clearly showed that what I thought they were understanding they did not understand." Mazur had been teaching at Harvard since 1984 and until then had depended chiefly on the traditional lecture format. But after this event, he quickly reformed his teaching ways. Among his innovations is Peer Instruction, a pedagogical method that actively involves students in the teaching process.

Mazur breaks down his Physics 11 lectures into a series of digestible snippets, emphasizing concepts and ideas and avoiding equations and derivations. Every 15 minutes an overhead projector flashes ConcepTests on a screen. These short conceptual questions generally require qualitative rather than quantitative answers. Example: a rubber bullet and an aluminum bullet, equal in size and mass, are fired at a block of wood at the same speed. The rubber bullet bounces back, while the aluminum one penetrates. Which is more likely to knock the block over? answer

The students get one minute to think about each ConcepTest, then punch their answers into hand-held computers. They have another minute to convince their neighbors of their logic. Chaos erupts as everyone starts to argue: bodies twist, heads shake, hands gesticulate. The buzz continues until Mazur calls time and asks students to plug their revised answers into the computer. They also record their confidence levels - pretty sure, not quite sure, or just guessing. Responses are uninhibited because the ConcepTests are not graded.

The computer instantly shows the percentage who answered correctly. It also displays a map of the seating arrangement with green seats for students who chose the correct answer, red seats for those with incorrect responses. By clicking on the seats, Mazur can identify each of his two hundred-odd students by name. "It's like having Big Brother watching you," he says, with a grin. "But Big Brother has so much data that the details are lost. I use only global information, like 'How many had the right answer?' - not 'What did Mary answer?'"

The convince-your-neighbor arguments systematically increase both the percentage of correct answers and students' confidence. Mazur sees the greatest improvement when about half the students are correct initially; there can be gains of a large as 40 percent. If Mazur is not satisfied with the students' performance, he slows the pace, providing more detail on the current subject, and then reevaluates with another ConcepTest. When the results indicate mastery of the concept, Mazur moves on the next topic.

During argumentation periods, Mazur runs up and down the aisles of the lecture hall, eavesdropping on conversations. "I listen to students making the wrong argument," he says. "I know the right answer; it's very hard for me to think of wrong answers and wrong reasonings. I can learn those only from experience." Because students are vividly aware of the difficulties involved in grasping a concept, they are often more effective at explaining an idea than teachers are.

"I feel awkward if I don't get feedback," Mazur says. "It's so much more satisfying when you have a response from your crowd. I think that any teacher, once exposed to this method, would find it hard to resist."

Apparently so. Word of Mazur's teaching technique began to spread after a favorable review in Sheila Tobias's 1992 book, Revitalizing Undergraduate Science Education: Why Some Things Work and Most Don't. To promote Peer Instruction, Mazur has mailed out three hundred copies of his teaching manual, the outcome of a contract with the National Science Foundation. Recently, he created a World Wide Web server on the Internet that enables teachers to submit and retrieve ConcepTests. Many are now experimenting with Mazur's technique - some of them in disciplines other than physics.

"It's really changed the culture of the large lecture classes that I teach," says Arthur Ellis a professor of chemistry at the University of Wisconsin at Madison. "It makes the classroom much more lively. I feel like I'm not playing with a full deck if I don't use this method now."

Even more important is the response of Mazur's students, one of whom appears in a videotape produced by Harvard's Derek Bok Center for Teaching and Learning. She says that solving a problem by using the Peer Instruction method gives her "an 'Ah Hah!' kind of sensation," then adds, "It's not that someone just told me; I actually figured it out. And because I can figure it out now, that means I can figure it out on the exam. And I can figure it out for the rest of my life."