"I hear and I forget. I see and I remember. I do and I become."
That ancient Chinese proverb could serve as the slogan for a radically different approach to higher education advocated by Larry J. Leifer, director of the Center for Design Research and professor of mechanical engineering at Stanford University.
The approach, called product-based learning, takes the idea of "learning by doing" to its logical limit. The testbed in which Leifer and his colleagues have developed and refined this method for more than 20 years is a graduate mechanical design course, ME 210. Its students learn the subject by designing and building actual devices for corporate sponsors.
In a paper delivered by graduate research assistant Ade Mabogunje at the Seattle meeting of the American Association for the Advancement of Science on Feb. 15, Leifer argued that this approach has a number of advantages over conventional pedagogy and could be extended to undergraduate courses and other subject areas such as science, mathematics and the humanities.
Leifer defined product-based learning as a "problem-oriented, project-organized learning activity that produces a product for an outsider." The approach insists that the projects involved be defined by outside groups.
In the traditional approach to learning, instructors pick a series of problems that usually have a single correct answer. In project-based learning, the problems are usually larger, less structured and tackled by student teams. "As long as the problems are academically chosen," Leifer said, "their purpose is almost always to make some issue self-evident, a situation that may discourage student initiative and active exploration."
Product-based learning takes a different approach. When the project comes from an external source, a major part of the students' task is to figure out what they need to know in order to solve the problem. There is considerable research indicating that active learners who pose questions retain the knowledge they gain better than passive learners who answer questions posed by others. While the learning is deeper, there is a trade-off: Fewer "fundamentals" may be covered in a given quarter. But Leifer says that a portfolio demonstrates mastery, whereas a transcript emphasizes the number of subjects to which a person has been exposed.
Another major advantage of product-based learning is motivational. "Industry involvement gives the work relevance," Leifer said. "Interestingly enough, many educators are not concerned about external validation of what they teach. That may be why relevance continues to be a major problem in education."
Products that student teams in ME 210 have successfully developed in recent years include a commuter bicycle for BMW; a special bumper for Ford to decrease pedestrian injuries; a portable, wireless banking appliance for NCR; and a surgical simulator for Immersion Corp. to help train doctors and nurses how to insert a spinal-tap hypodermic needle correctly.
A side benefit of industry involvement is that companies absorb part of the cost involved in this type of training. In ME 210 they provide about $500,000 per year to help cover the costs of product development and assign industry liaisons who work directly with students. "This is operating money, not philanthropic money. They aren't giving it to us because we are a worthy cause, but because we can provide them with something of real value," Leifer said.
There is no reason why undergraduates could not produce valuable products just like the graduate engineering students in ME 210, he added. College freshmen rank academically higher than three-quarters of the working population. Likewise, science and mathematics majors could produce products of value to companies and their communities, he said.
"Science and math education suffers deeply from a lack of context. It is generally delivered as if the material has come down from on high. Product-based learning could give it real relevance," Leifer said.
The primary unit in product-based instruction is not the individual student, but the team. Generally, the 40 to 50 students in ME 210 are organized into three- and four-person teams. Teams are graded as a unit. As a result, the students learn to integrate their social as well as technical skills.
"When projects fail, it is rarely for technical reasons. It is usually because of a failure in team dynamics - members fail to get along productively," Leifer said.
The ability to work as a member of a team is increasingly important for professional engineers for economic and technological reasons, forcing engineering companies increasingly to rely on interdisciplinary design teams that collaborate for the life of a project and then disband. This is also true in a number of other professions, Leifer said.
Another defining characteristic of the Stanford approach is its emphasis on documentation. The quality of a student team's documentation is given equal weight to the quality of the prototype it makes. Not only do student team members prepare extensive design documents quarterly, but also their e-mail messages, notebooks and sketches are all captured on computer, through special software that has been developed for the course.
This material provides a wealth of examples for assessing team performance and also provides a treasure trove of information for studying the effectiveness of the approach and identifying ways to improve it. In his doctoral thesis, for example, Mabogunje used these records to search for objective measures of team performance.
"Teams with members who have similar working styles tend to work more efficiently than diverse teams but explore a very limited range of possibilities. Teams consisting of people from varied backgrounds and personal preferences may spend extra energy collaborating, but they also explore a greater range of possible designs and produce products of higher quality," Mabogunje found.
Leifer also uses course documentation to raise the bar for each successive class. New students are shown the best projects from previous years and told they are expected to equal or surpass them.
"We don't yet have the objective evidence lined up, but I think this strategy is working," Leifer said. "For example, we used to spend considerable time teaching oral presentation skills. But now all the students 'get it' very quickly on their own, so we can spend more time on deeper issues."
Some external validation of the increasing quality of the course's design projects exists. Between 1977 and 1990, 29 percent of Stanford entries won awards in the Lincoln Foundation Graduate Design Competition. Since then, the percentage has jumped to 60 percent; in 1995, the class took 11 of the 12 awards issued.
Although product-based learning is inherently more labor intensive than standard lecture courses, Leifer contends that it has evolved in a manner that makes its widespread adoption financially feasible for undergraduate, graduate and continuing education courses.
"We should no longer be in the business of helping students develop transcripts. Instead, we should be helping them develop portfolios," Leifer said.