Science, Technology, Engineering & Math | Feature
Arizona State Tries Practice over Theory in Engineering Education
- By Dian Schaffhauser
When it comes to learning engineering, real-world problem solving has greater appeal to students than theory--it can also improve test scores. These are initial findings suggested by a recent experiment at Arizona State University performed as research for NASA. In two junior-year courses in the university's aerospace engineering program, faculty shifted how they were teaching to put less focus on theory and more emphasis on simulation and visualization through the immediate use of engineering software.
In 2009 the National Aeronautics and Space Administration awarded a $182,512 grant and the potential for a year's renewal to Valana Wells, associate professor in the Tempe-based Department of Mechanical and Aerospace Engineering, to lead work on developing what NASA said it hoped would be a "a revolutionary approach" to teaching aeronautical engineering. As Wells explained at the time the grant was issued, this modernization effort would incorporate "the use of computational tools for aircraft analysis that were not available when many of today's textbooks were written." The specific "computational" tools were MATLAB and Simulink from MathWorks.
Faster Iteration of Potential Solutions
Wells was joined in the work by Professor Kyle Squires, chair of the department, and lecturer Praveen Shankar, who taught one of the courses in the study. Students had previously given Shankar, a relatively new instructor, feedback that the class he was teaching, aircraft dynamics and control, wasn't very interesting because "it had a lot of theory," he recalled.
For example, one of the concepts he covered in the class was placement of the aircraft tail to define airplane stability. "Before I started using this new method, what I would do was derive the equations for the stability of the aircraft by putting the wing on the airplane, then the tail, and showing that placing the tail at such a distance from the wing would make it stable--all on paper," he explained. "While I could tell them, 'If you put it here, it's going to be stable, and if you put it there, it's not going to be stable,' they're not going to know it unless they work on the problem. And there are only so many combinations of designs you can test."
Under the new method, he didn't touch paper. Using MATLAB, Shankar created a program that would quickly allow him to visually demonstrate the impact on changes to the aircraft configuration. Shankar projected his program onto a screen in front of the group to show the impact of the plane's stability as the tail was moved from the wing backwards. "At some point it changes from unstable to stable," he said. "I'd tell them, 'Remember that.' Then I'd go back and show them the theory. It makes them say, 'Oh, I've seen this before. Now I know why it happens.'"
Trying to have students work out that same number of iterations on paper as part of homework isn't the answer, Shankar noted. "Just because you give them more homework doesn't mean they learn better."
MATLAB is both a design environment and the term used to describe the text-based language that people use in that design environment. The environment can work with a series of add-ons called toolboxes that expand the language to be application-specific. Simulink, used in a class taught by Wells, is a design environment that allows the user to graphically describe the behavior of a system and then use that specification to generate executable code.
Shankar said that MATLAB isn't a new addition to the course he's teaching. Previously, he'd use the application to "draw plots and other simple things, but it wasn't the fundamental tool used to teach dynamics and control."
Consistent Usage of Technology
Tom Gaudette, principal academic evangelist for MathWorks, said he would bet there's not an engineering program that doesn't teach the company's tools "at some level." What he says he has found unique in the work undertaken at Arizona State is the idea that software was introduced to students not for the sake of learning the basic concepts of programming, but because it's a program that could be used throughout a student's entire college experience. "When we talk about integrated curriculum, we're really talking about introducing a tool in the beginning of the student's progression through college and then using that same tool in the follow-on courses. This lets [that person] concentrate on just the topics that the course is really trying to teach," he explained.
At the same time, Gaudette added, the continual use of the software allows the student to work on more complex assignments. "Students use the tools in the particular course to solve the problems of that course, and that really enables the professors to be able to go further with the theory because they have an industry-grade tool in front of them that can solve these problems," he said. "Normally what happens is that [the instructor] simplifies the problem. With the right tools you can say, 'This is a real problem. This is what you will solve as an engineer when you get out in the real world.' It gets students excited about becoming an engineer."
Shifting the use of the software to the front end of the lessons at Arizona State has had multiple outcomes. Students have become more collaborative, "which is a good thing," Shankar said. As they're doing homework, for example, visualizing aircraft motion, they may come up with answers that are different from the answers their friends in class come up with--and they start talking to understand the discrepancies. "That's where you see that collaboration," he noted.
Along the same lines, Shankar said, students have become less reliant on the teacher for answers, because they can use the tool to help them quickly generate responses to whatever questions they have.
Along with MATLAB, Shankar also used Digital DATCOM, a computer program used by the U.S. Air Force to design aircraft, and FlightGear, an open source flight simulation game. In both cases the students use those programs within MATLAB. "MATLAB is ideal because it not only provides a simple framework to integrate various third-party software applications, but also allows users to execute them in a familiar environment," Shankar explained.
Measuring the Impact of the Change
As part of the grant the research team also included Jenefer Husman, an Arizona State associate professor of educational psychology, who was wooed to the project to evaluate the effectiveness of the new teaching approach. Her input on the project was invaluable, Shankar said. "When you implement anything new, it's hard to come up with a metric that tells you, OK, you've done a good job," he observed. "How do you compare two of the same classes--one taught the traditional way and one using the new method?"
Husman worked with the aerospace instructors to develop a range of metrics by which to gauge the effectiveness of the new instructional approach. She surveyed students throughout the semester to understand how they felt about what they were learning from the class. "[The surveys] asked questions like, 'Do you think you'll use this concept in industry?' Or 'Do you think you're going to use this in your next class?' It was more about what students feel about their own learning ability," Shankar said."
The outcome, while not dramatic, was still notable, he added, "which gives us confidence that maybe we're going in the right direction."
Then there were the final exam scores. The mean score of the finals in Shankar's course and another one taught by Wells, under the traditional mode was 67. The mean score under the new approach was 79, a 12-point improvement.
"You need to take those numbers with a grain of salt," Shankar noted. As he repeated, he's still a fairly new instructor, and he'll continue improving. But he says he believes under the new mode of instruction the students are more engaged in class. "There were certainly a lot more questions from students when I was using the new method," he said. Plus, he says he expects those juniors who were "test pilots" for the new approach will go into their senior year wanting to use the same software for their senior designs.
Shankar said he doesn't anticipate rushing ahead and implementing the new approach in other classes that he's teaching. "At this point, I want to perfect this," he said. "But, yes, the idea is that we'll take this forward."
MathWorks' Gaudette says he is surer of the outcome. "One of the problems in engineering education is that faculty are constantly being asked to increase the breadth of things that they're teaching these students. At the same time, they're supposed to provide students with deeper knowledge in different technologies. It's really pretty hard to do both," he pointed out. "You can't get people to know more about different subjects and also be deeper in one particular subject without making your program longer--like going from four years to five years. So everybody tries to find the place on that curve that does the best. Our belief is that by including the necessary tools from the industry in the curriculum, you can take that whole curve and shift it, allowing you to both increase the depth and the breadth of the students without increasing the time in a particular program. When you have the right tool to solve the problem, the students are going to get it, and then they'll want to get another one and another. That's what the integrated curriculum is going to provide the students. When they graduate, they'll have the right tools in their toolbox to pull out and solve problems that they'll see in the real world."