We’re Teaching Problem-Solving All Wrong
Engineers must be able to solve problems. What does that actually mean, and how do we best prepare students?
Opinion by Eric W. Burkholder and Carl Wieman
In typical engineering courses, students are taught concepts and equations and then asked to solve simplified practice problems to apply them. However, this approach eliminates key variables that exist in the real world—for example, the need to make decisions with limited available information. Thus, when graduates start their professional careers, they are prepared with content knowledge but often lack problem-solving abilities.
A fundamental issue in education is that we cannot teach what we cannot measure, and we cannot measure what is undefined. Therefore, we first need to know what it means to solve an authentic engineering problem—what actual cognitive skills are required—if we are to train better engineers.
We studied the detailed problem-solving process used by highly skilled scientists and engineers in their work across many different subdisciplines. Surprisingly, we found that all these experts framed their approach around a set of twenty-nine common decisions made with limited information. For example, What are the criteria and scope for a solution? What are similar problems whose solutions are relevant? What information is needed to solve the problem and how best to obtain it? How reliable is the information I have? What are the potential failure modes of a proposed solution? Skilled engineers use similar reasoning to make each decision, combined with their discipline-specific knowledge. Making this set of decisions with the limited information available lays out their path to a good solution.
Cognitive science tells us that to develop expertise a student must practice and master the desired skills (in this case our newly identified decisions) with careful feedback from an instructor. As it stands, undergraduate students receive very little practice in decision-making. Problem-solving in these courses often centers around textbook-style problems that make many of the decisions for the students, such as identifying the goal of the problem, determining the information needed, and even establishing the concepts that apply. Our work measuring the problem-solving skills of engineering students provides evidence for this lack of practice: even at top universities, graduating students’ scores on our problem-solving assessments range from 40 to 70 percent of the scores of practicing engineers.
Project-based courses offer more opportunities than standard content-based courses to practice making decisions (and learn from failed attempts). In reality, however, controlled types of decisions are made in these courses, and they often build upon previously learned content. In order for students to develop expertise, problem-solving must be practiced at the same time as new material is taught, so that students can learn what knowledge is needed to solve a problem and how to find it.
Our research suggests that content-focused courses must be transformed to teach effective problem-solving. For instance, in a fluid dynamics course, we might ask students to design an experiment that would allow them to figure out the pressure drop needed to pump a certain kind of fluid. They would need to decide on the goal (flow rate to pump the fluid), the information they need (friction factors, viscosity, density), how to measure that information, and the factors that might distort their measurements. In current courses, students might instead be given a flow rate and the physical properties of the fluid and pipe, and then simply be asked to use a given diagram to size the pump.
Students best learn problem-solving skills through consultation with their peers and regular targeted feedback from the instructor on the quality of their decisions and ways to improve them. Additionally, though it may seem counterintuitive, having students struggle with an authentic problem before instructors provide the relevant content is a more effective way for students to learn transferable problem-solving skills than providing the information first.
We may need to sacrifice some material to spend time solving authentic problems in content-focused courses, but in the long run students will be better educated. Most of the students’ future careers will require them to seek out and apply relevant new knowledge—their education should teach them that skill rather than simply furnish the content to them with no training on how to use it.
The problems facing the world today are too great to be tackled by students who only know how to solve textbook engineering problems. We must fundamentally shift the way we think about undergraduate education to prepare engineers for the challenges of the real world.
Eric W. Burkholder is an assistant professor of physics and adjunct assistant professor of chemical engineering at Auburn University with a focus in physics and engineering education research. Carl Wieman is a professor in the department of physics and the Graduate School of Education at Stanford University and is the 2020 recipient of the Yidan Prize for Educational Research and the 2001 Nobel Prize in physics.
Learn more about the authors’ research in CBE Life Sciences Education, “A detailed characterization of the expert problem-solving process in science and engineering: guidance for teaching and assessment,” https://bit.ly/3KqG98t.