In the classroom, failure can be more productive than success.
A new study published in Review of Educational Research finds that STEM students are more successful when learning a new concept if they engage in problem-solving followed by instruction rather than if they receive instruction and then engage in problem-solving. Students perform even better when the initial problem-solving is designed with the principles of productive failure in mind.
While the potential for productive failure has long been recognized in the education field, researchers were surprised by the extent of the benefit they saw in this study.
“These are breakthrough findings,” says the study’s lead author, Dr. Tanmay Sinha, a postdoctoral researcher and lecturer at ETH Zürich in Switzerland. “The effect size just by reversing the order of problem-solving and instruction is about two times that of what you would gain with an outstanding teacher over the course of a year. And if you implement productive failure principles, the effect is about three times stronger.”
Effective Productive Failure Design
For the study, Sinha and co-author Manu Kapur analyzed 53 studies that looked at STEM students globally. The students ranged from second grade to the postgraduate level, though a majority of the studies were conducted with students grades 6-10 from North America, Europe, and Asia.
Sinha and Kapur found successful productive failure strategies were guided by a few key principles. The type of question asked during the initial problem-solving exercise is important, Sinha says. “You won't pick up a standard textbook problem and just give it to students beforehand,” he says. “The problem has to be designed in a particular way, the language has to be nontechnical. It has to be intuitive, there has to be some kind of contrasting case so that students can compare between different data sets.”
Group work can also help make pre-lecture problem-solving more effective because it allows students to learn from one another, and then together they can increase their problem-solving abilities.
In addition, the type of instruction given after the initial problem-solving phase should not be a standard lecture. “If you give this kind of a problem to novices, they will make mistakes, and they will try out different ways of solving the problem,” Sinha says. “So the instruction then has to build on these incorrect solutions or sub-optimal solutions rather than just telling students, ‘This is the right way to solve it.’”
Productive Failure: Limitations and Takeaways for Educators
There are limits to productive failure’s efficacy. Problem-solving proceeding direct instruction did not hold the same advantage for second and fifth graders as it did for older students, and a direct instruction-first approach was equally as effective for teaching procedural skills.
“If you have to learn how to compute a procedure for say, standard deviation, instruction first will work as well as a problem-solving first design,” Sinha says. But for students to understand why they performed the steps to reach a final answer or to apply standard deviation to a different problem, beyond the current context, a problem-solving first approach appears to work much better.
Utilizing productive failure in your classroom does not require you to revamp your instruction. At Sinha’s university, he works with instructors to come up with strategies for incorporating the approach into their classes with small-scale changes. For example, many educators assign students to practice questions after a class, but he’ll ask them to assign the questions before a lecture with some tweaks to enhance their adherence to productive failure principles.
“The students are more prepped for the lecture that's going to follow,” he says. “It’s just shifting some of the exercises back, but you keep the structure of your class the same. You don't create radical changes, you don't go in and ask the instructor to completely flip what they have been doing.”