Lesson learned? Analysis finds lectures still dominate STEM education

An analysis of more than 2,000 college classes in science, technology, engineering and math has imparted a lesson that might resonate with many students who sat through them: Enough with the lectures, already.

Published March 29 in the journal Science, the largest-ever observational study of undergraduate STEM education monitored nearly 550 faculty as they taught more than 700 courses at 25 institutions across the United States and Canada.

The University of Nebraska-Lincoln’s Marilyne Stains and colleagues, including Michelle Smith, MacKenzie Stetzer and Erin Vinson from the University of Maine, found 55 percent of STEM classroom interactions consisted mostly of conventional lecturing, a style that prior research has identified as among the least effective at teaching and engaging students.

Another 27 percent featured interactive lectures that had students participating in some group activities or answering multiple-choice questions with handheld clickers. Eighteen percent emphasized a student-centered style heavy on group work and discussions.

The predominance of lecturing observed in the study persists despite many years of federal and state educational agencies advocating for more student-centered learning, the researchers said.

“There is an enormous amount of work that has demonstrated that these (student-centered) strategies improve students’ learning and attitudes toward science,” says Stains, the study’s lead author and associate professor of chemistry at Nebraska.

“It’s not just that they understand it better, but they also appreciate science more. They’re not as scared of it, and they engage more easily with it. When you see that kind of effect, it makes you say, ‘Why are we still doing it the other way?’”

One potential culprit captured by the study: Though smaller class sizes and open classroom layouts often are considered essential to student-centered learning, they do not guarantee its adoption.

Lectures did occur more often in larger than smaller classes, the study found, and open layouts did correlate with more student-centered learning. But about half the courses with those advantages still featured more conventional lecturing than interactive or student-centered teaching styles.

“When you talk to faculty, you often hear, ‘I teach in an amphitheater. I could never do group work; it’s just not practical. But if I had a small class, I could do it,’” Stains says.

“But just because you have the right layout doesn’t mean you’re actually going to (promote) active learning. You need to be trained in those kinds of practices. If there’s not a budget for professional development to help faculty use those environments, they’re going to default to what they know best, which is lecturing.”

Smith agrees. “Faculty professional development is most effective if it is a long-term engagement where instructors have the opportunity to discuss and implement changes as part of a supportive community,” she says.

“Making sure there is support for these types of faculty learning communities is critical to promoting instructional changes.”

The study did show that many faculty adopt multiple teaching styles throughout a semester. Among the faculty who were observed at least twice, 42 percent demonstrated two styles. Based on its data, the research team concluded that three or four classroom visits are needed to reliably characterize an instructor’s approach.

“This outcome is important for thinking about how STEM departments evaluate the teaching portion of tenure and promotion applications,” says Smith. “Faculty should be evaluated based on a combination of observations from three or four class periods.”

Much of the previous research into STEM instruction has relied on surveying faculty about their practices. Though the resulting data has proven valuable, Stains says, the flaws of human memory and perception inevitably find their way into that data.

“Surveys and self-reports are useful to get people’s perceptions of what they are doing,” she says. “If you ask me about how I teach, I might tell you, ‘I spend 50 percent of my class having students talk to each other.’ But when you actually come to my class and observe, you may find that it’s more like 30 percent. Our perception is not always accurate.”

So the research team decided to observe and document STEM classroom practices with a commonly used protocol that Smith helped to develop. That protocol involved documenting 13 types of student behavior and 12 types of instructor behavior that were codified for every two-minute interval throughout a class.

An analysis based on the prevalence of four student and four instructor behaviors allowed the team to identify seven instructional profiles, which were then categorized into three broad teaching styles.

Those efforts also led to the creation of an app that runs essentially the same analyses conducted for the new study, giving users a prompt breakdown of teaching styles and the frequency of all 25 behaviors captured by the observational protocol.

“People can do their own measurements and see how they compare to this large dataset — see how either their department or college is doing — and say, ‘This is where we stand. This is where we want to go,’” says Stains.

Smith says even though she has been involved in this research and is familiar with observation protocols, the data has helped her as an instructor.

“I was once observed teaching a population genetics topic and the observation data revealed that I lectured a lot more in this class period compared to my other observations,” she says. “I was shocked how different it was and these data were a wonderful incentive to change that class.”

In the meantime, the study’s scale and interdisciplinary nature make it a “reliable snapshot” of how STEM gets taught to undergraduate students in North America, its authors say.

“There are many universities that are interested in integrating student-centered practices into their undergraduate STEM curriculum,” Stains says. “This could give them insights about what’s probably going on in their classrooms if they’re at a research-intensive institution.”

Colleagues from Auburn University; Simon Fraser University; the University of British Columbia; the University of Colorado Boulder; the University of Iowa; Armstrong State University; the University of California, Los Angeles; Otterbein University; the University of California, San Diego; the University of Michigan; the University of Calgary; the University of Virginia; and St. Mary’s University (Halifax) also took part in the study. The research team received funding in part from the National Science Foundation and the National Institutes of Health.

Contact: Scott Schrage, University of Nebraska–Lincoln, 402.472.4206; Beth Staples, 207.581.3777