From the Field

A Proof Point for the Value of Growth Mindsets

Researchers have long suggested that changing a student’s mindset toward learning can influence academic performance. A new study, released in Nature magazine, offers proof that a low-cost intervention encouraging a growth mindset, a resilience-engendering belief that abilities can be strengthened through hard work, can improve grades. The National Study of Learning Mindsets—which involved 12,000 9th graders in a randomly selected, nationally representative sample—also explored the types of schools and school climates where the intervention worked best. FutureEd spoke with David Yeager, a University of Texas professor who led the team of two dozen researchers, about the implications for schools and education research. 

What did your study involve? 

The intervention involves two self-directed, online activities that each lasts about 25 minutes, and that students are invited to complete toward the beginning of their 9th grade year. Students go to the computer lab, they log in, they read some information about the brain. First that the brain can be developed like a muscle. And second, that the brain’s connections grow stronger whenever students learn from rigorous activities in school. And then during the exercises, students write a letter to a future 9th grade student coming to high school who might be struggling sharing the knowledge that the brain can grow and be developed like a muscle.

Did the intervention work?

We found that kids who were assigned to read and complete the growth mindset activities at the beginning of high school reported more of a growth mindset, and then were more likely to earn a higher grade point average at the end of freshman year than the control group. And they were less likely to have earned an average of D or less in their core courses—math, science, English, and social studies. That’s hugely important because when students fail to complete 9th grade with a passing GPA, they are far less likely to graduate high school in time. We found a 3 percentage point reduction in the number of students with a D-or-less average, which is striking, because that reduction applies to 3 million students in high school each year. Thousands of students might benefit from this light-touch, low-cost intervention.

You’ve called this study “a major milestone for science.” That’s a bold claim. 

It’s showing definitively that growth mindset has important effects on educational outcomes. And that growth mindset can be used through relatively brief, scalable interventions to improve adolescents’ grades. Second, we identified the school contexts in which the effects were larger or smaller. That allows us to better target the intervention both to students and to schools. Third, this is the first time an education intervention for adolescents has been tested using a nationally representative sample and involving multiple methods that allow us to trust that the results are valid. Other experiments down the road could use the methodology that we set out. We think it could lead to a sea change in the way that education research happens.

You say you’ve identified the school contexts where the interventions work best. Did the impact differ based on students’ prior achievement levels?

We didn’t think the intervention would cause students with straight As to have even more As. And that’s what we found. Within the group of low-achieving students, defined as at the 50th percentile or below based on prior grades, we discovered some very interesting patterns. We identified the kind of schools where low-achieving students’ grades are lifted and the kinds of schools where they’re not.

What are the characteristics of these schools?

We asked, “What are a few things that we think have the best shot at explaining where intervention has significant effects and where it has weaker effects?” Our colleagues who are specialists in the sociology of education have led us to think that two kinds of factors of schools could matter. First is what we call the school’s formal resources. That refers to the school’s curriculum, the quality of the teaching, the finances of the school. And we started with a hypothesis that in schools where there’s an overabundance of resource and advantage, students’ grades wouldn’t be affected by the intervention because they’re getting so many supports from other sources.

Was that the case?

In the higher achieving schools, the top 25 percent, where student presumably had many resources supporting them, we didn’t find effects on grades. In the other 75 percent of schools where presumably the treatment was more needed, we did find improvement on grades. Still, we think our intervention was valuable in the high-achieving schools. We found that students in the top 25 percent of schools were inspired to take harder math classes in 10th grade.

You also looked at how much support those schools were giving these students. How did you measure that and what did you find?

The second factor that matters is the informal resources of the school—the culture and climate. Our sociologist colleagues  hypothesized that the intervention would be less likely to persist over time if the norms in the school were against it. So if the treatment tells you that you grow your brain by trying hard math assignments, but then your peers make fun of you for raising your hand and actually doing homework, then maybe the treatment won’t work. But if the norm is the school is supportive, if peers value high achievement and motivation and learning and the desire to get ahead, then maybe that is the kind of place where a small message about the benefits of learning can find fertile soil.

Did you find that was true?

What we found is that the largest effects were in schools where the peers were really supportive of growth mindset ideas. And the average student effect is strikingly large when you combine low-achieving schools with supportive peer norms.

There was a reduction of over 8 percentage points in the number of students ending 9th grade with D or F averages—a drop from 45 percent of students at that level to 37 percent. An 8 percentage point reduction is striking. Early college high school, which is a well-known reform that puts college teachers into high schools, improved the rate of 9th graders on track for graduation by about 3 percentage points. And it costs us more than $5,000 per student. Here, we have a treatment that most likely cost under $1 per student and is having larger effects in some schools. This doesn’t mean that a growth mindset on its own is all that is needed. But it’s exciting to imagine what could happen if we combined curricular reforms with a growth mindset approach, as programs such as the Carnegie’s Foundation’s Statway have done.

You chose 9th graders for this intervention. Why 9th grade?

Ninth grade is a real turning point for young people. They are often confronting new and more rigorous academic standards. Their grades start to count for college admission and applications, which means the stakes are higher. And they have to face these higher standards with an uncertain support network because their friends, teachers, counselors and mentors often change as they move from middle school to high school. And all of that uncertainty can add up to stress and difficulty, which we think makes adolescents especially likely to benefit from a treatment that tells them that the struggles they’re facing are normal and can improve.

Freshman year is a time when kids start developing new habits and start getting into new course tracks that are more difficult. And so the right motivational boost at the right time might instill either habits or course choices that would have the chance to be self-reinforcing throughout high school.

[Read More: Reducing the ‘Toxic Stress’ of Starting High School]

In your intervention, why is it important that the students write about it to another student? 

Adolescents don’t tend to internalize information when they’re just told it by adults and told that they should believe it. Instead, we think it’s more powerful to invite young people to try to persuade others of the truth of some idea. First, because we know teaching someone else something causes us to learn it better ourselves. Second, because often the arguments that we generate to persuade someone else tend to be the ones that would have persuaded us the most.

Given your results, what are the implications for schools? Should we be instituting this in every school?

I think there are three implications. One is that the enthusiasm and work put into growth mindset is, at some level, justified and that could and should continue, although with some qualifications about how it can be done better and in concert with traditional school improvement.

The second is that, for the first time, we have an intervention that has been evaluated in a truly generalizable way, and that can be delivered at scale. And so I don’t see any reason why U.S. public high schools couldn’t use this right away as a part of freshman orientation. It’s available at None of the researchers involved will charge for that or have any financial conflict of interest. We just think it’s a good idea to be used nationwide, given the strong results we’ve seen.

But an online intervention is not a magic bullet for motivational problems among adolescents. It’s the start of a process. We can’t just rely on the treatment to change the mind of the kid, and then do nothing ourselves as adults to make the schools ready to benefit from that treatment. Another way to say it is that a treatment is like a seed. It’s still up to us, as teachers, parents, superintendents, and so on, to till the soil so that the seed can grow.

What are teachers’ roles in this work?

We think that the new era of mindset research, that will hopefully be spurred by these results, will focus on teaching practices that support students’ mindsets. Teachers’ mindsets matter—influencing whether their students can grow and learn, structuring assignments so that students can actually put their growth mindset into practice.

If we tell a student that mistakes are your friend and you can learn from mistakes, but then the teacher never gives students credit for fixing the mistakes, how can the students grades go up even with the strong growth mindset? I hope that the study results not only justify more work on growth mindset and scale-up of the treatment, but motivates us as a field to create the kind of schools and culture and climate that help young people.

[Read more: Teacher Mindsets: How Educators’ Perspectives Shape Student Success]

You mentioned earlier that this research could spur a sea change in the education field. What do you imagine that looking like?

The typical approach in educational research is to hand-select schools that are maybe down the street from the university where you’re doing a study, or in one large suburban or urban district where you can quickly and easily pull the grades for lots and lots of students. Because our interest was in understanding how the treatment styles vary from school to school, we didn’t do that typical approach. Instead, we generated a list of all 12,000 public high schools and randomly selected a subset of them, visited them, and asked them to join our study, with 65 ultimately participating in the project.

Just to illustrate how unique that sample is, a recent evaluation of the presence of rural schools in education trials found that only 4 percent of schools in previous randomized trials have been from rural areas. In our study it’s more than 42 percent, which is far closer to the national average for high schools.

We can’t presume that an intervention that shows effects in suburban Florida will generalize to rural South Dakota. At least some of the times, we need to do the randomized experiments with random samples so that we can justify the implementation of interventions at scale.

What’s next in mindset research?

The national study is the end of the old era of mindset research, which asked questions like “Is mindset a real effect? Can it change grades? Can it be replicated?”

The new era is the era of understanding how interventions work in different contexts, the era of collaboration across disciplines, the era of large data sets and partnerships with practitioners. And we’re excited to see where that goes. Researchers are pursuing this new work through an early career fellowship program supported by the Bezos Foundation and the Mindset Scholars Network. And in the next 12 months, we hope to have a data set available to any scholar who would like to work on mindsets.

Photo courtesy of Allison Shelley/The Verbatim Agency for American Education: Images of Teachers and Students in Action.