Students who learn to see their minds as muscles can better flex their brains for classroom success.
That’s the finding of a landmark research project. The National Study of Learning Mindsets examined how a brief online program encouraged students to adopt a growth mindset — and affected their grades in early years of high school. The researchers focused on that critical period of education when teens’ performance “is a predictor of their future educational attainment, financial success, health, and well-being,” according to the research summary.
Texas McCombs researchers were among those involved with the study. In his role, Jared Murray, assistant professor of Information, Risk, and Operations Management, took the collected data and examined the results, while guarding against results that weren’t what the researchers initially set out to find. That approach helped confirm the strength of their findings.
The research differentiated between a growth mindset versus a fixed mindset. What’s the difference?
The basic idea is whether or not you think that you have the ability to acquire new skills. In the mindset study, we were thinking about math. There’s that stereotype, “I’m just not a math person.” A long line of research on mindsets shows a particularly pernicious effect for mathematics. Are you a person who is not a “math person”? Or are you not a math person yet? If you take the time and put in the work, you can get to a point where you’re doing well at math.
How did you encourage students to break that stereotype?
At the beginning of ninth grade, the treatment group read and listened to material in an online module about how the brain is plastic and has the ability to change. The students were presented with the metaphor “Your brain is like a muscle — the more you work it, the stronger it gets”. That explanation was very clever, and is engaging to a high school student. They were asked to think about positive learning strategies beyond just working harder, like actively seeking alternative approaches from teachers and peers to solving a difficult math problem.
Then the students did a reflection exercise. They wrote a letter to a student struggling to develop this growth mindset. Having our participants go through that reflection exercise and reinforce what they’ve learned was really important, too.
It’s remarkable that this light touch intervention has such dramatic effects. On average, treated students had higher GPAs at the end of ninth grade. Those treatment effects were larger for students who were in lower achieving schools.
Why is that?
One possibility is that students who are in higher achieving schools are just in better environments with more resources, so struggling students in those schools who didn’t receive the treatment are likely to get other interventions that keep them on track. Another is ceiling effects: if you applied this intervention to a student who would get an A without the intervention, they can still only get an A. In the higher achieving schools, we have more students who are at or near that ceiling, so if we’re looking at an outcome like GPA you’d expect to see smaller effects. Students in higher achieving schools also often already have higher levels of mindset beliefs before the intervention, on average, but another finding from the NSLM is that having higher levels of peer mindset norms actually make the intervention much more effective in lower achieving schools.
The study also encouraged students to enroll in harder classes.
If you look at whether students take advanced math, we find that the mindset intervention significantly improves that across the board.
Once the data collection was nearly complete, what role did you play?
The greater context is that we’re in the middle of a replication crisis. It’s particularly acute in the social sciences. When researchers try to conduct a new study testing the same hypotheses, they’re often not able to replicate the results of the first one. There are a lot of reasons why that could be happening. One is if you’re coming up with your hypotheses after seeing your data. Maybe there’s little or no evidence for your original hypotheses, so maybe now you go back to the data and looking for evidence of other relationships between variables. You might iterate through that process multiple times.
As statisticians, we know if you repeat that loop enough times, you’re going to find something more or less by accident. Every time you test a hypothesis, there’s some chance that you’re going to get a false positive result, and that chance increases as we do more tests. The situation is even worse when you’re using the data to suggest interesting relationships between variables to form your hypotheses. One of the guards against that is before you collect any data, you write down all the different hypotheses you want to test, the relationships you want to estimate, and precisely how you want to do all of that. That’s helpful — if that pre-registration plan didn’t exist, then nobody would really know if you actually had this hypothesis going in, or if instead you sneaked a peek at the data and changed your analysis based on the data.
But it’s also limiting. The pre-registration for the NSLM is almost 200 pages long, and thinking about every analysis you want to do, how you’re going to define every subgroup, how you’re going to handle every contingency in data collection, it’s very difficult. It also limits your ability to be surprised by the data. My work is focused on developing modern statistical and machine learning methods for easing the pre-registration burden and also allowing for more unstructured discovery from data, while guarding against false or overconfident discoveries, data snooping, p-hacking, and other contributors to the replication crisis. In the NSLM, our methods give essentially the same inferences as the carefully designed pre-registration plan, and I could pre-register our new method in a few pages.
So that approach strengthens your findings about the importance of mindset. Can adults benefit from a similar approach?
The intervention was carefully designed over a long period. It had been to lots of pre-testing, and it was particularly tailored to modern beliefs in high school students. Teenagers can be sort of immoveable objects when it comes to some things. We had to be clever in designing the intervention.
If you wanted to build your own growth mindset beliefs, I’d think of the metaphor used in the mindset study: your mind is like a muscle. When you’re working on something and getting stuck, that doesn’t mean that you’re dumb or can’t do it. It means you’re pushing the boundaries of your abilities.
Story by Jeremy M. Simon