New Paper: Mindset Interventions Are a Scalable Treatment for Academic Underachievement

When we started PERTS in 2010, our first goal was to figure out whether mindset interventions could be used widely to help students become more motivated and successful in school.

The results of our recently published study suggest they can. We’re currently working to replicate these findings with larger, more representative samples in preparation for making the interventions freely available to all schools. In this blog post, we describe what inspired the recently published study and what we’re doing to build on it.

Why did PERTS run the study?

When we ran this study in 2011, several scientific articles already showed that carefully designed interventions could help students develop a growth mindset and become more motivated and resilient learners by doing so. In these studies, students who were taught that intelligence is something they could grow earned higher grades (Aronson, Fried, & Good, 2002; Blackwell, Trzesniewski, & Dweck, 2007) or test scores (Good, Aronson, & Inzlicht, 2003) compared to randomly assigned control groups. With results like these, the then-naive founders of PERTS (Carissa Romero, Ben Haley, Chris Macrander, and I) wondered: Why isn’t every school using growth mindset interventions to improve student outcomes?

We quickly discovered two good reasons schools weren’t implementing mindset interventions:

1. Schools didn’t know how to implement mindset interventions, and
2. Schools didn’t know whether mindset interventions would work for their students.

We had something important in common with them: We didn’t know either! To turn mindset interventions into something that schools could (and should) practically use, we first needed to develop a mindset intervention that schools could easily implement. We also needed to test whether this easy-to-use intervention was effective for various kinds of students.

What did we do?

We built two internet-based mindset interventions:

1. A growth mindset intervention, designed to instill the idea that intelligence is something you can grow, and
2. A purpose for learning intervention, designed to instill the idea that learning in school can help you become an empowered adult — a person who can have a meaningful impact on the people and causes they care about.

Each intervention was built as a short (less than 45 minute) online module. Each online module involved students reading and writing about the ideas at the heart of the intervention. For example, for the growth mindset intervention they read about intelligence being something they can grow, and they wrote about these ideas in their own words. Once we built the modules, we recruited schools to help us test out whether they would actually work.

We got a sample of over 1,500 students from 13 heterogenous high schools (8 public, 1 private, and 4 charter schools, with free and reduced lunch rates ranging from 0% — 90%), and we collected participating students’ academic transcripts from before and after the study. Each student was taken to the school computer lab by their own teacher (PERTS staff never physically went to any of the schools) and then signed into the study website. At that point, each student was individually randomly assigned to a control condition or an intervention. The students came back for a second online session about two weeks later, at which point we assessed changes in their mindsets.

We found that the interventions successfully changed how students thought about school: Students became more likely to endorse the belief that intelligence is malleable (growth mindset) or to see schoolwork as meaningful (purpose for learning). The interventions also improved the grades of lower-performing students:

Students who earned GPAs of C or below prior to the intervention (a key risk factor for high school dropout) became 6.4% more likely to earn satisfactory grades following the interventions, and they earned higher GPAs across all four core courses.

These results showed that mindset interventions can be effective in a format that is easy for schools to implement and that they can be effective in a variety of schools.

This was an important finding for a number of reasons. First, it’s very rare for an educational intervention to have any detectable effect under realistic administration conditions (i.e., when tested outside the context of a closely monitored research trial in which expert researchers are on hand to make sure an intervention is administered “perfectly”). Second, these interventions are essentially free to administer to additional students both in terms of financial cost and in terms of class time because the intervention just involves spending 45 minutes on a website. Third, the internet-based nature of the intervention also means that they can be delivered almost anywhere in the country (over 99% of U.S. high schools have internet access). That means that, if the results of this experiment held up when scaled across the whole country, American high school students would pass an additional 1.18 million courses per year.

What’s next?

We’re excited about this breakthrough, but it’s only one step. These results are limited in two important ways that we’re currently working to address.

First, our sample was too small (n=1,594) and non-representative to make us confident these interventions would work everywhere (though we’re much more confident than we were before this study that they would at least work in a lot of places!). To expand our understanding and to increase our certainty about where these interventions work, PERTS has been working with the newly formed Mindset Scholars Network to replicate these findings with larger, more representative samples of students.

A replication study conducted in Fall 2014 was successful, and it is in press right now at the Journal of Educational Psychology. We’re busy now helping a team of researchers run yet another replication — this time with a sample of over 10,000 9th graders drawn from a nationally representative sample of schools. Once we figure out where this intervention is reliably effective, we will make it freely available to all schools.

Another limitation of this study was the short, online format of the tested intervention. That format had some important benefits: It enabled us to test the effects of two mindset interventions on a large scale and at a low cost very quickly (because the internet makes it easy and cheap to get a carefully crafted message to a lot of students regardless of where they live). However, it’s hard to imagine that these brief, online modules could ever be as effective as helping teachers support a growth mindset and a purpose for learning in class every day of the school year. After all, our interventions were delivered directly to students through a 45-minute online exercise, while the average American high school student is in class for 74,400 minutes a year. Surely students would benefit even more if we could help teachers weave these powerful ideas into their teaching.

That’s why we’ve been partnering with teachers to develop, identify, and promote classroom practices that build students’ mindsets and enhance their motivation and resilience. It’s still early days for this work, but you can see — and contribute — to our progress at mindsetkit.org. We hope you’ll join us!