Drivers of student performance: North American insights

Drivers of student performance: Insights from North America

Our analysis of OECD PISA results from North America highlights four insights on student mindsets, teaching practices, instructional time, and early childhood education.

Policy makers, educators, and parents across North America want to raise a generation of students who can thrive amid the relentless change wrought by technology and globalization. Yet improving educational outcomes has proved elusive. Some countries, states, and municipalities have made great strides, but many continue to struggle. Educators continue to debate what matters and what works.

In this report, we take a data-driven approach to consider a few of the most active debates in North America today: around mindsets, teaching practices, the length of the school day, and early childhood education. Our data come from the Programme for International Student Assessment (PISA), administered by the Organisation for Economic Co-operation and Development (OECD). Since 2000, the OECD has regularly tested 15-year-olds around the world on mathematics, reading, and science. The most recent assessment in 2015 covered more than half a million students across 72 countries, including nearly 30,000 students in North America. What makes PISA particularly powerful is that it goes beyond the numbers, asking students, principals, teachers, and parents a series of questions about their attitudes, behaviors, and resources.

The report’s findings include the following four insights:

Student mindsets have triple the effect of socioeconomic background on student outcomes.

It is hardly news that students’ attitudes and beliefs—what we term their mindsets—influence their academic performance. But how much? And which mindsets matter most?

By analyzing the PISA data, we found that in North America mindset factors explain a greater proportion of a student’s PISA score (at 37 percent) than even the home environment and student socioeconomic factors combined (at 12 percent), controlling for interactions between variables (Exhibit 1).

Our research shows that mindsets eclipse even the home environment in predicting student achievement.

Some mindsets are more important than others. In the 2015 PISA assessment, the most predictive mindset is the ability to identify what motivation looks like in day-to-day life (including doing more than expected and working on tasks until everything is perfect). We call this motivation calibration, as it involves a student “calibrating” what types of behaviors motivated students.

Other mindsets that are predictive of student outcomes include having low test anxiety and believing that one’s school science work will be useful for one’s future career. We also found that students with a strong growth mindset (those who believe they can succeed if they work hard) outperform students with a fixed mindset (those who believe that their capabilities are static) by 16 percent.

To be clear, mindsets alone cannot overcome economic and social barriers, and researchers debate the extent to which parental or school-system-level interventions can shift student mindsets. Our research does, however, suggest that mindsets matter a great deal, particularly for those living in the most challenging circumstances.

Students who receive the right blend of inquiry-based and teacher-directed instruction have the best outcomes.

High-performing and fast-improving school systems require high-quality instruction. We evaluated two types of science instruction to understand how different teaching styles affect student outcomes. The first is teacher-directed instruction, in which the teacher explains and demonstrates scientific ideas, discusses questions, and leads classroom discussions. The second is inquiry-based teaching, which includes a diverse range of practices, from conducting practical experiments to understanding how science can be applied in real life to encouraging students to create their own questions.

Our research found that student outcomes are highest with a combination of teacher-directed instruction in most-to-all classes and inquiry-based teaching in some classes (Exhibit 2). If all students experienced this rebalanced blend of instruction weighted in favor of teacher-led instruction, average PISA scores in North America would be 4.4 percent (or 22 PISA points) higher, equivalent to more than half a school year of learning. Currently almost half of North American students are receiving too little teacher-directed instruction.

Finding the 'sweet spot': The best student outcomes combine both teaching styles.

It’s also important to note, moreover, that some kinds of inquiry-based teaching are better than others. More structured inquiry-based activities yield higher PISA scores. Understanding how a science concept can be applied and conducting and drawing conclusions from scientific experiments improve scores significantly. Less structured methods of inquiry, however, such as allowing students to design their own experiments, result in lower scores across the board.

Given the strong conventional support for inquiry-based pedagogy, this seems counterintuitive. We offer two hypotheses. First, students cannot progress to inquiry-based methods without a strong foundation of knowledge, gained through teacher-directed instruction. Second, inquiry-based teaching is inherently more challenging to deliver, and teachers who attempt it without sufficient training and support will struggle. Better teacher training, high-quality lesson plans, and school-based instructional leadership can help; so can giving principals and teachers the confidence to focus on fewer incidences of well-planned inquiry, rather than trying to use these methods exclusively.

While teacher-directed instruction has the most positive impact on PISA scores, inquiry-based practices do better in promoting students’ “joy in science” and instilling the belief that doing well in school will help them have a brighter future. We believe that is why blending teacher-directed instruction with inquiry-based teaching produces the greatest overall benefit across North America.

Increasing the school day can improve student outcomes, but significant gains can also be made from using existing time better.

Across the United States (the sole country focus for this finding), federal, state, and district officials have been considering the merit of longer school days, pushed to some extent by advocacy groups such as the National Center on Time & Learning.

According to the PISA data, student outcomes improve by 3.9 percent for each additional half-hour of instruction, up to 6.5 hours per day. If all students performed at the level of those currently receiving 6.0 to 6.5 hours of instruction per day, this would boost average science achievement by approximately 4.4 percent (or 22 PISA points). But extending the school day is expensive and has knock-on effects on commutes, after-school activities, and school start times. It’s also worth noting that some European systems—such as those in Finland and Germany—achieve superior results to the United States despite having a shorter average school day. This suggests that lengthening the school day is only part of the answer. Improving the quality of every hour in school also remains critical.

Early childhood education has a positive academic impact in most regions, but not in the United States.

Many studies have shown that high-quality early childhood education (ECE) improves social and academic outcomes, producing gains that sustain for many years. But research also shows that low-quality early childhood programs produce neutral outcomes at best and can be detrimental at worst.

Our findings suggest that ECE is indeed having a positive impact on outcomes at age 15 in almost every region in the world. For example, students who had some ECE score 19 to 21 PISA points higher in Asia and Europe, controlling for student socioeconomic status, school type, and location. Canada reflects this global trend with a 15 PISA point (or 3 percent) lift for children who received ECE.

In the United States, however, students with ECE actually score 8 PISA points (or 2 percent) lower than those without. What is going on with ECE in the United States? While the PISA data cannot provide an answer, it does provide hints. For example, it seems that children with a lower socioeconomic status are the ones who are benefiting the least. It also seems that children who start ECE at age one or younger score dramatically lower on PISA than those who start later. This suggests that there could be a significant ECE quality problem in the United States for low-income and very young children, potentially linked to the lack of government provision compared with many other developed nations.


As we share these four findings, we are mindful of their limits. One cannot construct definitive answers from a single source, no matter how broad or well designed. The direction of causality, sample sizes, missing variables, and nonlinear relationships are all potential issues. Many questions still need to be resolved through a thoughtful research agenda and longitudinal experimentation. That said, we believe that these four findings provide important insights into how students succeed—and that North American educators should heed these lessons as they develop learning agendas and school improvement programs to deliver the progress that their students deserve.

Download the full report on which this article is based, Drivers of student performance: Insights from North America (PDF–12 MB).

About the author(s)

Jake Bryant is an associate partner in McKinsey’s Washington, DC, office, where Paul Kihn is an advisor, Mona Mourshed is a senior partner, and Jimmy Sarakatsannis is a partner; Emma Dorn is a practice manager in the Silicon Valley office; and Marc Krawitz is an associate partner in the New Jersey office.

The authors deeply thank the many people who supported us in bringing this report to fruition. We are grateful for the invaluable guidance of our analytics leadership: Rafiq Ajani, Taras Gorishnyy, and Sacha Litman. We thank our dedicated engineer and data-scientist colleagues: April Cheng, Sujatha Duraikkannan, Roma Koulikov, Devyani Sharma, and Avan Vora; and we are grateful for the substantial contributions from our colleagues Anne-Marie Frassica, Joy Lim, Esteban Loria, Miriam Owens, Corinne Spears, Amy Tang, Rachel Valentino, and Paul Yuan. We further acknowledge the external thought leaders and experts who provided counsel and expertise. Finally, this report would not have been published without the support of our editor Cait Murphy, and the design creativity of Nicholas Dehaney at Spicegrove Creative.

Related Articles