Heat isn't just a nuisance to students; it's a barrier to learning

Heat isn't just a nuisance to students; it's a barrier to learning
© Getty Images

Less than a month ago, a heat wave in the Midwest forced schools in Minnesota, Michigan, Illinois and other states to cancel classes or dismiss students early. To help with the heat, officials in St. Paul gave students water bottles and delivered hundreds of fans to schools, most of which lack air conditioning.

This example, though extreme, points to a larger pattern. Most teachers and administrators know that sweltering classrooms are both physically dangerous for students and not conducive to learning. Overheated students and teachers become fatigued, lose focus and find it hard to do their educational work well.


In our recently released research, "Heat and Learning," we quantified just how bad heat might be for student learning. To do so, we examined the PSAT/NSMQT scores of 10 million students across the United States from the high school classes of 2001 to 2014.


We focused on students who took this standardized test twice, once in the fall of 10th grade and again in the fall of 11th grade. By doing so, we measured a typical student’s improvement in knowledge, as measured by math and verbal scores, over one year’s time.

We then used federal weather sensors to measure the temperature experienced by students on school days in the year leading up to each exam. Finally, we asked the question: Does a student’s PSAT score improve by a smaller amount following a hotter school year? We reported three remarkable findings.

First, hot school years substantially impede student learning. On average, each 1 degree Fahrenheit increase in the temperature of the school year prior to the exam causes students to learn about one percent less than they otherwise would have.

On average, extremely hot school days (those in the 90s and above) reduce learning roughly twice as much as somewhat hot school days (those in the 70s and 80s). Unlike school-day heat, heat on weekends and during the summer has no discernible effect on PSAT scores, suggesting that heat directly interferes with the effectiveness of instructional time.

Second, these average impacts of heat mask large differences by income and race. For students living in the poorest 20 percent of ZIP codes, heat is three times as damaging to academic achievement as it is for students in the richest 20 percent of ZIP codes.

Similarly, the impact of heat on achievement is three times as large for black and Hispanic students as for white students. Hot classrooms thus appear to exacerbate educational inequalities.

Third, school air conditioning appears to largely solve the problem. To show this, we surveyed tens of thousands of students and school counselors to construct the first nationwide measures of air conditioning by high school.

Fewer than 60 percent of students reported that their classrooms were adequately air conditioned, with particularly low rates of school air conditioning in northern, cooler parts of the country.

Students in schools that don’t have air conditioning see their test scores nearly five times more negatively affected by heat than students who learn in air-conditioned schools. The schools attended by low-income students and minority students are less likely to be air conditioned, partly explaining why heat has larger impacts on such students.

That the physical comfort of students and teachers affects the learning process is not terribly surprising. Why then don’t all schools have air conditioning? Such infrastructure is clearly expensive to install and operate, especially in older buildings.

In 2017, New York City allocated $29 million to install air conditioning in 11,000 classrooms within five years. Such costs must be paid up front while the benefits have been historically hard to quantify and come only in the longer run.

We estimate, however, that the long-run economic payoff to students’ increased learning would be substantially larger than such costs to schools. Such results suggest that we may currently underinvest in this particular aspect of school infrastructure, especially in warmer parts of the country.

What would those of us who work in offices do without air conditioning on a hot day? If you’re anything like we are, you’d get a lot less done and probably go home early.

Because our employers know this, they provide an air-conditioned environment that keeps us comfortable and functioning productively. What is surprising is our failure to apply these same expectations to the work environments of our children and their teachers.

Joshua Goodman is an associate professor at the Harvard Kennedy School of Government. Michael Hurwitz is a senior director at The College Board. Jisung Park is an assistant professor at UCLA. Jonathan Smith is an assistant professor at Georgia State University.