Steve Levitt, my co-author on the Freakonomics books and an economist at the University of Chicago, has spent a lot of time helping his four teenagers with their math homework in recent years. The experience has left Levitt with questions he can’t find good answers to. Questions like: Does anyone actually use this kind of math in their daily life? Is there any benefit, at all, to learning this stuff? And are there not more useful things they could be learning?
So, Levitt wrote up a grant proposal and convinced Schmidt Futures to give him funding to study those questions. In this guest column, Levitt tells us about his mission to modernize high school math.
I use mathematical thinking, statistics, and data analysis constantly, whether I’m writing economics papers, trying to get better at golf, or hoping to pick winners at the race track. But the math tools I actually use, and the math tools I see people around me actually using, seem to have nothing to do with what my kids are learning in school.
Jo Boaler, a professor of mathematics education at Stanford University and an expert on reforming math education, says that one of the biggest problems with the math curriculum in the U.S. today is that it’s a relic of a bygone era.
“It was a long time ago that somebody in the U.S. decided to teach what I think of as the geometry sandwich — a course of algebra for a whole year, followed by a course of geometry for a whole year, and then another course of algebra,” Boaler says. “I don’t know any other country that does that, and it’s part of the problem.”
In the United States, it was elite universities that first spurred the teaching of higher-level math. In 1820, for instance, Harvard began requiring knowledge of algebra to gain admittance. As a result, secondary schools started teaching algebra. Fifty years later, Harvard added geometry to its requirements, and the secondary schools followed suit.
After the Soviet Union launched Sputnik in 1957, math education became a matter of national security. The math curriculum was overhauled with two different goals in mind. The first was to increase the number of engineers, mathematicians, and scientists. This led to the introduction of higher-level, more abstract math in the high school curriculum. The second goal was to develop a workforce that could do the complex calculations required to support the military and space efforts. Remember, this predates the age of modern computers.
“In a way, we’re teaching kids to be computers, and they’re learning to do what computers do,” Boaler says. “So, high school, in particular, has lots of antiquated methods that students sit and work out by hand. They will never do that again in their lives. What kids can be doing that computers cannot be doing is creative, flexible thinking.”
Boaler has devoted her academic career to developing new ways of teaching math that generate creative, flexible thinking. Her work has had a massive impact. A website she helped develop, youcubed.org, translates academic research into actionable ideas for teachers to use in their classrooms. Youcubed has tallied 32 million page views over the last three years.
But actually changing how math is taught in American classrooms is difficult. Boaler knows this firsthand. During the early 2000s, she was working on implementing a new curriculum and found herself caught up in what’s known as the “Math Wars” — a debate between reformists and traditionalists over how math should be taught. The battle was fierce; Boaler says one opponent even staged a hunger strike.
“I interviewed some of the parents working to stop the new curriculum,” Boaler says, “and I remember one of the mothers saying to me, ‘You know, I’m not traditional about anything else, but if you can change math, anything can change.’”
The rate at which data are being created is mind-blowing. Every second of every day, enough data are created to fill 50 Libraries of Congress. Ninety percent of the data ever created by humanity was produced in the last two years. The labor market is having a hard time keeping up. According to LinkedIn’s 2018 Jobs Report, seven of the 10 fastest-growing job categories in the United States are data-centered: machine-learning engineer, data scientist, big-data engineer, full-stack developer, to name just a few. And these are well-paying, exciting, challenging jobs. And other in-demand jobs like software engineer and finance analyst require data fluency as well.
“When we look at the world out there and the jobs students are going to have, many students will be working with big datasets,” Jo Boaler says. “So, we haven’t adapted to help students in the most important job many people will do, which is to work with datasets in different ways. So, statistics is really important, as a course, but is underplayed. This is a fifth of the curriculum in England and has been for decades. But here in the U.S., it’s sort of a poor cousin to calculus.”
Most high school students are barely exposed to data: only 10% take a statistics class, and most statistics courses are primarily theoretical rather than requiring students to get their hands dirty with data. Nor is the American way of teaching math producing impressive outcomes. On the most recent Program for International Student Assessment (PISA), a standardized test administered in 70 countries, the U.S. placed 39th in math, just behind Hungary and Slovakia. American math performance is substantially worse than on either science (25th place) or reading (24th).
But it isn’t clear how we could ever transform a high school math curriculum that has proven so resistant to change. A natural place to start is by changing the incentives faced by teachers and schools. One of the strongest incentives is the high stakes testing done in almost every state. How students perform on these tests can determine school funding, teacher bonuses, whether a principal is fired, and even whether a school will be shut down.
David Coleman is the CEO of the College Board, the nonprofit that administers the SAT, PSAT, and Advanced Placement exams. The College Board has quietly been leading the charge on data fluency. “The really big new idea of the SAT is the only thing you’re allowed to put on it is that which is most widely used,” Coleman says. “So, we survey first-year math teachers and first-year college professors not only in math but outside of math, and we analyze which math is most used in their courses. That’s a knowable question.”
According to Coleman, there are just a few core math concepts that high school students need to understand: arithmetic (subtraction, multiplication, division, addition, and fractions); data analysis and problem-solving; and linear equations.
The College Board also started integrating data science into the reading and writing sections of the test. “There are five passages. Two of them are always a passage from science that includes numbers, data, and a passage from social science, like economics, that includes data,” Coleman says. “And so you can no longer be perfectly verbal without being able to read and analyze data from charts, tables, and graphs.”
For the record, everything Coleman says is true. Twenty percent of the SAT math questions test data fluency; and, amazingly, 10% of the questions on what used to be the verbal section are also data questions. A decade ago, those numbers would have been close to zero. The College Board has quietly been leading the charge on data fluency. And that matters, because the SAT is becoming an even more powerful force for change in the education system. Students have, of course, always cared deeply about their SAT scores. But the SAT is also starting to be adopted by a few states as their high stakes test for teachers and principals as well.
A few months ago, my team used the Freakonomics Twitter account to gather some data on my hypothesis that most of what we are teaching kids in high school math is pretty useless. We put together a survey about how often people use different kinds of math in their daily lives and sent it out to Freakonomics Radio listeners.
Our findings support my hypothesis: Only 2% of people said they used calculus daily, fewer than 2% said they used trigonometry daily, and about 4% said they used geometry daily. By contrast, almost 70% of respondents reported using Excel or Google spreadsheets on a daily basis.
We also asked respondents which subjects, both traditional math subjects and data-related math topics, they wished they’d learned more about in high school. Virtually no one mentioned calculus or trigonometry. On every single data-related topic, however, over 40% of our respondents said they wished they’d learned more. Almost 65% of respondents said they wished they’d learned more about analyzing and interpreting data to uncover hidden insights. Sixty percent of people said they wished they’d learned more about visualizing and presenting data to make an argument.
So, how might schools actually go about teaching these skills that people say they want? Here’s one example: In 2010, UCLA — in partnership with the Los Angeles Unified School District — received a National Science Foundation grant to design a high school course that would teach statistical thinking. It was piloted in the 2014–2015 school year with 10 teachers in 10 different high schools. The Introduction to Data Science (IDS) class covers just about everything you might want to do with data: creating a dataset, cleaning the data, analyzing it, and creating a visual presentation. The course even teaches how to design randomized experiments.
Saylee Garcia, a 12th grade student taking the course, chose to study crime rates in Los Angeles; she learned about more than just crime rates. “I’ve noticed how on the news, they keep on talking about crime. This is bad. It’s all negative,” Garcia says. “So most people are going to think Los Angeles is really unsafe. South Central is really unsafe. They don’t really show you crime rates have gone down in the last five years. And that’s what I like about this class because it shows me how to detect, like my teacher would say, liars.”
This school year, there will be 125 IDS classes taught. That’s an amazing accomplishment. But to fundamentally change things, that number of classes needs to be multiplied by 100 or even 1,000. It is hard work building to that scale from scratch. It struck me that the College Board, which is responsible for the Advanced Placement (AP) tests, might be able to use its enormous reach to jump-start this process. So I asked David Coleman whether the College Board had given any thought to an AP data-science test.
“I think the more profound thing we’ve done… is to include data science in the core exams we give like biology, like AP Government… to make data analysis something you encounter over and over again…” Coleman says. “If I weave data analysis into AP biology that’s widely given, or if we weave it into AP Government and Politics, which 400,000 kids take, that will touch kids in public schools in all levels of our society.”
About this PODCAST