Forum on Education in America

November 11, 2008
Prepared remarks by Bill Gates, co-chair and trustee

Editor's note: Read Melinda French Gates' comments at the Forum on Education, which preceded Bill Gates' remarks.

Good morning. Big advances only come when committed people study the same problems and build on each other’s work. It accelerates discovery, and I’m optimistic about what all of us can accomplish together.

Melinda and I believe that providing every child with a good education is the only path to equality in America. A good education means completing a postsecondary degree. And yet when we began our work eight years ago, the level of high school dropouts made even starting a postsecondary degree impossible for millions of students.

We were determined to find ways to work with our partners to turn around rising dropout rates, and increase the number of high school students who graduated from high school ready to succeed in college. We hoped that if we could build a model of a high-achieving school, it would be picked up by other schools. So we focused on 8 percent of schools, hoping that the lessons from our work in the 8 percent would scale to the 92 percent.

As Melinda said, we are determined to follow the evidence. So let me describe what we’ve found, what we make of it, and what we’re going to do about it.

There were some highly encouraging results—but I’ll start with the disappointments. In the first four years of our work with new, small schools, most of the schools had achievement scores below district averages on reading and math assessments. In one set of schools we supported, graduation rates were no better than the statewide average, and reading and math scores were consistently below the average. The percentage of students attending college the year after graduating high school was up only 2.5 percentage points after five years. Simply breaking up existing schools into smaller units often did not generate the gains we were hoping for.

On a more positive note, we saw encouraging successes in some of the new, small schools we supported, including some in New York City. Their graduation rates were nearly 40 percentage points higher than the rates in the schools they replaced. In 2006, the small schools' graduation rates exceeded those of comparable schools in the district by 18 percentage points. Chancellor Klein is here this morning, and I want to thank him and Mayor Bloomberg for their leadership.

There were a number of small school replications that were also encouraging: KIPP, High Tech High, Green Dot in Los Angeles, Hidalgo Early College High School in the Rio Grande Valley, YES College Preparatory Schools in Houston, Aspire High Schools in California, the Noble Street network in Chicago, IDEA Public Schools in Texas. I’ll highlight just one of these: YES College Prep.

YES has done an impressive job with demographics that are the same as the lowest-performing public schools in Houston. Ninety-five percent of the students are African-American or Hispanic. Eighty-eight percent will be the first in their families to go to college. Eighty percent are economically disadvantaged.

For the eighth year in a row, 100 percent of their graduates were accepted into four-year colleges, including some of the top universities in the country. Ninety-one percent of YES alumni have either graduated from or are still enrolled in a four-year college.

Those are the top-line results of our work. They have shown us that all kids can succeed. But since our goal was not only to turn around schools, but to find good models and take them to scale, I have to add: we did not get the results we were seeking in scaling. We wanted to reach all schools indirectly, by showing clear gains and inspiring other schools and districts to replicate those models. Largely, this has not happened.

At our foundation, we believe that success ultimately means that at least 80 percent of low-income and minority students graduate from high school college ready. According to our data, the number of low income and minority students graduating college ready today is 22 percent, and that figure is increasing far too slowly. It’s unacceptable. We need to do better.

So let me describe what we make of the evidence, and what we plan to do next.

The disappointing results showed how hard it can be to convert large, low-performing high schools into smaller, more autonomous schools. To be successful, a redesign requires changing the roles and responsibilities of adults, and changes to the school’s culture. In some districts, we got tacit agreement to move forward, but then the schools weren’t willing to do the hard things—like removing ineffective staff or significantly increasing the rigor of the curriculum.

In New York City, many schools reorganized the school day to get students more time with math and reading, and they reduced the size of the school to improve relationships between students and teachers. Results showed that smaller, more personal learning environments and strong, caring bonds between students and adults can increase graduation rates dramatically. We see these structural changes as necessary, but not sufficient.

We saw that there is a big difference between graduating from high school and being ready for college. In New York City, less than 40 percent of the class of 2007 met the City University of New York's standard for college readiness on the Regent exams. And the percentage of students from small schools was no better than the rest of the city.

It’s clear that you can’t dramatically increase college readiness by changing only the size and structure of a school. The schools that made dramatic gains in achievement did the changes in design and also emphasized changes inside the classroom.

For example, YES and other models include a longer school day and a mandatory Saturday school and a summer school program, which is in line with some of the changes in the New York schools. But YES also sets academic standards that line up with the expectations of top universities.

In general, the places that demonstrated the strongest results tended to do many proven reforms well, all at once: they would create smaller schools, a longer day, better relationships—but they would also establish college-ready standards aligned with a rigorous curriculum, with the instructional tools to support it, effective teachers to teach it, and data systems to track the progress.

These factors distinguished the schools with the biggest gains in student achievement. Interestingly, these are also limiting factors in taking these gains to scale. A model that depends on great teaching can’t be replicated by schools that can’t attract and develop great teachers. A school that has great instructional tools cannot share them with schools that don’t use the rigorous curriculum those tools are based on.

We will continue the part of our work that is dedicated to improving the structure of schools, because it can help promote achievement.

But the defining feature of a great education is what happens in the classroom. Everything starts from that and must be built around it. So we’re going to sharpen our focus on effective teaching—in particular supporting new standards, curriculum, instructional tools, and data that help teachers—because these changes trigger the biggest gains, they are hardest to scale, and that is what’s holding us back.

We’re not the first people to focus on effective teaching to improve education. We’re not even the first people in this room. A growing body of evidence tells us that teacher effectiveness is the single most important factor in student achievement. If you take two classrooms from the same school, both starting out at the 50th percentile, and assign one to a teacher in the top quartile and another to a teacher in the bottom quartile, there will be a 10 percentile difference in achievement at the end of the year.

In fact, research shows that there is only half as much variation in student achievement between schools as there is across classrooms in the same school. We’ve known about these huge differences in student achievement in different classrooms for at least 30 years. Unfortunately, it seems that the field doesn’t have a clear view on the characteristics of great teaching. Is it using one curriculum over another? Is it extra time after school? We don’t really know. But that's what we have to find out if we're going to not only recognize great teachers, but also take average teachers and help them become great teachers. I’m personally very intrigued by this question, and over the next few years I want to get deeply engaged in understanding this better.

The first step in identifying effective teaching has to be setting fewer, clearer, higher standards that are aligned with the goal of graduating students from high school college-ready. You can’t compare teachers if they’re not pursuing a common standard. I believe strongly in national standards. Countries that excel in math, for example, have a far more focused, common curriculum than the United States does.

Every student is capable of a college-ready curriculum; that has to be the standard everywhere. On our trip to Texas that Melinda mentioned, she and I spoke to a group of teachers who were working at KIPP Academy schools. When we asked them why they chose KIPP schools over public schools, one Latina teacher said: “I wanted to teach at a school where everyone believes you can go to college, even when you look like me.”

That ought to be every school in America.

Our foundation will keep working with states and districts to develop a core set of priority standards that students need to succeed in higher education, and getting states and districts to sign on. Members of the Common State Standards Coalition have built momentum on this issue—we now have governors and state education chiefs leading the effort. So we’re optimistic about this.

As states begin to embrace common standards, technology will help us create the next-generation models of teaching and learning. With interactivity, we can provide software to qualify a student or to bring a subject to life. We need to have the best lectures available online and for free on DVDs. Microsoft did this in India with math courses and saw that it was beneficial in a number of ways. They held contests to pick the most effective teachers from their lectures on the DVDs. Then they distributed the DVDs. Some students watched the lectures. In other cases, teachers watched outside of class to improve their teaching, or they would assign the lectures to kids who were ahead or behind.

And we’re not doing enough to provide data for teachers. Amazon.com knows every book you’ve ever bought from them. They can recommend five more based on what you like. But we have no such tool set for teachers. On the first day of school, a ninth-grade teacher has absolutely no idea which of her students can calculate the area of a circle or identify the elements of a short story. Teachers should know this.

The education sector desperately needs an infrastructure for creating better instructional tools—always with measurement systems in place so we have evidence that the new way works better than the old way. Without evidence, innovation is just another word for “fad.”

We need to be able to determine which curricula, which software and other instructional aids are most effective in helping teachers teach and students learn.

Doctors aren’t left alone in their offices to try to design and test new medicines. They’re supported by a huge medical research industry. Teachers need the same kind of support. We will help build the infrastructure for testing and evaluating the tools developed by others.

Teaching is a hard profession. As we heard in the video from Texas, a lot of great teachers leave school far sooner than they want to because they’re exhausted. Offering this kind of help could not only improve student performance; it could help make good teachers into great teachers, and help keep great teachers in the classroom longer.

We will also be helping states and districts build data systems that provide teachers timely feedback about student learning. One of the great benefits of No Child Left Behind, whatever its flaws, is that it requires states to track data about the achievement gap. That’s crucial information for addressing inequity, and we need to build on it.

A principal should be able to see at a glance how each student in a school is doing, and ask about those who are falling behind. We have seen people oppose this kind of data system on behalf of privacy; I don’t think that argument holds. I’m optimistic that very advanced data systems can be built that provide indispensable information on student progress while preserving legitimate privacy concerns.

Data systems, of course, will tell us which teachers are getting the biggest achievement gains every year. If we’re going to retain them, we’re going to have to reward them. It’s astonishing to me that you could have a system that doesn’t allow you to pay more for strong performance, or for teaching in a particular school. That is almost like saying “Teacher performance doesn’t matter”...and that’s basically saying: “Students don’t matter.”

If we don’t pay great teachers more, we won’t develop and keep more great teachers. This isn’t computer science; it’s common sense.

There are two extreme sides in this debate. According to the caricature, one side just wants to turn teachers into commissioned salesmen, so their whole salary is based on how much the scores improve. The other caricature says that teachers don’t want to be held accountable, so they will reject any system that ties pay to performance. In truth, designing an appropriate incentive system is difficult, but possible.

We believe in incentive systems, but we understand the concern that without the right design, they could seem arbitrary or incent the wrong things. They need to be transparent, they need to make sense, and teachers themselves need to see the benefits of the system and embrace them.

That’s why we’re going to set up partnerships in three to five areas to design a system that offers the training and tools that help every teacher improve; recruits, rewards, and retains effective teachers; and gives them incentives to work in the schools where they’re needed most. Then we will measure whether it leads to significant improvements in student achievement.

We’re going to choose districts that have strong leadership, that have a base level of data systems in place, and that have demonstrated support from teachers and the local teachers union. If the teachers don’t embrace it, it will fail.

We’re excited by certain models around the country—including Green Dot schools in Los Angeles and the schools in Prince George’s County, Maryland. We’re also encouraged by the model in Denver, where the teachers union and district administrators designed and adopted a system based on performance incentives. Teachers could choose their old pay formula, or join a new system that gives raises and bonuses for meeting test score targets and teaching high-need subjects, or working in high-need schools. The district funded the additional pay from a $25 million annual property tax levy that was approved by voters—but the only way the teachers could draw down the salary is if they enrolled in the new system.

It showed that taxpayers are willing to pay more to support their local schools—if the extra money is tied to higher achievement.

Money is tight. We need to spend it wisely. We’re now spending $8 billion a year for teachers with master’s degrees, even though the evidence suggests that master’s degrees do not improve student achievement. We’re spending billions on a seniority system, even though the evidence says that seniority, after the first five years, may not improve student achievement. We’ve spent billions to reduce class size, even though there is no strong evidence that spending money to reduce class size in high school is the most impactful way to improve student performance.

And the last thing we can afford—whether the economy is good or bad—is to pay teachers who can’t do the job. As President-elect Obama and others have pointed out: We need to give all teachers the benefit of clear standards, sound curriculum, good training, and top instructional tools. But if their students still keep falling behind, they’re in the wrong line of work, and they need to find another job.

Anyone who opposes dramatic change in our schools has to make an impossible case. Either they have to deny that our schools are failing, or they have to argue that the kids are to blame. Either view is wrong. If you believe every child can learn—and the evidence strongly supports this—then if the students don’t learn, the school must change. It won’t be easy, but it’s essential.

I am optimistic. We have better technology than we've ever had to help us identify great teachers, support their work, and spread their methods. We have ingenious ways to tap a kid’s desire to learn. We have political momentum that is bringing teachers and districts together. And we’re going to have a dynamic new president who’s committed to education. The country is ready for change. Let’s use the moment to accelerate the change in our schools.

Thank you.