2012-02-08

Co.Exist

Personal Tutors And Paying For Good Grades: Roland Fryer's Experiments On Children

Fryer—a Harvard economist and MacArthur genius—doesn’t care how we get underperforming students to do better. He just wants to find ways to test any education hypothesis as quickly as possible, until we find what works.

Roland Fryer doesn’t want to guess at how to help kids who are stumbling through school. He doesn’t believe in "going with his gut" or heart or any other body part or intuition. The Harvard University economist is neither insecure or unusually modest: His personal history is an intellectual’s rags-to-riches story that gives him more than enough street cred to tell the world how to help poor, black inner-city kids with messy home lives.

But Fryer is a true empiricist: He wants to test and measure and validate every theory about academic interventions. His goal is big: Nothing less than changing life outcomes for black people. And his drive for significant, research-validated data could make him one of the most pivotal people in the American school movement.

Fryer, 34, is an economist at Harvard University who also runs the four-year-old Education Innovation Laboratory (or Ed Labs). Last year, Fryer was awarded one the "genius" grants handed out by the John D. and Catherine T. MacArthur Foundation. He’s also gaining a reputation for systematically asking tough questions—then designing experiments that have real scale to answer those questions. A project he and his team started last year in Houston, which they nicknamed Apollo 20, is already providing data on the effectiveness of five broad interventions. And in the coming year, he’s hoping to expand that research, possibly with the help of a still fledgling group, the League of Innovative Schools to dive into operational details.

"People sometimes cringe at my use of the word 'experiment,'" Fryer says. "It’s not like we’re shocking kids. … Schools do this all the time." A school may try one approach with 3rd graders and a different strategy with 4th. What Fryer wants to do: Apply a research rigor to that kind of experimentation and then share the results. And do it fast.

"Instead of taking 20 years to figure out the right way to use data, let’s do that within one year," he says. "These kids aren’t going to get 4th grade back. I don’t have the time to wait for 20 years. We need to do it right now."

Even with billions of dollars of both government and philanthropic dollars pouring into education, many choices have relied more on anecdotal observations rather than evidence-based research. That’s in part because the existing "evidence" is a patchwork—too often not statistically significant, not conducted with classical methodologies (such as established control groups), and so on.

"Do we really know, of all the levers available to us, which ones we can pull for a particular school or district?" says Fryer. "Merit pay, smaller schools, charter schools, after-school programs, instructional coaching, culturally sensitive curriculums … of all these, what can really close the gap? What works?"

There’s another problem, too. Education reform has always been a cauldron of simmering social concerns. Now the pot is boiling furiously: Are our public schools failing us? Is America unfairly demonizing teachers? Why can’t Johnny read? Should we pay teachers more? Are charters the solution—or do they suck resources out of communities? How do we improve test scores? Should we let kids use iPads in the classroom? Are we raising a nation of Tweeters? Does the technology usefully "engage" kids or distract them? Who’s leading the conversation on reform?

Just about everyone who went to school uses their personal experiences to inform their opinion; if they’re a parent of a school-age child, they often feel they’ve got a double-dose of expertise. (How many comments on education pieces begin with something like "When I was in school …" or "My daughter’s 2nd grade teacher …"?)

Just about everyone—but not Fryer. That’s impressive in itself because his childhood was far more like the lives of the students he is studying than is often the case with Ivy League professors. Fryer grew up in Florida and Texas. His mother was absent, his father in and out of trouble; his grandmother had a steadier hand. Eight out of 10 family members either died young or did jail time, according to a detailed profile of Fryer that ran in 2005 in The New York Times.

But instead of going with his own hunches, Fryer relishes breaking hypotheses down into testable nuggets. Here’s an example: A few years ago, Fryer roiled educators by proposing to pay students for doing school work and getting good grades. Fryer wanted to investigate whether structuring such payouts in different ways would help or hurt outcomes.

Over two school years, Fryer’s team handed out about $6.3 million to 38,000 students in 261 schools in four "prototypically low performing urban school districts." They tested a wide range of scenarios. For instance, some students received $2 for every book they read; others received $5 for taking a quiz; middle schoolers could earn $50 if they got a perfect score.

Their conclusions: Paying kids to get good grades had virtually no effect. Most of the students didn’t have a clear idea of what to do to "improve their grades." Paying students to do things they could control—hand in their homework, read, dress better—did indeed bolster those activities.

Fryer’s next question: What activities make a difference to how students score on tests? His team spent a million dollars gathering data on the practices of the most effective charter schools, such as the Promise Academy in the Harlem Children’s Zone. They concluded there were five common elements:

  • Increased instructional time
  • Data-driven instruction
  • Increased feedback for teachers on their performance
  • Tutoring
  • A culture of high expectations

But which of those elements really matters? And when? And how should it be implemented? And are there technological tools that can help?

In 2010, Fryer’s team began working with the Houston public school district to begin to explore those questions. After the first year, they saw significant gains in student performance, particularly in 6th- and 9th-grade math when they introduced tutors.

"Our next step is go deeper into the black box of school operations. How do we make best practice out of the things that we do routinely every day?" Fryer says. What are effective ways to do professional development? To involve parents? To use available tools? And what is genuinely involved in doing effective "data-driven instruction," he asks.

The League of Innovative schools, coordinated by Digital Promise, is reaching out to districts that are intrigued by Fryer’s proposition to test fast—and get the information back to them quickly. Fryer sketched out his plans to a group of superintendents in late January in a webinar. Some interventions, he suggested, could be tested in as little as two weeks’ time.

That kind of fast turn around, he hopes, will make economic research relevant to decision makers.

"What typically happens [among academic research] is that someone gets two-year-old data, they spend four years crunching the data, and then show up at a superintendent’s office," Fryer says. "By that time the district’s had two or three superintendents and everyone’s moved on. This is about doing best practice research at the speed the schools are already operating."

Not every question can be answered in two weeks, Fryer concedes. Even so: "There are a zillion questions we can answer that will increase the everyday learning of our kids in schools, right now."

Add New Comment

2 Comments