For decades, most efforts to improve opportunities for high-needs students have resembled snowflakes; they come down from above, are completely different from each other, complicate routines, and rarely stick.
However, experts gathered at EWA’s annual conference in Los Angeles this year said at least one kind of reform has a good chance of making long-lasting gains: “school improvement networks.”
The session, “When Schools Network to Improve,” drew a packed house, likely in part due to Bill Gates’ announcement last October that the Bill & Melinda Gates Foundation is earmarking 60 percent of its $1.7 billion investment in education over the next five years to support school improvement networks.
School networks aren’t new; there are hundreds of them, said panelist Robert Hughes, the director of the Gates Foundation’s K-12 education programs. But too often, said Hughes, most networks don’t dig deeply enough to identify and understand the root cause of a problem, and they don’t use data effectively to determine if their new ideas are actually helping students who have fallen behind.
The type of networks that Hughes and the other panelists described are designed to be collaborative, with coordinated groups that typically would include teachers, school and district leaders, scholars, and outside consultants. Their work can be guided by a set of principles for applying a rigorous, research-based methodology that’s known variously as continuous quality improvement or improvement science.
“I think for teachers, it’s important that they have a [network] launch with improvement science principles,” said panelist Joseph Espinosa, the coordinator for elementary mathematics at the Los Angeles Unified School District (LAUSD). The principles provide a methodology for networks to follow, which is necessary because improvement work “is not a linear process, and it’s messy,” he said.
The first principle is a question that’s fundamental to everything that follows Espinosa said. It asks network members to determine “What specifically is the problem we are trying to solve.”
The question calls for teachers to investigate what’s not working from several perspectives, including one they’re not always comfortable with – asking students for their feedback. At Hillcrest Drive Elementary in LAUSD, where half the students are not meeting academic goals, each teacher interviewed a student “to find out what it’s like to learn math in the classroom,” said Espinosa. They also asked students to draw a picture to illustrate the problem. The network settled on problem-solving as its priority, with a specific goal of having 80 percent of students able to solve a word problem in more than two ways.
“When the textbook is driving instruction, [teachers] don’t see what’s happening with students. But when they use the students’ feedback, they feel that how they teach really does impact the students,” said Espinosa.
Time is not of the essence
Continuous improvement is a time-consuming process that can be at odds with a world where schools and teachers are under pressure to show gains for all students within one academic year. Yet, taking it slow is part of the reason that changes developed by school improvement networks often succeed where others fail. It took months to agree on the problem at Hillcrest Drive.
“Spending a lot of time marinating in a problem is very important,” Espinosa explained. “You can gather data, you can interview students, and you can develop a plan or strategy based on that.”
LAUSD is one of eight districts in an improvement network started by the California Office to Reform Education (CORE), which reaches more than one million students. CORE’s ambitious goal with improvement science is to close the achievement gap in math for Latino and African American students.
It also takes time to train teachers in the principles of improvement work, said Kirk Walters, the principal investigator of the Better Math Teaching Network, a project launched in 2015 by the American Institutes for Research (AIR).
“We’re trying to get teachers to be improvement scientists,” said Walters. The AIR network consists of 40 high school math teachers at public schools in New England serving a predominantly low-income population. They are putting research on student-centered math instruction to the test, Walters said. The goal, he noted, is to increase the number of students who understand how algebra works.
Frequent mini experiments
Typically, reforms last for most of the school year, so by the time teachers know whether or not the changes helped improved scores, it’s too late to do anything about it for those students, said Sarah Duncan, the co-director of the University of Chicago’s Network for College Success. Another problem with the usual way of doing things, said Duncan, is that schools “do a lot of [reforms], and don’t know what worked.”
A distinctive element of school improvement networks is that they regularly evaluate changes to see if they’re successful, Duncan said. Think of it as quick cycles of mini experiments, she said. For example, teachers at a Chicago high school started holding lunchtime tutoring sessions to see if that would help students who weren’t passing math. After two weeks, the teachers compared grades of students who had been failing and attended tutoring sessions, with those who had been failing and didn’t take advantage of the tutoring. The school found that tutoring helped quite a bit, but not for everyone.
Before the teachers started on another cycle, they talked to the students who didn’t get anything out of tutoring and used that feedback to develop and test another strategy. Each successive mini experiment provides new insights into how different types of students learn and what they need to be successful in school, said Duncan.
“The idea is to learn as much as you can, cycle after cycle, and then share that with folks in the network,” explained AIR’s Walters. “Part of the reason I’m excited about this work is that the promise of this method more directly connects research with the problems on the ground in schools.”
Building a systemwide infrastructure
Improvement cycles in classrooms are an important aspect of the work, but ongoing network success requires buy-in from schools, and especially, district leaders.
“We only work with schools where this is a principal’s priority,” said Duncan. That’s crucial for several reasons. Network members have to meet regularly – in person or virtually – to compare and analyze data from their improvement cycles. School leaders must be willing to rearrange schedules to free up the teachers. Unfortunately, in Chicago, the principals’ support was undercut by the district’s decision to reduce teacher planning time.
There has to be a shift toward systems thinking, said Hughes. “Continuous improvement involves thinking about the systems and resources we have in our schools and how to use them to improve student outcomes.” That includes more professional development that helps teachers learn to apply improvement science in their particular classrooms. “We’re really building a different kind of infrastructure in public education that honors teachers and gives them a different context.”
The panelists offered suggestions on how to tell if school improvement networks are gaining traction.
“The story is in the schools,” said Duncan. “Ask kids, ‘Do you know what you’re working on?’ Notice how engaged the kids are. We now know that learning looks like a level of discourse between kids and teachers where the kids are really working on the problem themselves.”
Look at “how people solve complex problems in schools,” added Walters. “Do they use these principles? Do they work or not?”
“The goal of this strategy is to see young people improve,” said Hughes of the Gates Foundation, adding that the story is how the incremental changes are building on each other to reach that goal. “Over time, we are going to see innovation.”
We’re not there yet, said Espinosa. When the right infrastructure is in place, teachers “will learn these routines and they [will] become routine.”
During the Q and A portion of the discussion, a question was raised about the capacity of school systems to create and sustain a reform that requires significant buy-in from teachers, administrators, and school boards.
“Twenty years ago, people tried networks … and ran into problems with implementation. What’s different now?” asked Alexander Russo, editor of The Grade.
Hughes responded that educators did learn from that previous experience. He noted that the learning curve for network reform is slow and that professional development is expensive and has to be done right to make continuous improvement work.
“You have to play the long game,” said Hughes, who urged journalists to let the story unfold over time. “The story is getting young people to succeed, every day, every year, and these add up to changes.”
Duncan said she believes there will always be skeptics and some teachers resistant to change, but she said that attitude won’t last when all the classes around them are showing strong improvements. “Once you see the results,” she said, “you won’t go back.”