For all the debate over education, there is one point on which almost everyone agrees: Good teachers make a difference.

Just how much teachers can do in the face of deeply impoverished communities and limited government funding may still be controversial. But neither the staunchest union booster nor the greatest champion of school choice and accountability is likely to argue that teachers don't matter.

Study after study has shown that high-quality teaching results in more student learning. The effect is magnified for low-income and minority students and amplified when a student receives consecutive years of good teaching.

However, few teachers have developed the technique to make the impact that is necessary in low-income schools. Research conducted by The New Teacher Project (TNTP) in several large, urban districts has found that only 20 percent of teachers - termed "the irreplaceables" - made a substantial positive difference for their students.

Why are so few teachers prepared to move the needle for the children who need it most?

It's an issue that is getting attention nationally. On July 28, New York Times columnist Joe Nocera wrote about the new book Building a Better Teacher: How Teaching Works (and How to Teach It to Everyone), by Elizabeth Green, editor of a nonprofit news organization that covers education. The punch line was simple: We do a terrible job of teaching our teachers.

Good teacher development is about technique. It's about repeatedly practicing the same very specific skills, getting a little better with each repetition, and then putting your own spin on the material only when you've become a master. As my ninth-grade English teacher, Mr. Bohn - a master teacher in his own right - used to say, "Show me you know the rules perfectly; then you'll know the most effective way to break them."

TNTP is using this practice-heavy approach to increase the effectiveness of young teachers. And, as Nocera writes, there is a new breed of graduate schools of education, such as Sposato in Boston and Relay in New York, that are preparing teachers this way before they enter a classroom full-time.

Teacher training that includes frequent feedback and high-repetition practice of tightly defined techniques should not be restricted to graduate schools or summer institutes, though. It should be happening in between classes in principals' offices, teachers' lounges, and classrooms after the students have left. There is time for this type of development during the school day, and it does not require additional resources.

The teachers I work with, both new and veteran, gladly give up 15 minutes of prep time three or more times each week to practice new techniques and build their skills. The principals and coaches I work with find ways to reorganize their schedules, make tradeoffs, and de-prioritize other tasks, all so they can devote more of their time to coaching teachers. They know there is an effective way to help teachers get better in their own school buildings. And they know why it's important: It can be a game-changer for their students.

To put great teachers in every classroom, we need great school leadership and a new focus for principals - not on the role of administrator and manager that they are used to, but on the role of expert coach. When school leaders and principals see themselves as - and develop themselves into - experts at getting teachers better faster, then on-the-job training will work for teachers. More importantly, it will work for our most vulnerable kids and underserved communities.

Paul Dean is cofounder and executive director of Jounce Partners, a Philadelphia-based nonprofit focused on increasing skill development among teachers in urban schools. paul.dean12@gmail.com