Does Every Teacher Need a Coach?

A new analysis of 60 studies on teacher coaching finds that benefits dissipate as coaching programs attempt to reach more teachers

Just like athletes become stronger under the guidance of a good coach, so do teachers. There are many ways to design a coaching program, but the general idea is that a veteran educator observes a teacher in the classroom and then gives constructive feedback on issues ranging from managing student misbehavior to framing open-ended questions that push students to think harder. The teacher then tries to incorporate what the coach suggests and the cycle of observation and feedback repeats. The number of sessions and frequency vary.

Researchers began rigorously studying coaching in the late 1990s and have been ramping up in the past 10 years to see how well coaching works and if coaching programs are any better than the kind of training seminars that teachers typically attend to further their “professional development.” Now, a team of researchers has come to a frustrating conclusion: coaching can help but no one has figured out how to successfully expand coaching programs so that they reach many teachers.

“We see coaching is less effective as the programs get larger,” said Matthew Kraft, a professor at Brown University. “It raises real questions on whether coaching should be scaled to be for everyone.”

Kraft began his research into coaching along with classmate David Blazar when they were students the Harvard Graduate School of Education. They were perplexed by the results of a study they were working on in New Orleans, where teaching improved during the first year of coaching in 2011 but not so much in the second and third years. They wanted to understand what other coaching studies had found and if there were lessons for policymakers.

Kraft and Blazar, now at the University of Maryland, have been collecting every well-designed study on teacher coaching. They tallied up results from 60 programs and found that, on average, coaching greatly improved the quality of classroom instruction as measured by how outside observers evaluated teachers’ performance and interaction with students. But the average improvement in students’ academic achievement, as measured by reading or math assessments, was small. Their latest study, also co-authored by Dylan Hogan, was published last month in the Review of Educational Research.

The small academic bump may disappoint coaching proponents but it isn’t all bad news. These small gains are better than what researchers have seen after teachers attend training seminars or summer workshops — which cost school systems millions of dollars every year. The increase in student achievement from coaching is on par with the gain researchers typically see from students of a veteran teacher with five to 10 years of experience compared to a novice teacher.

Kraft points out that it would be unrealistic to expect student achievement to improve as much as instructional quality. Coaching directly targets teaching practices, such as managing disruptions. That allows teachers to spend more classroom time on academics but the connection to student achievement can often be indirect.

As researchers drilled down into the results, they noticed that some programs showed much better results than others, both for instructional quality and student achievement. Surprisingly, the benefits of coaching didn’t improve with the number of sessions. In some cases, coaches met with teachers as few as three or four times. In other cases, it was 15 sessions.  “The quality of the feedback may be more important than actual quantity,” said Kraft.

Focus matters, the researchers found. Coaching everything a teacher does at once was not as effective as using coaching to reinforce a new curriculum or a specific teaching technique. Summer workshops with follow-up coaching appears to be “particularly potent,” said Kraft.

Another interesting result: smaller programs work better. For example, one study of a low-income middle school in California randomly assigned eight teachers to weekly coaching sessions. The coaches specifically worked on helping teachers to express their internal thought processes aloud, while they are reading, so that students could grasp how to be active, critical readers. All the teachers had attended numerous professional development sessions to improve reading at the school. But students of the coached teachers had much larger reading gains.

By contrast, a large statewide Florida coaching program had inconsistent and small results in a 2010 RAND Education study.

As coaching programs get larger, Kraft noticed several common implementation problems. The first is that coaching quality deteriorates. Small coaching programs are often developed by academic researchers who do the coaching themselves or train a small corps of teachers to coach in a specific way. It becomes harder and harder to find great coaches and have them stick to a coaching protocol in an expanded program.

A second problem is enthusiasm. In small, successful programs, researchers found that the teachers liked the coaching process and wanted to improve their teaching techniques. When coaching becomes mandated across a whole educational system, not every teacher is open to hearing critical feedback and changing.

A third problem is scheduling. It’s hard to find mutually convenient times for teachers and coaches to meet. As the numbers of teachers who need coaching grows, so does the scheduling nightmare.

What should be done? Kraft says that one option is to limit who gets coaching and not attempt to provide it for all teachers.

Kraft is more enthusiastic about developing peer coaching instead of relying on expert coaches. With the peer approach, teachers at each school would rotate between mentor and mentee roles with teachers taking the lead in areas that they are relatively more adept at, based on their teacher evaluations. For example, a teacher who is really good at discipline could observe colleagues and give feedback in that area. Another teacher could take the lead in the best way to teach fractions. “I think there’s untapped potential in every school,” said Kraft, who began his career as an eighth-grade English teacher. “It’s less about finding individual experts than about tapping expertise across a teaching staff.”

Technology might be an answer as well. Kraft is currently studying whether expert teachers can coach more teachers by eliminating the logistics of in-person observations and feedback sessions. Mounting cameras in the classroom to videotape lessons would allow a coach to review footage at leisure and provide feedback later by video conference. Whether this will be as effective as in-person coaching is unknown. Other models also include earpieces where expert teachers whisper suggestions to a teacher while he or she is teaching, just like a television news producer guides an anchorman on the evening news.

One thing is clear: The best way to improve education in the United States may be to help our 3.6 million public school teachers become better at their jobs. Unfortunately, the research evidence on how to do this is disappointing.

[Cross-posted at The Hechinger Report]

Jill Barshay

Jill Barshay is the founding editor and writer of Education By The Numbers, The Hechinger Report's blog about education data.