Introduction

This quote strikes many educators and parents as shocking and counterintuitive. Isn't giving teachers evaluative feedback an essential part of a principal's toolbox for improving teaching and learning?

But when I ask groups of educators what helped them improve in their early years in the classroom, their responses (via anonymous polling) tell a different story. Here's what participants in a recent webinar had to say:

I see similar results every time I ask this question, with “supervisors’ formal evaluations” often getting zero votes. Far more likely to improve teaching and learning, say educators in a wide variety of settings, is informal input from colleagues, mentors, coaches, supervisors, students, various forms of professional development, and a modest acknowledgment of their own training and talent.

Facing Facts

This begs the question of whether teacher evaluation can be a player in improving teaching and learning in K–12 schools. As I've coached principals, given presentations, and read research for the Marshall Memo in the new millennium, several hard truths have emerged:

Hard Truth 1. Students learn a lot more from some of their teachers than from others. The egalitarian teacher norm described by Susan Moore Johnson (2012)—we're all equal in a very tough job—is belied by major differences in achievement from classroom to classroom. The results of a Tennessee study summarized here show a fifty-two-point spread in achievement between students who spent three years with the least-effective and most-effective teachers.

Source: Sanders and Rivers (1996).

What made the difference? It was the cumulative impact of specific teaching practices used hour by hour, day by day, week by week, month by month. Books like The Skillful Teacher by Jon Saphier et al. (2008) and Teach Like a Champion 3.0 by Doug Lemov (2021) have unpacked the techniques that explain why students learn so much more in some classrooms than in others.

Hard Truth 2. Every school has a range of teaching quality from highly effective to not so effective. Variation can be represented by a simple bell curve, which has a slightly different shape from school to school but conveys the same basic idea: there's always a range of teaching effectiveness.

British researcher Dylan Wiliam said it well (2018, p. 183): “Today in America the biggest problem with education is not that it is bad. It is that it is variable. In hundreds of thousands of classrooms in America, students are getting an education that is as good as any in the world. But in hundreds of thousands of others, they are not.”

In hundreds of thousands of classrooms in America, students are getting an education that is as good as any in the world. But in hundreds of thousands of others, they are not.

Hard Truth 3. Vulnerable students have a greater need for good teaching than their more-fortunate classmates. Yes, a rising tide of effective instruction lifts all boats, but the maritime metaphor doesn't convey an important characteristic of schools: skillful teaching makes a bigger difference for students who walk in with any kind of disadvantage, including poverty, neighborhood violence, quarreling parents, learning disabilities, health issues, and ineffective teaching the year before. The study summarized here compared the impact of effective and ineffective teachers on students with different levels of preparation as they moved through fifth, sixth, and seventh grades:

Source: Bracey (2004).

On the left are three cohorts of students who were lucky enough to have effective teaching three years in a row. They achieved at similarly high levels, even though some (the left-hand bar) started with much lower achievement than others.

The three student cohorts on the right are a matched sample who had three years of ineffective teaching. Those who started out with high and middle achievement (the two bars on the right) were still doing quite well at the end of seventh grade despite lower-quality teaching, but those who started out with low skills fell way behind.

With mediocre and ineffective teaching, we see a widening proficiency gap—the so-called Matthew effect, with the rich getting richer and the poor getting poorer.

This study and others like it show that vulnerable students disproportionately benefit from good teaching—the so-called equity hypothesis (Fullan, 2003). With mediocre and ineffective teaching, we see a widening proficiency gap—the so-called Matthew effect, with the rich getting richer and the poor getting poorer. Conversely, with effective and highly effective teaching we see more-equitable student outcomes and narrowing proficiency gaps.

Hard Truth 4. Traditional teacher evaluations rarely have an impact in teaching and learning. This figure lays out the components of the “clinical supervision” model that's been standard practice in K–12 schools for almost a century:

We'll go into more detail on the design flaws in this model in Chapter Three, but one problem jumps out: if a single evaluation takes four hours, a supervisor with twenty to twenty-five teachers (a typical caseload) is spending a lot of time each year on an activity that does very little to improve teaching and learning.

In some schools, these four hard truths converge in a perfect storm, with grievous effects on students’ education:

• Many students have academic and other disadvantages.
• Too much teaching is mediocre and ineffective.
• Teachers are not effectively supervised and evaluated.

Tragically, this scenario is most common in economically embattled communities where the need for good teaching is greatest. The pandemic heightened these equity issues, and even though the emergency has passed, its lingering effects on student behavior and learning loss continue to erode teachers’ and administrators’ morale. It's more urgent than ever that schools use the most effective methods!

The Search for a Better Process

None of this is news to seasoned educators and policymakers, and experts have been hard at work looking for ways to improve teacher evaluation and student outcomes. Here are ten theories of action that have been used in some schools around the US, each followed by my concerns about its viability:

• Double down on the traditional teacher evaluation model, investing heavily in training supervisors to ensure inter-rater reliability. Spending more time preparing administrators on a deeply flawed model will not improve outcomes. There might be more uniformity in write-ups, but they will remain an ineffective method for improving teaching and learning, taking up large amounts of supervisors’ time that could be better spent, and adding to their cynicism about the process.
• Use detailed rubrics to 4-3-2-1 score individual lessons. Rubrics are helpful descriptions of the many facets of teaching, but they're not suitable to evaluating a single lesson, during which a teacher can demonstrate only a small part of the overall palette of effective instruction. Rubrics are best used to evaluate each teacher's work at the end of the school year - more on this in Chapter Seven.
• Bring in outside evaluators to backstop principals’ evaluations. The idea is to have supervisors with more objectivity to supplement on-site supervisors, but educators from outside don't know the culture, curriculum, and personalities of a school when they parachute in and can't possibly visit classrooms often enough to give fair and accurate evaluations. Better to put the resources into supporting school-based supervisors with manageable caseloads and a better evaluation model.
• Use anonymous student surveys as a significant part of teacher evaluations. Although students speak the truth about their teachers and their input can provide valuable pointers (and sometimes stinging rebukes), making surveys high-stakes (in Pittsburgh schools they were 15 percent of teachers’ evaluations) can corrupt the process and prevent teachers from listening to their students’ helpful suggestions.
• Inspect lesson plans and classroom artifacts to ensure quality teaching. Yes, teachers need to be prepared for each lesson, but lesson execution is what matters. The time supervisors spend reading and commenting on lesson plans is better spent visiting classrooms (they can spot-check the lesson plan) and talking to teachers about how each observed lesson went. Asking teachers to submit “evidence” of their planning and assessments is also a poor use of their time—and a poor use of administrators’ time going through reams of paperwork or digital files.
• Use “real-time coaching” with supervisors intervening during problematic lessons. This idea will be discussed in more depth in Chapter Twelve; suffice it to say that this runs the risk of undermining teachers’ authority with students and making teachers dread every visit by their supervisor. Except for dire emergencies, why not wait till after the lesson to talk to the...