The Teacher Wars

by Dana Goldstein

  When I visited Harrison District 2, the Colorado school system that created standardized tests for every subject and grade level, I accompanied crusading superintendent Mike Miles on a round of classroom observations at Fox Meadows Middle School.*1 There is no collective bargaining in Harrison, so Miles had incredible autonomy in shaping teachers’ working conditions. Teachers had to keep their classroom doors open at all times. They were told to expect up to sixteen surprise observations each semester from administrators, instructional coaches, or outside consultants. Teachers told me they felt constantly watched. A TFA corps member who generally supported Miles’s accountability efforts described Harrison as “a high-anxiety district to work in.”

  This all seemed a little extreme. And since I hadn’t been impressed with the quality of the district’s paper tests in art or physical education, I was prepared to find the classroom observations less than compelling. But I was wrong. In spot observations of about ten or fifteen minutes each, Miles was able to make a series of insightful critiques of teachers’ performance. He traveled through the school’s hallways with a six-person team of administrators and consultants, some of whom were being trained in these new observation methods. When the group stepped into the classroom of a science teacher in her early twenties, the young woman began to tremble ever so slightly. She was conducting a lesson on “hypotheses, graphs, and data.” But the activity she had assigned the seventh graders, reading a graph and answering questions about the values on it, had nothing to do with hypotheses, which Miles thought was the most important concept in the lesson. Out in the hallway, he discussed with his team what they had observed. “Did she get to the idea of using data to construct a hypothesis?” he asked. “No.” He also noted that while the students worked in small groups, the teacher could have been moving around the classroom more actively, making sure each child was participating.

  In a social studies class, Miles was unimpressed with the teacher’s assignment for students: “Using geographic facts, which Western European country most resembles Colorado?” It was vague, kind of boring, and far too easy. Plus, there was no map displayed of Europe, making it hard for the students, most of whom had never left Colorado, to visualize what they were learning about.

  For a math teacher working with students on circumference, the superintendent suggested that instead of simply writing equations on the whiteboard, the teacher could have demonstrated the concept using a physical object, like a basketball. A second math teacher spent ten minutes—far too long—explaining that the word “denominator” referred to the bottom part of a fraction. “She’s not a superstar,” Miles said. He knew of an exemplary math teacher whose classroom this teacher should observe.

  Before he left Fox Meadows that day, Miles took some time to evaluate the evaluators. Leafing through the notes the administrators and consultants had taken on each teacher, he said, “I’m not seeing enough validating comments. Every room I went into, I saw positive things.” Later on he told me he thought the school had two to three “distinguished teachers,” the district’s highest designation, and perhaps four who would be dismissed at the end of the year, one of whom had tenure. In a school of thirty teachers, that was a roughly 14 percent ineffective rate.

  There is nothing new about the idea of an administrator taking a detailed look at a teacher’s classroom practice. Progressive Era reformers promoted “efficiency” observations, in which supervisors used lengthy rubrics to rate teachers according to measures like how many children were late to class or how many seconds it took to hand out worksheets. After World War II, a “supervisory visit” to a teacher’s classroom might have entailed a principal judging whether sufficiently “democratic ideals” were being promoted in the lesson. By 1980 many school administrators used a “clinical” model to observe classrooms, a system based on medical rounds, popularized by Robert Goldhammer of Harvard’s graduate-level teaching program. Principals would conduct pre-observation and post-observation conferences with teachers to reflect on their practice, areas for improvement, and long-term goals. But because Goldhammer had not defined specific characteristics of effective instruction, principals who used his model often failed to provide concrete, helpful feedback to teachers.

  Later, Madeline Hunter’s more prescriptive “lesson design” system became popular. Principals looked for whether a teacher’s lesson included several key components, such as a lesson objective written on the board, a “model” of successful performance, and an opportunity for students to practice new concepts. Mike Miles was clearly influenced by the Hunter system—a combination of direct instruction and student group work, similar to how Teach for America expects its corps members to work. A criticism of Hunter is that she focused too much on teacher-directed behavior, and not enough on whether the teacher helped students become self-directed learners. Today there are other, potentially more sensitive classroom observation tools, and because of Race to the Top’s emphasis on improving teacher evaluation, these methods are now being adopted in thousands of classrooms nationwide. The Classroom Assessment Scoring System, CLASS, was developed at the University of Virginia. It gives teachers numeric ratings based on whether they exhibit behaviors associated with achievement gains, such as “expanding on student talk”—repeating a child’s speech back to her, using corrected grammar and more sophisticated vocabulary.

  Another popular and detailed classroom observation model is embedded within Charlotte Danielson’s Framework for Teaching, first developed in 1996. Danielson was a former Washington, D.C., public school teacher who went on to become an education researcher at Educational Testing Service, the test-maker. She knew that during the “competency” craze of the 1970s there had been a lot of vague talk about asking students “higher-order questions.” Danielson wanted to watch teachers work, look at their students’ performance, and figure out exactly what effective, higher-order instruction looked like. According to her findings, an effective classroom discussion question has more than one correct answer. (Not “When did Hitler come to power?” but “What social, political, and economic factors led to the Nazi Party gaining power in Germany? Which factors do you think were most crucial? Why?”) Danielson found that an excellent teacher asks students to explain concepts to one another, instead of repeating herself ad nauseam. She highlights connections between disciplines—for example, by giving students background information on Elizabethan England before assigning a Shakespeare play. She chooses books and works of art from a broad range of cultures, including the cultures from which her students hail. She allows students to debate one another in class, and requires them to cite evidence to support their claims. If the teacher is really skilled, the students can talk among themselves for many minutes about the topic at hand without her interrupting. She assesses her students throughout a unit, not just at the beginning and end, through pointed questioning and the use of tricks like “exit slips”—quick problems students must solve, on paper, before they leave the classroom.

  In New York City, principals must now use the Danielson framework to observe each teacher’s classroom at least four times per school year, and to conference with each teacher for professional development. If there is a downside to systems like these it is that most research-driven observation rubrics require administrators to rate teachers on many, many different competencies—twenty-two in the Danielson framework. Historically, evaluation systems with heavy time and paperwork burdens have not been viable in the long term, because principals either go through the motions without making meaningful distinctions among teachers, or they find ways to use the great number of subjective variables in these rubrics—for example, “compliance with standards of conduct,” in Danielson’s framework—as a way to target disfavored teachers for dismissal, regardless of more objective measures of performance.

  One of the most frequent complaints I heard from teachers about classroom observation in Harrison District 2 (and many of the other school systems I’ve visited) was that consultants and ot
her non-teacher observers had a poor understanding of curricular content and so failed to provide relevant feedback. Bringing respected teachers into these observation processes as peer coaches and evaluators can not only lighten administrators’ workloads, but also greatly increase teacher buy-in. To do that, accountability reformers and teachers unions are taking a second look at teacher peer review, the practice first developed in Toledo in 1981.

  Bob Lowe, a thin, short man in his fifties, was a little jittery when he arrived at a drab suburban conference room at the headquarters of the Montgomery County Education Association, his union, on a rainy morning in June. He was there to fight for the continuation of a thirty-three-year teaching career.*2

  Lowe had to make his case to a panel of eight teachers and eight principals who had been selected by their unions and the county school district as expert practitioners. The panel had already heard from Lowe’s boss, the principal of a large, diverse suburban high school. She had tried everything to help Lowe improve as a teacher, she said, even assigning one of her deputies to spend 60 percent of her day in his classroom. But he was disorganized—he turned his grades in late and lost student work. In class he lectured in front of the room, sharing anecdotes from history that often amused students. But he offered little sense of the lesson’s objectives or what students would have to do to perform well on an essay or test. The teenagers lost focus and wandered around the room. Even his honors kids did poorly on final exams.

  It wasn’t easy for the principal to recommend Lowe’s dismissal. He was a big part of the school community. Every year he hosted a cultural exchange with a Russian high school, whose students visited Maryland. “He is an extremely good person, a good man,” the principal said. “Students generally like him. He has a passion and deep knowledge of his content area. But he can’t teach this to his students.”

  The panel had also heard a report from Lowe’s “consulting teacher” (CT), part of a fleet of peer coaches, one of whom Montgomery County assigns to every novice teacher and to veteran teachers flagged by a principal for underperformance. Consulting teachers work full-time as coaches and evaluators for three years, earning a $5,000 bonus above their regular teaching salary, and then return to classroom positions. The report from Lowe’s CT included an actual script from several of Lowe’s lessons, and it didn’t sound good. Lowe’s directions to students were so vague that he had to re-explain one assignment three times. During a five-minute warm-up activity, he demanded the students’ attention seven times. He spent nine minutes having kids rearrange their desks to transition to a new activity, and he allowed struggling students to evade participation in class discussions. At his coaching sessions, Lowe had attempted to explain away his poor performance by recounting family tragedies and medical problems.

  By the time Lowe appeared to plead his case, the outlook was grim. He spoke about his love for the county’s public education system, which he had attended as a student and where his own children went to school. “I never wanted to be anything other than a teacher,” he said. “I’m a third-generation teacher. My sisters are teachers. I don’t know what I’d do if I weren’t a teacher.” But it was clear that expectations had risen at his school and he hadn’t kept up. Tracking student data had not been a demand earlier in his career, nor had aligning lesson plans to specific curriculum standards. His lectures were, in his own words, “ad hoc.” Yes, he admitted, more of his students were failing this year than at any time in the past six years. He seemed to blame the kids. “Am I unclear, or are they deer in the headlights?” he pondered. “Some kids are like that most of the time.” He demonstrated the terrified deer, stiffening his entire body and staring wide-eyed into the distance. In the end, he simply pleaded for mercy. “I just hope that I can end my career with some dignity.”

  After Lowe’s presentation he left the room. Since his principal was actually a member of the peer-review panel, she recused herself, as did another member, a teacher who had taught Lowe’s daughter. There were fourteen judges left, solemnly perusing huge binders full of information on Lowe and his students, including grades, test scores, e-mails from parents, and photographs the principal had taken of his messy classroom. How many of his students were classified as special education, the panel discussed? Had that made his job tougher? Lowe had been in peer review once before, five years ago. Had anything improved since then?

  One young teacher on the panel seemed to have made up her mind. “I don’t believe more support would help him,” she said. “He’s not leading the students strongly enough. I see no strategies or methodologies for improvement.” A principal agreed. “He rambled on. If he teaches like this, how can there be clarity for students?”

  In the end, the panel voted unanimously for dismissal. Union vice president Christopher Lloyd, a lanky middle school teacher, went upstairs to tell Lowe his career in Montgomery County was over. When Lloyd returned to the conference room, his face was drawn. But he gave the peer-review panel a pep talk. In Toledo teachers can remain in the peer-review system indefinitely, as long as they show some small amount of improvement. “That’s not the philosophy of our system,” Lloyd said. Struggling teachers are owed at least one year of intensive coaching on lesson planning, the use of student data, and classroom management. But ultimately, “teachers need to be able to support themselves on their own.… We own this process, and it’s something very important to us as an association. We have to be about protecting the profession, not just about protecting anything that breathes.”

  Peer review has sometimes been portrayed as a sort of union front; a faux accountability plan labor leaders use to distract policy makers from demands to end tenure and use test scores to evaluate teachers. It is true that in Montgomery County, a district of more than nine thousand teachers, only forty-one were dismissed or pressured to resign in 2013 through the peer-review process. At least eight of those were tenured teachers, including Lowe. These low numbers are typical of many of the older peer-review systems across the country. Journalist John Merrow ran the numbers in Toledo in 2010 and found that the nation’s first and most prominent peer-review system terminated 8 percent of first-year teachers annually, but only an average of two-tenths of 1 percent of tenured teachers, likely a much smaller figure than the true number of ineffective veteran teachers in the district.

  In 2004 Montgomery County commissioned an outside report on peer review, which included a survey of principals. One of the questions the district wanted to answer was why so few tenured teachers per year were being put through the process. A number of principals replied that they were overburdened by paperwork and had missed deadlines for referring teachers. Others seemed emotionally reluctant to punish ineffective teaching, telling researcher Julia Koppich, “It’s gut-wrenching. These [teachers] are not strangers. You know their stories. They’re not evil people.”

  Koppich is not the only researcher to mention that principals seem reluctant to fire teachers. Economist Brian Jacob found that in the years after the Chicago Public Schools significantly reduced the paperwork burden for dismissing nontenured teachers, 30 to 40 percent of principals chose not to dismiss a single staffer. Why? It may be that principals fear replacement teachers will be just as bad as ineffective ones who are laid off. Kati Haycock, the president of the Education Trust, has heard another theory: “Maybe the problem is that your best principals don’t give up on a single kid, and they don’t give up on a teacher, either.”

  Supporters of peer review say it is a mistake to judge these systems solely on how many teachers they dismiss; rather, the ideal outcome of the process is helping a struggling teacher develop into an effective one. “The worst thing you can do is fire someone,” says Dennis Van Roekel, the former NEA president. “It means you’ve lost all of your investment in recruiting and training.”

  Peer review is also criticized as too little concerned with student achievement outcomes. Montgomery County was initially denied Race to the Top funding because it refused to give value-added s
cores a fixed weight in its teacher evaluation process. But “data” was one of the most frequently spoken words in that June day of peer-review deliberations. In one case, a tenured pre-K teacher had adjusted special-ed students’ learning plans without presenting supporting data to parents—a big problem. The panel voted to enroll her in peer review for the following year, to help her develop student assessment skills. If she did not improve, she could be dismissed.

  The next case was of a nontenured teacher in a fourth-grade language immersion program. She was a native Spanish speaker in her midthirties who had moved to Montgomery County after teaching in another Maryland district. Her principal reported that the teacher had increased the reading skills of her gifted and talented students by only 1.84 book levels—less than he thought was reasonable—and in math, test scores showed 26 percent of her gifted students had made no progress at all. In that case, after interviewing the teacher and looking closely at the principal’s supporting documents, the panel chose to return her to the regular teaching pool. It turned out the teacher had been assigned an especially large class of twenty-nine fifth graders, including all of the grade level’s special-education students, some of whose parents had requested her. She had worked successfully with one boy with selective mutism, a condition in which a child will not speak in social situations. That boy now participated in class discussions. “I’m really happy about that,” the teacher said. But she left the room knowing she had serious work to do on differentiation—making sure that her class is as challenging as possible for both special-education and gifted kids. She would have the opportunity to apply for tenure in the future, though there was no guarantee she would earn it.