A teacher reviews her performance review

An English and journalism teacher for six years, Coleen Bondy ranked as low average in her effect on students’ test scores this year. The value-added scores — based only on her least-motivated students — are “practically useless in evaluating teacher performance,” she writes in a Los Angeles Times op-ed.

It’s hard for those who finished high school 20 or 30 years ago, as I did, to fathom the conditions in a typical L.A. Unified high school classroom these days. Classes are huge. Students face overwhelming family and social issues. Drugs are rampant. Students are incredibly disrespectful, testing authority constantly at the beginning of the year. Teachers must be able to get a strong grip on their classes all by themselves because consequences for bad behavior in class are often nonexistent outside it.

. . . Today’s teacher must be highly skilled in her subject matter just to make it into the classroom, more so than at any other time in the history of education. She also must play the role of parent, custodian, psychologist, drug and alcohol interventionist and parole officer, to name a few.

“Society has decided to blame many of its failings on teachers,” Bondy writes.

If teachers can’t be evaluated fairly based on their students’ progress (compared to their previous progress’ rates) and they can’t be evaluated based on classroom observations, how can they be evaluated?

About Joanne


  1. While I have heard good arguments against value-added scores, most teachers I have taught alongside don’t understand how it supposedly works. (Actually, many understand how it works but the ones that complain about the system usually do not.)

    The valued-added score is based on how well your students do compared to their historical trend. Unmotivated students generally have low growth each year therefore the teacher is compared against this low expectation. If a teacher has motivated students then the expectation is higher for their overall performance. It is generally harder for teachers who teach the top students to show growth because these students are already performing at the top of low expectations state tests.

  2. Fair point, Joanne.

    And for a long time, I would have defended the teacher more. For she is correct. The conditions of some schools and the “quality” of the kids is truly shocking at times.

    Yet, as a society we have to argue that these schools simply need teachers who give them the greatest opportunity to break the cycle. These schools need truly inspiring leaders and consistently competent teachers who will not give up on any classroom. And I am not saying I am that person.

    But anyone who says “I just can’t make a difference with these kids” needs to leave that classroom and district.

    I don’t know what the answer is, but you are right in that we have to expect success.

    • “These schools need truly inspiring leaders and consistently competent teachers”
      May as well ask for unicorns. Any profession will have a group of individuals of varying ability (except for NASA and similar organizations). Wishing for a magical group of individuals is futile and distracts the public from demanding real change for the better.

      • No, it happens where people are committed. Abraham Lincoln HS in Denver was a wasteland for years until Antonio Esuibel took over. He set the vision and the community responded. It’s happening all over Denver, and not just in charters. Charters make a difference, but plenty of inspirational leaders are changing the urban neighborhood schools, too.

        We can’t, SuperSub, afford to be so pessimistic, so fatalistic when talking about the lives of children. Committed leadership in a school and classroom is not a fantasy. It’s not unicorns. It’s tough as hell at times, but anyone who believes it is impossible needs to get out of public education post haste.

      • Except that good teachers aren’t mythical creatures although for the degree of importance the public education system places upon them they might as well be.

        Lacking that institutional motivation the public education system depends on good luck and the driven individual to overcome institutional indifference. Obviously that’s a sufficiently high mountain that few are good enough and enjoy the proper circumstances to climb it and those that do, the lucky, stellar performers, don’t become beacons guiding other teachers.

        Then there’s the political nature of the public education system which, while not overtly hostile to student achievement, relegates education to the role of just another consideration and a consideration without a powerful, political constituency and no obvious reason to elevate in importance.

        So Micheal’s Antonio Esuibel may have a salutory effect on the balance of the district but more likely, not. If his example isn’t followed, what’s the harm? To the professionals, there is none. To the kids? Sure but that’s not going to get anyone fired or result in any schools being closed down.

  3. Classes are overcrowded and they have a lot of disrespectful students in them. Solution: Expel 20% of the students using disrespect for authority as the criteria. Classes are smaller. Disruption is reduced. Remaining students take in a life lesson. School is a gift. If problems persist, repeat.

    • Ah, yes. The Joe “Batman” Clark solution.

      Not necessarily a bad idea – though expulsion for “disrespect” is a bit extreme.

    • Reluctant Teacher says:

      Great idea. It would instantly change the whole culture of the school.

      But this would require that administrators enforce the rules of their schools across the board for all stuents. Which rarely happens in my experience.

  4. Cranberry says:

    I teach 10th-grade English and journalism. My “10th grade” English classes are actually made up of ninth-, 10th-, 11th- and 12th-graders. The 11th- and 12th-graders are repeating the class because they failed it the first time. The ninth-graders are students who didn’t pass enough classes the first time they went through ninth grade to be promoted to 10th.

    With my scores, the district also sent a notice that, for reasons not explained, the 10th-grade scores were not considered reliable at this time, and so my overall score had been derived solely from the ninth-graders who happen to be in my 10th-grade English class. Because these happen to be my least motivated students, I was therefore judged not on my best students but on my worst.

    I’d say she’s doing well to receive “low average” when the only students she’s judged upon were those who couldn’t complete 9th grade the year before. To have any validity, teachers’ value-added score should be based on all the students she’s taught–not just the weakest.

    • “To have any validity, teachers’ value-added score should be based on all the students she’s taught–not just the weakest.”

      If the value-added statistics work correctly, it should not matter what students scores are used. It is based on how students do compared to what they were predicted to do by past performance.

      • Of course it matters which student scores are used. Any one teacher can only teach a finite number of students. If one removes the better teachers from the pool, the teacher’s performance is artificially depressed.

        Prediction is not as reliable as test scores. If it were, Wall Street would not have imploded under securitization. (I’ll note in passing that, in my opinion, it’s easier to predict whether someone could repay their mortgage, based upon income, than whether a student will be able to master Algebra.) Also, the principle of “GIGO” would apply: Garbage In, Garbage Out. If the statistical models are not based upon all students for all teachers, comparisons between teachers are garbage.

        • The whole stated goal of value-added analysis is to make it so that it does not matter which students a teacher is teaching. That way a teacher who has high ACHIEVING students and a teacher who has low ACHIEVING students can be fairly evaluated by not looking at achievement but PROGRESS, which relies on comparing predicted performance with actual performance. If value-added is successful in fairly comparing the two teachers with different sets of students, then it shouldn’t matter which students a single teacher is evaluated by as long as there are enough students to get an adequate evaluation.

          And no, not all students’ scores are used. Students must have previous test scores in order to make the prediction and I’m sure the value-added analysis throws out outliers.

          “If the statistical models are not based upon all students for all teachers, comparisons between teachers are garbage.”

          Why? Call up a random assortment of people in Texas and ask if they will vote democrat or republican in the next presidential election. I bet you can make some very, very good conclusions about how Texas will vote without calling everyone in the state.

          • Cranberry says:

            Election polls have been wrong. The more people you call, the more reliable your results would be. However, just surveying people doesn’t make the results predictive. Research has shown how the structure of a list of questions can bias the results. Also, you can’t predict the probable outcome of next year’s election from this year’s election results.

            You wouldn’t get good results in an election poll from a few dozen people. If the teacher in this case has about 200 students, and her class includes 9th, 10th, 11th, and 12th graders, it makes no sense to judge her performance only on the test results of the 9th graders. If half her students are on grade level, and the remaining half are evenly distributed by grade level, her performance would be judged on the results of 33 students?

            This is a variant of blaming the teacher if a student won’t learn. Yes, there are some bad teachers, but you don’t need to turn to a statistician to determine who they are. Receiving a value added prediction from the central office doesn’t mean the prediction is reliable.

          • No system of evaluation is perfect and none ever will be. Should we just give up on evaluating teachers?

            The more data the better, of course. But the data has to be the correct data.

            Creators of value-added modeling say that one years worth of data is not enough to evaluate teachers. They also warn against using the data to say there is a difference between say a 3 and 4 on five point scale. It should also not be the whole evaluation of the teacher.

            My original problem with the article and many comments about value-added modeling is that people don’t actually discuss the realities of how the method works. They simply start assuming that having better students will ensure better scores which is not true.

            I don’t know about you, but I have actually seen many years of value-added reports for myself and others. Amazing how little flux there is year to year. Teachers that score outstanding one year, do it again the next, and the next. Others with mediocre scores, get mediocre scores.

            It is not blaming the teacher. Some teachers will be better than others. Why not try to figure out which ones through evaluations and value-added analysis? I, as a teacher, can’t go to the students house and whip the parents in shape, but I sure as hell can find ways to do my best with the students I have. One way is to find teachers who perform well (value-added analysis?) with the same students and see what they do.

            You are correct that we can tell who bad teachers generally are and value-added data generally shows that those teachers are bad. Now we have two ways to show that the teacher is not performing. Why throw that away?

          • “would be judged on the results of 33 students?”

            There are ways to determine how many students is enough to create an accurate evaluation. I don’t know what the number would be. Is 33 students enough to produce a satisfactory confidence interval? I don’t know.

            This is not the argument that the teacher in the article made. That type of argument would show that she understands how value-added works. The teacher obviously thinks that if her motivated students were included her scores would be better.

  5. I don’t know exactly what they are doing in L.A. but some of what is being passed off as “value-added” isn’t really being done the way many of us, particularly those familiar with the Tennessee model, understand value-added. Some states that are sort of trying to pretend they are going to use value-added measure for teachers don’t even test every year. How valid are the growth projections going to be based on projections from a couple of years before or based on fewer years of previous scores?

    (There’s a pretty big motivation drop off from say, third to ninth grade. How can you account for that in VA measure if you don’t track the kids every year? )

    We need to watch the actual methodology closely, which is hard when some states won’t release their formulas.

  6. “If teachers can’t be evaluated fairly based on their students’ progress (compared to their previous progress’ rates) and they can’t be evaluated based on classroom observations, how can they be evaluated?”

    This is just my guess, but if some sort of survey could be given out to determine the circumstances of a student and his/her family regarding the environment within and outside of the home, then that could give an idea of how much of a student’s learning can or cannot be improved upon. To some extent, that would give a clearer idea of how much of a student’s progress is due to a teacher and how much is precluded by external circumstances.

  7. Homeschooling Granny says:

    Aren’t we like the drunk looking for his car keys under the street light because he can see there rather than looking in the dark where he dropped them? We criticize teachers because we can. It would be “blaming the victim” if we criticized a more significant source of the problem: homes and parents.

    Sociologist Brian D Ray has studied the demography of homeschool families. He describes families that look a great deal like those in the public schools where students do well: stable families that value education in stable living arrangements. There is also research showing that children in families that eat a meal together around a dinner table do better in school, a positive effect comparable to having books in the home.

    Children apparently can do well in school, whether public or private, or homeschooled, whether school-at-home, eclectic, or unschooled, but they don’t do well without adults who have time and interest for personal, intimate, committed attention that keeps them prepared to learn.

    Doesn’t all the blame-the-teacher media noise risk letting less responsible parents off the hook? How about a campaign to make adults aware of what they can do in their homes and communities to support success in school.

    • Teachers have to deal with what comes through the door. Some will be better at working with low-income students. Why not find out who they are and what they are doing?

      Could it be that the teacher in the article might actually be a below average teacher? 50% of teachers have to be below average.