Many of us have spent more time in hospital emergency rooms than we’d like. The one nearest to where I live serves part of Harlem, and I’ve seen people arrive on foot or in emergency vehicles in pretty bad shape. In some cases, there’s an acute condition that has just emerged; when watching the triage process, however, it becomes clear that many patients have a long history of chronic disease, and the ER visit is just a snapshot in a longer trajectory of growth or, sadly, decline.
The ER doctors and their associates diagnose the problem conditions, and use their training and experience to prescribe a course of treatment. In some cases, the patients are admitted, and after a few days are discharged back to their daily routines. But the severity of the ailment matters. When a patient arrives with chest pain due to a heart attack and has a balloon angioplasty two hours after arrival, there’s an eight percent likelihood that she or he will die in the hospital, and a 20 percent chance of death within seven years.
No one I know blames ER doctors for the fact that their sickest patients often die. To be sure, some ER doctors may be more skilled than others at diagnosing problems or recommending treatments. And some hospitals may be overcrowded, or lack up-to-date equipment. But there’s a widespread recognition that the best predictor of a patient’s prospects for survival and a good quality of life after treatment is the patient’s condition when he or she arrives in the emergency room.
The state of New York recognized this when it first sought to compare the performance of hospitals and physicians across the state. New York’s “mortality report cards” looked at the mortality rates for hospitals performing coronary artery bypass surgery, adjusting for the mix of risk factors for each patient undergoing the procedure. Patients in poor health are likely to be readmitted quickly or to die within 30 days, regardless of which hospital they go to, and it doesn’t seem fair to penalize a hospital in its rating for the fact that high-risk patients come to it.
The notion of risk adjustment is central to value-added measurement in education, the effort to identify the unique contribution of a teacher or a school to a student’s measured academic achievement. Value-added models are “risk-adjusted,” in the sense that they take into account factors outside of the teacher’s (or school’s) control, such as a student’s prior academic achievement or socioeconomic background. (How well they do this is a matter of debate, and given the imprecision of these models, I am on record opposing their naïve use for high-stakes decisions regarding schools and teachers.)
New York Gov. Andrew Cuomo, a Democrat, doesn’t seem to grasp the concept. The Opportunity Agenda book accompanying his State of the State address on January 21st states:
“Last year, less than one percent of teachers in New York State were rated ineffective; but state test results show that statewide only 35.8 percent of our students in 3rd through 8th grades were proficient in math and 31.4 percent were proficient in English Language Arts. We must ask ourselves: how can so many of our students be failing if our teachers are all succeeding?”
A few years ago, the governor would not have asked this question. In 2011, 53 percent of students in the state of New York in grades 3-8 were classified as proficient or above in English Language Arts, and 63 percent were proficient or above in math. Just two years before, in 2009, 77 percent of students were judged proficient or above in English Language Arts, as were 86 percent of students for mathematics.
The composition of the teaching force hasn’t changed appreciably since then; all that’s changed are the academic standards and the assessments tied to them. Setting standards is a political process infused with values. Teachers across the state of New York haven’t suddenly gotten worse; rather, their students are being asked to do more.
The governor’s diagnosis, however, is that the problem lies in our state’s teachers. The treatment? Increasing the role of standardized tests in annual evaluations of teacher performance, and requiring that teachers have five consecutive ratings of “effective” or “highly effective” to be eligible for tenure. Both are highly speculative. There is, for example, no research basis for weighting value-added measures of teachers’ contributions to their students’ test performance as 50 percent of the overall annual evaluation. The Measures of Effective Teaching project funded by the Bill & Melinda Gates Foundation cautioned that weighting any one component in a teacher evaluation system too heavily may make it hard to identify teachers who are contributing to a broader set of learning goals than those represented on the standardized tests. (Disclosure: the Gates Foundation has been among the many funders of The Hechinger Report.)
And the governor’s proposal stacks the odds against any new teacher obtaining tenure. We know that beginning teachers often need time to develop in the classroom; struggling for a year or two as they adapt to school and classroom realities is not at all uncommon, but might well result in an annual rating of “developing.” Those years wouldn’t count toward tenure.
If the governor’s proposal were enacted, and the annual distribution of ratings paralleled the existing value-added component—with 1 in 6 teachers rated as “developing” or “ineffective”—I estimate the odds of a teacher earning the due process protections of tenure within six years at just under 50 percent. In the meantime, the governor’s proposal would allow the firing of any probationary teacher at any time without cause. This would make entering teaching in the state of New York a very high-risk career choice.
If low student achievement is the problem, what’s a better diagnosis and treatment? We could start by honoring the state’s obligation to fund school districts at a level adequate for a sound basic education. Since 2007, the state legislature and a series of governors have ignored the New York State Court of Appeals’ ruling to direct several billions of dollars of funding annually to the state’s neediest school districts.
Anyone who doubts that poverty and district finances matter in the achievement equation need only look at a scatterplot of the percentage of third-graders in a school who are proficient in English Language Arts and the percentage of students in that school who are eligible for a free or reduced-price lunch. In schools where 90 percent or more of the students are eligible for free or reduced-price lunch, 15 percent of the students are proficient. Conversely, in schools where fewer than 10 percent of students are free-lunch-eligible, an average of 53 percent are proficient. This is not because all of the good teachers are in low-poverty schools.
To be sure, correlation does not imply causation. But in this case, the correlation should cause some skepticism about the governor’s diagnosis and treatment plan. Adequate funding and support for New York’s public schools are the best medicines.