The Folly of linking teachers pay to test scores

From the blog school matters

by Rachel Squires Bloom

Part One: Students are Not Widgets: The Unfairness of Using a Business Model to Measure Teacher Effectiveness

Evaluating educator effectiveness by linking salaries to test scores would destroy the relationships on which sound teaching and learning are built.
While teachers are unquestionably responsible for the progress of their students, creating monetary incentive to “teach better,” as defined by the narrow confines of standardized test scores, undermines relationships among students, teachers, schools and systems. A model which relies on standardized test scores as indicators of teacher quality measures student performance very narrowly, and fails to account for relevant factors such as economic level, attendance, special educational needs, and language ability of students, all which influence student progress.

Tying test scores to salaries is based on the business model and its reliance on numbers (the “bottom line”) to assess and subsequently reward effectiveness. Numbers, after all, make great sound bites. They convey an often superficial tone of authority and trustworthiness, particularly if one does not dig too deeply, inquire about cofounding variables, or examine them in context. Numbers may be taken at face value by those who are reluctant to examine statistics more closely and who are comfortable with facile treatment of issues. The business model may be useful when gauging the efficiency of production of cars or widgets, but such a pay-for-productivity model is not appropriate for measuring student achievement. If widgets (or car parts, cans, bolts of fabric) are produced by machine, how many that are produced per hour can be measured, and a person running that machine can be said to “make production” or not. Schools, however, are not factories and students are not widgets. Teachers do not operate machinery, nor are they themselves machines. Transposing a math-based business model onto the human and humanity rich field of education is neither valid nor fair.

Flawed attempts to tie “trustworthy” numbers to quality, bean-counting masquerading as valid assessment, encroaches in other fields as well. Tying professional effectiveness to one particular measure in any field is flawed. Similar uses of contextless data occur increasingly in health care. Attempts are being made to use numbers generated from patient data to determine whether or not a physician is a “good doctor.” Statistics, such as blood pressure or sugar levels, are used to assess physician effectiveness without accounting for variables such as compliance, follow-up appointments, co-morbidity, and variables which, as in education, stem from poverty.

This would be similar to comparing the blood pressure statistics of a general practitioner with that of a specialist. A specialist’s numbers may appear different, since he or she is likely to see patients with more severe symptoms, and who may be more (or less) compliant with medications.

These patient numbers create a metric; metrics are easy to generate and may reveal information about and individual case. The danger in medicine, (as in education) is that because numbers are easy to generate, that the metrics themselves become part of the definition of quality, rather than being only part of the bigger picture of the quality of a doctor’s care or of a particular patient’s outcome. Like students, patients are individuals who arrive with histories that include factors beyond what can be remedied in classrooms or medical offices. Poverty is one such enormously negative influence on a child’s academic and physical health.

Enhancing the pay in many professions, including education, by relying on statistics is particularly misguided when teachers are responsible for student populations which do not traditionally succeed on standardized tests. Any profession contains multiple variables which interact and which enormously affect outcomes. While numbers are useful and sometimes necessary measurements, they can be misused or, worse, easily manipulated by those determined to simplify a complicated assessment of progress.

Quantitative data from state-mandated tests can provide useful information when taken in context with other variables affecting student performance. Valid uses include viewing student performance as one component in determining how well districts and schools align curricula with state frameworks. Data can provide limited insight to teachers as to how individuals or students groups perform in a subject based 40-50 questions. Using such scores to garnish or set salaries, however, is not valid. It is especially specious when this is done with the allege purpose of improving teaching and learning.

Standardized tests are part of the fabric of students’ and teachers’ lives, including the Scholastic Aptitude Test (SAT) and state tests such as Massachusetts Comprehensive Assessment System (MCAS) and Louisiana Educational Assessment Program (LEAP). Standardized test scores allow college admissions committees to sort and rank students from among thousands of applicants. Standardized testing has valid critics, including those who say that such tests discriminate by class because wealthier parents pay for expensive test prep, and that the tests themselves allows students to demonstrate understanding in a very limited way.

What would a formula tying teacher salaries to test scores look like? For teachers of lower elementary grades, whose students are not yet subjected to standardized tests, would teacher salaries be tied to reading fluency tests such as the Dynamic Indicators of Basic Literacy (DIBELS) or other assessments that are administered as early as kindergarten. How feasible is it to create a formula to reward a teacher when so many teachers influence a particular student at a particular time? In many cases, it is a confluence of teachers who are responsible for a student’s performance.

Struggling students in may work with a classroom teacher as well as with resource room, skill support, or English Language Learner teachers.
If salaries were tied to test scores, how would such factors as economic status, home support, parent engagement, culture, and English language skills be calculated? Adequate nutrition, sleep and exercise affect ability to pay attention, to process and retain information, and to sustain task engagement are factors in achievement and in test-taking ability. Children who lack these basic securities, which research has linked to performance and achievement, are unlikely to attain the same test scores, even with the same teacher and curriculum, of students who take these things for granted.

Teacher effectiveness includes quality of curriculum, instructional methods, knowledge of students, and the development of lifelong learners.

Standardized testing already compels teachers to reluctantly narrow curricula. Teachers are pressured to race through units before students thoroughly understand concepts, to encourage memorization rather than thoughtful understanding, and to teach items out of context because “they might appear on the test.” In the weeks before a state-mandated exam, students are most likely doing review work or being force-fed topics which appear at the end of the school year after tests are taken. What is lost is enormous. Students are deprived of opportunities to explore topics in depth, with curiosity and reflection. Insufficient time is given to activating background knowledge, interest, and relevance, all which increase retention, engagement, and the desire to learn.
Good teachers provide instruction that creates well-rounded, curious students. How is teacher effectiveness and quality currently measured?

All states have mandated requirements in order to be licensed. Exams are in the areas of reading, general knowledge, and in content areas. In Massachusetts, to receive an initial teaching license an individual must possess a Bachelor's degree, passing scores on the MTEL (Massachusetts Tests for Educator Licensure) and must have completed an approved educator preparation program, which includes pre-practicum and practicum work. To gain a professional license, teachers work for three years under an initial license and must complete an approved teacher induction or mentorship program. Teachers apply for relicensure every five years, during which they must demonstrate evidence of relevant coursework or professional development to assure that they remain current in the field. These standards are part of what define a teacher as “highly qualified.”

In addition to licensure, teachers are regularly evaluated by principals or administrators. These evaluations, which become part of a teacher’s permanent personnel file, may make recommendations for professional teaching status or may recommend areas for remediation. It is the role of the principal or administrator to assure that teachers who struggle receive the support that they need in order to be effective. Of course, poor and ineffective teachers exist and it is crucial that administrators understand the power and the necessity of honest evaluations. Such evaluations may document poor teaching, provide remediation for struggling teachers, and in necessary cases, provide a route out for teachers who are not effective or who are harmful to student learning.

Evaluation of educators will never be a perfect science because much that teachers contribute cannot be measured by numbers. An effective teacher is one under whose tutelage students make progress, progress being far deeper than can be measured by a multiple choice test. Tying salaries to scores is an inaccurate, unfair, and ultimately injurious, creating neither incentive for sound teaching nor sound learning. This model is particularly inappropriate when the “products” are not products at all, but are human beings with unique minds, histories and learning needs.

This is what “effective” teachers know.

http://www.schoolsmatter.info/2011/03/folly-and-unfairness-of-tying-teacher.html?spref=fb