Indispensable Tests: How a Value-Added Approach to School Testing Could Identify and Bolster Exceptional Teaching

Issue Brief

Executive Summary

Most parents want the comparative information that tests supply about students, schools, and school districts. Tests also serve as a hedge against grade inflation, and help ensure that education standards are more than mere suggestions.

However, high-stakes tests do raise legitimate concerns. By excluding large numbers of disabled or limited-English-proficient children from the testing cohort, school officials sometimes can game the tests to pad scores and reap favorable publicity. In addition, another kind of unfairness arises when raw scores from a school serving an affluent neighborhood are compared straight-up with those for a school serving the most impoverished families.

A technique of using test data to determine the value that teachers add to the learning of each student – that is, how much they help the child progress academically – could become an important tool for building better teachers and better schools. Value-added assessment makes it possible to isolate the impact of the individual teacher and to respond with appropriate rewards or corrective training. It ensures that teachers are not penalized for accepting the challenge of raising the learning level of the most socially and culturally deprived children. The deeds of those who add great value to children’s learning are brought to light, as is the work of teachers who make little or no impact.

Former University of Tennessee statistician William Sanders, who pioneered value-added research, has demonstrated that pupils’ ability to succeed in school is severely crippled when they draw poor teachers three years in a row. Value-added assessment offers a way to ensure that as few children as possible suffer such a disadvantage.

Details follow.——————————————————————————–

Indispensable Tests: How a Value-Added Approach to School Testing Could Identify and Bolster Exceptional Teaching

by Robert Holland

High-stakes testing of students has become the sharpest arrow in the quiver of would-be reformers of public elementary and secondary education. Not without justification. No matter how high standards are set for English and math and history and science, education requires some form of measurement to confirm that students are mastering the material. Unless there are consequences for failure on these tests, students and school staffs will not take them seriously. Without testing, standards are mere suggestions.

Moreover, as public opinion surveys have repeatedly shown, parents by substantial margins support standardized testing. A survey done for the Committee on Economic Development found that about three-fourths of respondents thought the quantity of student testing in their communities was either about right or not enough. (CED 2000). Parents believe testing yields valuable information. They want to be able to compare their children and their schools to other children and their schools. Information is power. Tests provide information to consumers, and results will be critical as education begins to operate more like a free market. Objective testing is like an annual doctor’s check-up, or an audit by an external auditor.

Tests also serve as a valuable confirming device for grading systems that vary from school to school and teacher to teacher. Bill Evers, a research fellow at the Hoover Institution and member of Hoover’s Koret Task Force on K-12 education, observes that a pupil’s grade is based not merely on achievement and exertion but what the teacher expects from students in his or her class. One teacher may have lower expectations for students than a second teacher. One teacher may want to grade students more leniently than another teacher. Grade inflation may result in many students receiving As and Bs without actual improvements in achievement.

“A divergence between grades from classroom teachers and scores on standardized tests,” notes Evers, “can be a wake-up call for parents, taxpayers, and school boards – telling us that students don’t really know the subject matter and that teachers are too soft in their grading practices. Getting rid of standardized tests is like getting rid of thermometers, X-ray machines, and blood-pressure gauges in a doctor’s office.” (Evers 2001)

Although data from mass testing may jolt schools into the first stirrings of reform, problems undeniably accompany so-called high-stakes testing. There is, for instance, the political element: How high a rate of school failure is politically acceptable? When Virginia launched its Standards of Learning testing in 1998, for instance, fewer than 3 percent of public schools initially could meet the seemingly modest requirement that 70 percent of each school’s students pass the tests of core-subject knowledge. The passing rate has gradually climbed (reaching 40 percent in Spring 2001) but Virginia still faces the prospect of stripping accreditation from hundreds of its public schools as the 2007 deadline approaches, or lowering the standards it has been pushing the schools to achieve. Similarly, the complex formula for Title I testing in the version of federal reform recently negotiated on Capitol Hill had Washington’s policymakers puzzling over just how many schools they could afford to fail. Moreover, the problem of failing schools pales in comparison to the problem of what to do when high numbers of students cannot pass end-of-course or graduation tests.

The bottom line is that high-stakes tests increase the likelihood of school officials and policy-makers gaming tests to attain the results desired. Simply by excluding from the testing large numbers of disabled or limited-English-proficient children, a local district or a state may pad the scores and reap undeserved praise in the press. Another kind of unfairness occurs when scores from a school serving affluent families are compared straight up with scores from a school serving children who’ve grown up in the poorest of neighborhoods.

Value-Added and the Difference It Makes

A fair and objective system would identify the difference that schools and teachers make with the individual children they serve – the value that they add to each child’s learning, so to speak, regardless where that child started in school or what kind of advantages or disadvantages he or she brings to the classroom.

So-called value-added assessment holds much promise of being that method of testing. Properly refined, it could have its greatest impact on the quality of teaching in American K-12 schools ø revolutionizing the way teachers are trained, hired, evaluated, retrained, rewarded or encouraged to enter another line of work.

Speaking before the metropolitan school board in Nashville in January 2001, Tennessee statistician William Sanders rattled the education establishment when he disputed the connection commonly made between poverty and low student performance. “Of all the factors we study – class size, ethnicity, location, poverty – they all pale to triviality in the face of teacher effectiveness.” (Long and Cass 2001) Such a pronouncement rubbed against the grain of doctrine instilled in many educators during their training and retraining. As the perceptive education historian Diane Ravitch has documented, the dominant education progressivists long have maintained that many children shouldn’t be pushed to absorb knowledge beyond their l