States that use large-scale testing to make high-stakes educational decisions should ensure the tests are aligned with state curricula and provide a fair measure of students' learning, urged psychologists at a March 20 congressional briefing on educational testing, sponsored by U.S. Rep. Bobby Scott (DVa.).
The briefing was part of a broader effort in Congress to address the growing use of large-scale testing to make high-stakes educational decisions--for example, to determine whether students can graduate, be promoted from one grade to another or enroll in advanced or remedial classes.
Despite concerns that have surrounded high-stakes testing, such tests can be invaluable in helping to diagnose gaps in learning, said Eva L. Baker, EdD, a psychologist at the University of California, Los Angeles, and co-director of the National Center for Research on Evaluation, Standards, and Student Testing (CRESST). Baker testified on behalf of APA and CRESST at the briefing.
If states use scientifically valid tools in designing and implementing trustworthy tests, Baker said, "I believe that it is possible to systematically improve the quality of education in this country."
Government addresses high-stakes testing
Large-scale, standardized tests of student achievement have long been a feature of K12 education. And about half of U.S. states use large-scale assessments to evaluate students' qualifications for graduation or promotion from one grade to another.
President Bush's education proposal portends even greater reliance on educational tests for such high-stakes decisions. Among other things, his plan calls for testing children in reading and math every year from grades three through eight and granting federal education funds only to schools that meet state-determined achievement standards.
Because growing national attention on accountability in education has heightened the seriousness of such tests' consequences--both for individual students and for schools and districts--federal lawmakers are proposing legislation to help ensure that states deploy such tests fairly and appropriately.
This spring, U.S. Sen. Paul Wellstone (DMinn.) proposed a bill that would authorize the National Research Council (NRC) of the National Academy of Sciences to study the consequences of high-stakes testing for students, teachers and schools. That investigation would complement a 1999 NRC study that set guidelines for the fair and appropriate use of tests in making high-stakes decisions concerning individual students.
Wellstone has also introduced a bill that would implement the NRC's 1999 recommendations. As the Monitor went to press, Rep. Scott planned to introduce a companion to that bill in the House of Representatives.
"These bills uphold the science of testing and the appropriate use of large-scale educational tests, while safeguarding against adverse consequences for students," says Ellen Garrison, PhD, APA's director of Public Interest Policy. "The congressmen's concern about fairness in testing particularly applies to minority, low-income, disabled and limited-English proficient students who may not have had an equal opportunity to learn or to be evaluated properly, if at all."
Crafting scientifically sound testing programs
As states increasingly rely on large-scale educational assessments, it's crucial that the best scientific information should be brought to bear in developing and implementing assessment and accountability programs, said Baker. She said states should ensure that:
Tests be used only for the purpose for which they were scientifically validated.
High-stakes decisions not be made using the results of only one test.
Tests are aligned with states' curriculum standards, so that teachers can prepare students to succeed.
Tests measure only the academic domain of interest, without unwittingly emphasizing extraneous factors.
Tests are sensitive to school quality differences.
Baker urged states to consult the guidelines set forth in "Standards for Educational and Psychological Testing," published jointly by APA, the American Educational Research Association and the National Council on Measurement in Education. In addition, states should consult new standards for educational accountability systems developed by CRESST, in partnership with other organizations. Finally, Baker recommended, states should evaluate the effects of ongoing testing efforts.
Addressing testing concerns
Other panelists at the briefing addressed concerns that critics have raised about high-stakes educational assessments.
First, Dianne M. Piché, executive director of the Citizens' Commission on Civil Rights, noted that most states are not yet compliant with a 1994 federal law mandating that they develop comprehensive educational standards and accountability measures. She argued that President Bush's education plan, which is even more ambitious, may be unrealistic.
Critics have also expressed concern that high-stakes tests, if designed or implemented inappropriately, may draw an inaccurate picture of student achievement and unfairly jeopardize students or schools that are making genuine efforts to improve. Others worry that overreliance on testing might paradoxically compromise educational quality by leading teachers to "teach to the test," focusing their classes on narrow test-taking strategies rather on than on broader, conceptual material.
Many of these concerns are addressed by the NRC's 1999 report on high-stakes testing, noted psychologist Pat DeVito, PhD, director of NRC's Board on Testing and Assessment, who described the report's main recommendations at the briefing.
Other speakers at the briefing discussed worries that high-stakes testing systems might treat students inequitably. They applauded congressional efforts to institute safeguards for minority students, students with limited proficiency in English and disabled students.
And even the most scientifically sound and equitable tests cannot alone solve educational problems, noted Scott: "Once you find that a school is dysfunctional or a student isn't learning, then what do you do?"