Neill, Monty
FairTest

Prepared by Lynn Davey and Monty Neill based on testimony presented by Monty Neill to the House Subcommittee on Select Education, this article argues that current efforts to establish a national test to measure progress toward the nation's educational goals will inhibit, not advance, educational reform.

U.S. policymakers and the public have been deluged with proposals for national testing. Several proposals called for single tests, primarily multiple-choice, in different subject areas in a number of grades. Among these were Educate America and a plan mentioned in "America 2000" to develop individualized tests based on National Assessment of Educational Progress test items.

However, the National Education Goals Panel and the National Council on Educational Standards and Testing have both endorsed development of a national examination system based on performance assessments. This idea was initially proposed by the Learning Research Development Center and the National Center on Education and the Economy, which created the New Standards Project. They hope to design model exams that states and districts can use to develop their own specific curricula and tests.

Dozens of national and local education, civil rights and advocacy organizations, including those associated with the FairTest-initiated Campaign for Genuine Accountability in Education, have consistently opposed any type of national exams at this time. They have argued that all national exam proposals put the cart of testing before the horse of educational reform and that the harmful effects of this effort will fall most heavily on low-income and minority-group children.

ISSUES

All national testing proposals are based on the fallacy that measurement by itself will induce positive change in education.

Proponents claim that other nations, whose students score higher on some international exams, have national testing programs. In fact, no significant economic competitor of the U.S. has a single national exam or exam system of the sort proposed for the U.S., and none of them uses tests for accountability purposes. If these nations out-perform the U.S., a point open to debate, it is not because of national tests. These nations do, however, have far smaller percentages of children living in poverty.

The multiple-choice testing option is especially dangerous.

Evidence from the standardized exam driven school improvement efforts of the 1980's demonstrates conclusively that this type of testing does not, in fact, improve education. (See the papers from the American Educational Reform Association special conference on national testing, Phi Delta Kappan, November 1991.) All multiple-choice/short-answer tests are severely limited. They cannot directly assess higher order thinking, problem solving abilities, creativity, or initiative.

As an accountability tool, a national multiple-choice test would lead to drilling for the least important skills and perpetuate damaging models of instruction and educational excellence. Learning how to use knowledge in academia, employment and society would be relegated to an even smaller part of the curriculum. Necessary school reform would be stymied.

A national standardized, multiple-choice exam would likely perpetuate sorting students by class and race. Under the guise of tracking for "ability," students are often segregated by race and class on the basis of test scores. This is partly because the tests make cultural assumptions through the language used and the experiences the tests treat as normative, assumptions that work to the detriment of minority-group and low-income children. While there are real differences in the educational opportunities of poor and rich students, standardized tests exaggerate these differences by their biases and confuse lack of ability with lack of exposure.

Performance assessment attempts to assess what students know and are able to do.

They are based on observing, documenting, and analyzing student work. Projects and products from in and out of the classroom, research, writings, and exhibitions are among the materials used in such assessments. Portfolios are tools for selecting from, analyzing, and summarizing these types of student work over time. Performance assessments can also be examinations developed and administered from outside the classroom. Examples include open-ended complex problems that require the student to explain how the problem was solved, science fair exhibitions, performances in applied arts, and Scout Merit Badges.

Like multiple-choice tests, however, performance assessments should not be transformed into a national examination system, at least not at this time.

There are many issues that must be resolved before it is reasonable to consider a national system:

Lack of consensus on educational practices and outcomes. There is no widely accepted set of skills and attributes that are expected as outcomes of education. This must be resolved before national exams would make any sense. Subject area groups suggest it could be a decade before complex debates over curricular content and instructional methods are resolved, if ever.

None of the proposals adequately address issues of equity. These exams must not become gatekeepers that perpetuate sorting students by race and class under the guise of "ability". Bias issues in performance assessments have not been addressed adequately.

Testing by itself will not improve education. Assessment reform must be only part of an integrated process of systemic change which addresses issues of curriculum, instructional practices, staff development, school structure and governance, textbooks, and schools of education. Equitable access to educational resources and other necessary prerequisites of student learning also should precede national exams.

Tests tell only part of the story. Any proposal must be part of an overall educational information system. Having outcome information is not useful unless we also have adequate information on the educational context (type of community, socioeconomic status of students, school climate), resources (expenditure per student, teaching staff, building quality, textbooks and materials), programs and processes (class size, curriculum, instructional methods, grouping), and other outcomes (graduation rates, employment, further education, "customer satisfaction"). All this information must be organized into an information system designed to ensure continuous school improvement.

The technology and benefits of comparing scores on different assessments from around the nation (calibration) is untried. Because regional exams will be based on national standards and model exams, supporters of the exam system claim each student's scores could be compared with those of all other students through a system of "calibration." We do not know if such calibration is possible, how expensive it would be, or whether the benefits will be worth the expense. Money would be better spent to support comprehensive educational reform and disseminating useful information.

We do not know who will be in charge or the consequences of one national body setting standards and approving exams. Will a national curriculum result? Will the system be sufficiently flexible to allow all students to develop and demonstrate their capabilities? How will those who think standards should change be able to influence a national body?

RECOMMENDATIONS

Instead of national testing, we suggest the following steps to improve education and assessment:

The federal government should assist states and districts with the development of performance assessments; teacher education and staff development; and the development and dissemination of model curricula, standards, and assessments. All these should be integrated into comprehensive reform strategies.

Re-examine cases in which the federal government requires multiple-choice testing, particularly for the Chapter I program.

Consider assessment not in isolation but as part of a comprehensive educational information system.

Finally, require that any assessment system be evaluated on the basis of extensive experience at the state and district level. Equity issues, in particular, must be adequately resolved before any decision is made on a national examination system.

CONCLUSIONS

U.S. students need school reform, not more testing. More test scores will not magically produce educational improvement. Resources should be spent on helping teachers teach and students learn, not on further sorting and ranking students, schools and states.

New assessments should be used to help student learning, guide educational improvement, and enhance equity. National testing will short-circuit these fundamental reforms.

REFERENCES

Medina, N. and D.M. Neill (1990) Fallout from the Testing Explosion Cambridge, MA FairTest.

Oversight hearings on Educational Assessment. Hearings before the Subcommittee on Education and Labor, House of Representatives, 102nd Congress, Serial No. 102-8. (ERIC TM 017 236)

Phi Delta Kappan, November 1991, special section on national testing.

Perrone, V., Ed (1991) Expanding Student Assessment: Alexandria, VA: Association for Supervision and Curriculum Development.

FairTest, 342 Broadway, Cambridge, MA 02139 is an education reform and civil rights group monitoring testing activities.