are actually being provided in the classroom and whether the curriculum being offered meets content standards. In addition, a series of smaller studies could address particular aspects of testing and learning.
FOR FURTHER ANALYSIS
-
How can the nation ensure that assessments are used appropriately and fairly?
-
Under what conditions is it appropriate to use assessments for high-stakes applications?
-
How can we extract reliable and useful information from heterogeneous data elements that emerge in performance assessment?
A NEW ERA OF EDUCATION FEDERALISM
THE ISSUE IN BRIEF
Goals 2000 and the Elementary and Secondary Education Act reauthorization legislation have far-reaching implications for the federal, state, and local compact on education. Goals 2000 establishes a framework for standards-based reform, codifies eight national education goals in federal law, authorizes funding and other incentives to encourage states to adopt and implement standards, calls for participating states to develop assessments aligned with standards, and authorizes federal money to develop and evaluate new assessments. The ESEA legislation revises the testing and accountability requirements of the Chapter 1 program for disadvantaged children (renamed Title I).
Both Goals 2000 and ESEA contain reassurances about the voluntary nature of national standards, vest primary control of standards and assessments in the states, and establish a partnership between local communities and the federal government. What types of governance relationships are implied by the new legislation? What are the potential impacts of the federal government on state and local policies? The workshop spurred new thinking about these questions.
VIEWS FROM THE WORKSHOP
Title I: A Major Influence
Included in the ESEA reauthorization legislation are the outlines of a new system for testing and accountability under Title I. This system, analyzed in Phyllis McClure's presentation and ensuing discussion, would replace the current Title I testing procedures, which are based on national aggregation of norm-referenced test data and which have been criticized for promoting undesirable instructional approaches for disadvantaged children and for producing information of questionable quality and utility. National aggregation of local Title I test data would be abandoned; instead national information would be obtained from a national assessment that used a matrix sampling approach.
◦ ◦ ◦ ◦
“Title I is really the 800-pound gorilla that is going to drive Goals 2000. What Title I says about assessment is what school districts are going to follow.”
Phyllis McClure
◦ ◦ ◦ ◦
In addition, the House version of the bill required states to adopt content, performance, and opportunity-to-learn standards for Title I children that are the same as those for all children and that are aligned with the Goals 2000 standards. States would also develop or adopt state assessments to measure the proficiency of Title I children in core academic subjects. These assessments would be administered at some point during grades 3-5, 6-9, and 10-12 and would provide individual student scores, as well as disaggregated results for certain subgroups. Assessment results would also be used to gauge the progress of schools and districts in helping Title I children meet performance standards. Sanctions and rewards stronger than those in current law would be tied to these evaluations.
With over $7 billion in federal dollars at stake and with three-quarters of the school districts in the nation participating, the Title I amendments may prove to be more consequential than Goals 2000 and, in effect, could set the parameters of a state's general assessment system.
Several concerns emerged from the workshop regarding the Title I amendments. One question revolved around the decision in the legislation to use the same system of standards and assessments for multiple purposes, from measuring individual student progress to enforcing institutional accountability. As an alternative, it was suggested that individual student assessments and institutional accountability were different functions requiring different measurements: for the former, schools could use multiple measures designed by teachers, and for the latter, standards-based assessments administered through matrix sampling.
The new provisions could actually increase the amount of testing attributable to Title I, it was argued, if multiple assessments are developed in all core subjects. Questions arose about how the multiple measures called for in the bill would be applied to state and local accountability decisions; whether the measures would meet reliability, validity, and other technical criteria; and whether the new assessments will be appropriate for high-stakes uses. Further questions focused on how to maintain baseline information on individual student achievement if assessments are administered only at certain grades, possibly beginning as late as grade 5. In many ways the new system could be more problematic than the one being replaced, warned presenter Michael Kean.
Other issues are whether the three-year period for developing new Title I assessments will be adequate and where the funding will come from to develop and pilot the mandated assessments and train teachers in their use.
◦ ◦ ◦ ◦
“In the name of reform, we are about to create a more complex, more technically problematic, more burdensome, and perhaps less useful assessment system.”
Michael Kean
◦ ◦ ◦ ◦
Certifying Standards and Assessments
Goals 2000 establishes a new entity, the National Educational Standards and Improvement Council (NESIC), with the main responsibility for “certifying” standards and assessments—a challenging and complex task, participants said. There is no single definition or widely accepted set of criteria that make assessments certifiable or uncertifiable. Rather, experts can only analyze whether assessments meet various technical and other criteria. It was suggested (although this option is not specified in Goals 2000) that NESIC might produce a range of judgments about the relative strengths and weaknesses of specific assessments on different criteria.
Several governance issues are left unanswered by Goals 2000. For example, what is the standing of voluntary national standards that are authorized by federal legislation, developed by national but nonfederal panels, certified by a federally established body that includes nonfederal representatives, and offered as a model to guide state standards but not control state curricula? How will these standards affect state governance systems, especially when the states must answer to constituents and potential litigants? How will they influence local behavior from several layers removed? To what extent are state standards expected to be aligned with