egorical programs might be lost, however, so it will be important that greater flexibility is accompanied by efforts to ensure that the needs of students with disabilities continue to be addressed in more integrated settings.
Accountability for the education of children with disabilities in integrated settings would be enhanced by both including these students in ongoing large-scale assessments as well as developing standards of practice. A 1997 report (National Research Council, 1997) extensively reviewed the issues involved in incorporating students with disabilities into standards-based curriculum and assessment reforms. At present, there are no generally recognized standards, linked to desired outcomes, providing benchmarks for determining what constitutes an appropriate education for students with various kinds of disabilities. The diversity of the characteristics of students with disabilities poses challenges to the development of professional standards, as does the fact that traditional categories of disability do not have a demonstrable relationship to specific outcomes or to prognoses (Epps and Tindal, 1987; Kavale, 1990; Kavale and Glass, 1982). Nevertheless, progress is being made. Efforts have begun to establish diagnostic constructs based on a child's placement along a number of continuous dimensions of disability, rather than an either-or dichotomy (National Research Council, 1997). At the same time, others (e.g., Reschly, 1996) are working on a service-based approach to identify outcomes that could be associated with certain levels of service investment for broad clusters of students with disabilities.
New approaches at the state and federal level to special education finance like census-based funding and less reliance on individual entitlement and classification pose potential risks to localities. Some of these risks will be magnified if school finance becomes more school-or pupil-based (rather than district-based) in the future. One risk is that local districts or schools may have unusual concentrations of students with disabilities, a fact that census-based funding would fail to address. We suggest that states that move to census-based funding ought to allow appeals when there is evidence of unusual concentrations of students with disabilities. A second risk is the financial drain that students with severe disabilities can pose for schools or small districts. These students with disabilities are, as we have noted, the easiest to identify. We suggest that states consider establishing risk pools to pay the ''excess costs" of such students.
A major question about special education finance is whether the federal government should in the future pick up a greater share of the unfunded mandate it has created for states and local districts. Not only is the federal share currently very low, but IDEA imposes a large regulatory overlay especially on districts and schools. At the same time, new approaches to special education at the state level pose both possibilities and challenges for federal officials. When states move to fund their special education responsibilities through census-based funding rather than individually identifying all students with disabilities, they lose out on federal aid under current arrangements. It is not clear that continued reliance on individual categorization is necessary at the federal level, though we recognize this is