A similar argument can be made for special education referral and assessment, though in this case some new costs are implied. Early universal screening and intervention for reading problems would impose additional costs. While the tests themselves are short and easy to administer during a regular class period, the training required for teachers to properly administer and interpret the test results, and to learn to work effectively in a classroom with multiple reading groups, represents a considerable up-front cost. Texas currently provides two days of training to every teacher administering the TPRI.2 We would point out, however, that this training could become a standard requirement of pre-service teacher education in the future. School districts currently spend resources on professional development. Were the proposed program adopted, some portion of existing inservice training resources could be devoted to this training. The same is true of training in classroom management and universal behavior interventions. And offsetting the costs of new assessment techniques are the savings from replacing a time-intensive assessment process centered around IQ testing.
The largest sustained cost of our recommendations is likely to be quality early intervention services for the highest risk children that begins before birth and is maintained throughout childhood. Experience with these interventions suggests that doing them well is expensive—on the order of $7,000 to $12,000 per child per year. Provision of quality preschool services to children from poor families would also require substantial resources, since only about one-third of those children now receive Head Start services. As Chapter 4 indicates, the average cost per child of Head Start services is approaching $6,000 per year. But the upfront cost of the program would be offset by reduced special education and grade repetition, and long-term benefits to society from higher reading and mathematics achievement, increased employment and earnings, and reduction in teen pregnancy and crime.
Similarly, changing the data collection done by the U.S. Department of Education need not cost more in the long run. If the data collected are expanded as we recommend, one would expect higher costs of additional reporting. However, duplication of reporting that currently exists would be eliminated, potentially offsetting any increase in cost.
Lastly, the research we propose both with regard to early childhood programs and general, special, and gifted education is not an incremental increase in existing efforts. Rather, it is a substantial investment in carrying through important research findings to classroom applicability; an effort that is likely to require substantial new resources. In virtually every other sector of the economy, research plays a substantially larger role in improv-