ratios. This finding has been reinforced by the results of a substantial education experiment conducted by the state of Tennessee and summarized by Hanushek.

Another important trend during the 1970s and 1980s was a rise in the level of teacher training, as measured by the percentage of staff with master's degrees. In addition, the average level of experience for teachers has risen with the decline in enrollments. Both contribute to higher teacher salaries and add to the costs of education. Again, studies of school performance show no effect of additional teacher training and experience on student performance. While Hanushek agrees with Smith and his colleagues that past education reforms have not been successful, the PEER report diagnosis is that the nation's school system has failed to provide appropriate incentives for teachers and students.

Hanushek, like Smith and his colleagues, emphasizes the importance of explicit goals and of developing measures of performance that relate to these goals. However, Hanushek also emphasizes the creation of specific incentives directed at improving student performance, a matter Smith would leave to state and local officials. Examples of performance-based incentives include merit pay and hiring, promotion, and retention of teachers on the basis of classroom performance. The core of this approach would be performance-based contracting between school districts and teachers. This would be phased in through a two-tiered contracting system, leaving the existing seniority-based contracts in place for teachers already in the system. Given anticipated rates of retirement, this could produce a rapid transition to the new system.

Some of the incentives would be for schools rather than individual teachers. For example, merit school programs could key school budgets to student performance. School choice could provide incentives for better student performance by engaging the interest of parents and students in good schools. Mechanisms that permit choice, possibly including private as well as public schools, could align better student performance with the allocation of school resources. Hanushek emphasizes that incentives for performance are largely untested, since they have been introduced so infrequently. One attractive approach is to experiment with different incentive systems, coupled with evaluation of the resulting impact on school performance.

A final point of disagreement between Hanushek and Smith is the measurement of school performance. Hanushek and Smith agree that better assessments are essential. Hanushek observes that schools resist evaluation; this is echoed in Smith's negative evaluation of existing education assessments. Smith and his colleagues propose to align these assessments with education standards established at the state level. Hanushek propounds a "value-added" approach, based on differences in performance between students entering the school system and those leaving the system.

Under Goals 2000, federal education policy charges states with standard setting. Hanushek would have states promote incentive schemes for performance as well as establish standards and set overall education policy. Education policy



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement