School-testing policies need to be reexamined and revamped to realize the goals for classroom assessment outlined in this volume. In terms of large-scale testing for program monitoring and accountability purposes, the Standards stress that “[A]ssessment policies and practices should be aligned with the goals, student expectations and curriculum frameworks” (NRC, 1996, p. 211). Designing quality assessments that assess the full range of science education is particularly important considering the significant impact “high-stakes” testing has on classroom teaching and assessment practice (Darling-Hammond, 1994; Gifford & O'Connor, 1992; Linn, 2000; Oakes, 1985,1990; Resnick & Resnick, 1991; Smith, Hounshell, Copolo, & Wilkerson, 1992).
Accountability is an important feature of our educational system, but the current models used for implementation of accountability policies are far from sufficient. Useful data often is generated at the expense of detrimental effects on classroom practice and student learning (Darling-Hammond, 1994; Gifford & O'Connor, 1992; Gipps, 1994; Goodlad, 1984; Smith et al., 1992). In particular, high-stakes external assessments often drive curriculum and classroom assessment practices (Frederiksen & Collins, 1989; Haney & Madaus, 1994; Linn, 2000; NRC, 1987).
External assessments should be criterion-referenced and aligned to the recommendations of standards, curriculum, and instruction. Not only will effective assessments match exemplary instructional practices, they also will assess what is important and valued, not solely what is easily and inexpensively measured. In science education, this includes assessing and supporting the development of inquiry as a key component of science education. To support the Standards' vision of quality science and to attend to inquiry will require that we reconsider ways in which to involve teachers in the large-scale assessment process.
One alternative solution may be for school districts to establish their own local standards-based accountability system in tandem with those of their state. Ideally, carefully designed tests would allow districts to develop and implement assessments in core curriculum areas, such as science, that match their learning goals, demonstrate student achievement of standards, inform instruction, guide professional development, and demonstrate program impact. Teachers can take a role in the development of these assessments. The assessment development situation in Delaware as described in Chapter 5 offers such an example at the state level. However, what works for a small state like Delaware may not be as feasible for larger states, especially where curricular decisions are