Recommendation 1: Funding agencies should support research on both: (1) the ways in which students’ understanding of the fundamental concepts of science develop over time with instruction, and (2) the ways in which students represent their understanding of these ideas as they develop greater expertise.
Recommendation 2: To assist states in their efforts to make more effective use of assessment results for improving curriculum and instruction and for diagnosing student needs relative to reaching the standards, funding agencies should support research on the ways in which tests could be designed to produce more useful subscores and the ways in which those subscores could be used effectively by teachers and others.
NCLB requires that state science assessments be aligned with state content and achievement standards and that they include multiple up-to-date measures of student achievement that are valid and reliable for the purposes for which they will be used. To meet these requirements, states will need both assistance with developing and validating new forms of assessment and with better incorporating and aligning all aspects of their assessment and education systems while meeting standards for technical quality when systems of assessment are involved.
For example, in an assessment system, assessments and combinations of them would need to be reliable and valid for every level and every purpose for which they are used. Current strategies for thinking about technical quality are not focused on thinking about systems of assessment. New methodologies for judging these concepts across different tests and across different levels (e.g., classroom, school, school district, and state) are needed. Similarly, strategies for conducting alignment studies among multiple components of an assessment system and the state standards are needed. Such strategies need to focus on the collective alignment of all the tests and tasks that constitute the assessment system, yet researchers are still struggling with ways to conduct such studies for a single assessment.
Recommendation 3: Research on the design and validation of science assessment systems should be conducted. Among the subjects investigated should be strategies for using classroom assessments for accountability purposes and instruction and procedures for determining the alignment, reliability, accuracy, and validity of assessment systems composed of multiple measures. Federal funding agencies and others should support these research efforts.
Most state science standards recommend that students understand and develop appropriate skills related to scientific inquiry, yet, as we have discussed,