Learning Progressions in Science: Analyzing and Deconstructing the Multiple Dimensions in Assessment

Assessments are a crucial tool to enable the success of teaching and learning in science classrooms. Hence, to realize the vision of current reform efforts assessments must be developed that (a) measure the higher order cognitive skills that are essential to those reforms, and (b) that report results in ways that are readily accessible and interpretable. Through this project, researchers will develop internet-based assessments that capture such learning outcomes. These assessments can influence policy and practice by providing tangible products that exemplify the kind of learning outcomes and performances expected in today's science classrooms. The measures will also ensure that all students are assessed fairly. This project has the potential to enable students to express their knowledge and skills in a variety of ways which are less demanding and more creative than typical in traditional assessments.

Researchers will develop assessment materials aligned to the Next Generation Science Standards practice of scientific argumentation, the cross-cutting concept of patterns, and content in each of the following foundational middle school science domains - physical, life, and Earth sciences. Researchers will work closely with administrators and educators to ensure the relevance and alignment of materials to teachers' needs. Data will be gathered from middle and high school students from ethnically, culturally, and linguistically diverse school districts in California and Arizona. Key measures to be developed include embedding questions about patterns into the nine existing scenarios (tasks) and in new scenarios about natural resources. The team will develop assessments that can be used in both open-ended and selected response formats, to enhance their usefulness to teachers for diagnostic understanding, and their efficiency for summative classroom use. Data will also be gathered from teachers, to help develop interpretational materials. Teachers will have access, in real time, to their own students' responses, and estimates of students' performance on learning progressions within each of the science practices, crosscutting concepts and domains mentioned. The team will use the BEAR Assessment System to develop and refine assessment materials. This system is an integrated approach to developing assessments that seeks to provide meaningful interpretations of student work relative to cognitive and developmental goals. The researchers will gather empirical evidence to develop and improve the assessment materials, and then gather reliability and validity evidence to support their use. In total, item response data will be collected from several thousand students across the two districts. Student response data will be analyzed using multidimensional item response theory models.