Supporting Coherence Across a System of Assessment for Next Generation Science Standards

• In Event: 77.020 - Next Generation Science Standards Assessment Design Across Use Cases: From Formative Classroom Assessment to Large-Scale Accountability Assessment

Abstract

If assessments are to support and monitor teaching and learning, they must be embedded in a system in which instruction, assessment, and policy all move toward the same goals. Given the complexity of the NGSS and the substantial shift in science teaching and learning that they require, [blinded project name] has addressed the need for multilevel coherence integration between classroom and external assessments (Gitomer & Duschl, 2007) by supporting an assessment “ecosystem” in California. This approach attends to coherence across classroom and external assessments in conjunction with the other critical elements of a robust system that supports science learning, such as professional development and instruction (NRC, 2015). In this presentation, we seek to describe how [blinded project name] has supported the development of this system by focusing on state-level policy, high-quality assessment, and capacity-building.

The NGSS are based on a new vision of science learning for states, and to successfully implement these standards they require a substantial change in how student learning is monitored and evaluated (Gorin & Mislevy, 2013). To operationalize this vision, we outlined a theory of action and a system of assessment that could support the vision of science learning underpinning the NGSS (Authors, 2015). The system has four components, and uses a combination of short items that contribute to individual student scores, matrix-sampled short performance assessments that contribute to system monitoring during federal testing years, and short and extended performance assessments in classrooms.

An assessment system that will support the vision of learning underpinning the NGSS must use assessments that are able to measure competency-based performances in multiple dimensions (NRC, 2014). To illustrate such tasks, we created and piloted a bank of model assessments, scoring rubrics, and sample student data. The assessment development process draws on a situated learning framework (Lave & Wegener, 1991) and uses a combination of evidence-centered design (Mislevy, 2003) and construct modeling (Wilson, 2005) to develop engaging performance assessments that assess students along multidimensional constructs. Content validation was done using expert reviews and construct validation was done using an iterative process in which assessments were piloted and revised until student data showed evidence of each dimension being assessed across the range of progress variables. These assessments demonstrate a shift from the measurement of declarative, or even applied science knowledge, toward assessing students on the strength of their evidence-based reasoning about a phenomenon.

Efforts to broaden capacity in using performance assessment to support teaching and learning with NGSS led to the development of hybrid online/face-to-face courses that support professional learning communities. The courses have been provided to over 1,000 active participants over eight months, including teams of teachers, coaches, instructional designers, and principals.

These three sets of resources were used to cultivate an assessment “ecosystem” in California by framing and aligning discussions about NGSS assessment across networks of policy-makers, administrators, teachers, and PD providers. In our contribution to this symposium we will describe our approach to supporting the development of a coherent system of assessment for science and how it is being brought to other states.

Authors

• Jonathan F. Osborne, Stanford University

• Jill Wertheim, Stanford University