Next Generation Science Standards Assessment: Lessons Learned From the Stanford Next Generation Science Assessment Project

• In Event: 47.088 - Overcoming Challenges in Developing and Implementing Next Generation Science Standards–Aligned Instructional Materials and Assessments

Abstract

In this presentation, we seek to present the challenges and issues that the Stanford Next Generation Science Assessment Project (SNAP) have met and dealt with in their work to support a research-based approach to the implementation of the NGSS in California. Funded by the Bechtel Foundation, we have faced three challenges.

The first has been how to encourage state leaders to develop a coherent system of assessments that could monitor and support the shifts in science teaching needed to successfully implement the NGSS (Gorin and Mislevy, 2013). SNAP’s approach was to work with an invited group of national and international experts, and California Department of Education, to outline a model coherent system of assessments for NGSS (Osborne et al., 2015). The proposed model blends individual student measures and matrix-sampled performance tasks to provide a mix of data on individual and system performance. This policy paper was used to frame discussions with state leaders, and ultimately informed the system that was chosen.

The second challenge has been how to develop a model of a new vision of science assessment. The NGSS demands the integration of science content, practices, and crosscutting concepts and shifts assessments toward a multidimensional competency-based measurement of performance (Koeppen, Hartig, Klieme, & Leutner, 2008). In approaching the development of assessments that can meet these requirements, we conducted an extensive search both nationally and internationally for exemplar items that might offer models and frameworks on which to build. However, the three-dimensional framework of NGSS poses a unique challenge which has not been addressed in other contexts (Wertheim et al., 2016).

The design of our assessments was grounded in a framework of situated learning (Lave & Wegener, 1991), where students take the role of scientist as they engage with real questions and problems, discourse, model building, and experimentation, peer and teacher feedback and revision. Assessment development followed a combination of construct modeling (Wilson, 2005) and learning-centered design (Soloway, Guzdial, & Hay, 1994) to define multidimensional performance outcomes and progress variables, and to develop tasks that would be engaging, educative, and a valid and reliable measure of each of the targeted outcomes. The resulting assessments are used as models to communicate with educators, administrators, and policy-makers about this vision of science assessment and how it can support a coherent science assessment system.

The third challenge has been capacity-building. If SNAP’s recommended system design is to support the substantial shifts embedded in the NGSS, all participants in this coherent system must know how to use them to support that vision (Gitomer and Duschl, 2007). To this end, SNAP developed innovative online/face-to-face courses to support professional learning communities in using the assessments. These courses have had over 1,000 active participants, including teachers, coaches, principles, and PD providers since November, 2017.

The three main components of SNAP’s work have served to provide concrete tools to inform the assessment system that California has developed. In our contribution to this symposium we detail the challenges that we have met, addressed, and those that still remain.

Authors

• Jonathan F. Osborne, Stanford University

• Jill Wertheim, Stanford University