Benchmarking assessments or curriculum-embedded formative assessments created in the context of a curriculum are designed to elicit student thinking and are referenced specifically to an interpretive framework. While few science-specific studies of benchmarking assessments have been completed, there is a large research base on benchmarking assessment systems in other subject matter areas. Some well-developed programs that are based heavily on benchmarking assessments have shown positive student learning effects. Success for All, for example, uses reading tests at 6-week intervals to determine the effectiveness of reading instruction and to regroup students for subsequent instruction. Instruction based on the principles of mastery learning, a system developed by Benjamin Bloom in which students are allowed to progress on the basis of demonstrating proficiency on a set of formative assessments, has been shown to have a significant positive effects for lower achieving students and for inexperienced teachers (Block and Burns, 1976; Guskey and Gates, 1986; Whiting, Van Burgh, and Renger, 1995).
There are a few published studies of science-specific benchmarking programs and others are in progress. Currently the Berkeley Evaluation and Assessment Research Center (BEAR) (2005) is creating embedded assessments for the Full Option Science System. The assessments are being developed to help teachers of students in grades 3-6 assess, guide, and confirm student learning in science. These assessments make use of construct maps, which model levels of student understanding of a particular construct (e.g., students’ ability to reason with evidence) on the way to developing proficiency (Wilson, 2005). BEAR has helped to develop and refine the associated assessment frameworks, items, scoring guides, and other elements of the system and will later provide support in the process of psychometric data analyses.
In a recently completed study, the Stanford Education Assessment Laboratory explored Black and Wiliam’s (1998) contention that formative assessment would increase student learning by developing curriculum-embedded assessments for the Foundational Approaches to Science Teaching (FAST) curriculum (Yin, 2005). The first unit of FAST guides students through a series of investigations to culminate in an explanation of floating and sinking on the basis of relative density. Assessments were embedded at key conceptual “joints” in the curriculum, following a developmental trajectory of understanding density that students were expected to experience. Twelve sixth and seventh grade teachers were selected from a pool of FAST-trained volunteers. Teachers were matched in pairs according to school characteristics, and one member of each pair was then randomly assigned to a control group, which would teach FAST as they normally did, while the other was assigned to an experimental group, which would implement the curriculum-embedded assessments. The experimental group teachers attended a 5-day