Formative Assessment as an Approach to Improving Education
Executive Summary
Robert E. Floden & Lorrie Shepard
The review was aimed at informing the Hewlett Foundation’s decisions about future investments for work on formative assessment. To gather information, we read the reports from Hewlett-funded work on formative assessment, followed up with phone interviews and web-based demonstrations of tools, read prominent reviews of research, interviewed experts on formative assessment, and discussed our preliminary ideas with Hewlett staff. As we considered the implications of the information, we considered the place of formative assessment in the context of other education reforms and of the current emphasis on assessment and accountability, especially through No Child Left Behind.
Context of the Problem
Among the pressing problems in US education are the gaps (between rich and poor; between White and Black) in student achievement and the overall low US achievement levels found in international comparisons. The national response to these problems, prominent in No Child Left Behind, has been an emphasis on accountability and an increase in the amount and frequency of standardized testing. While the increases in testing have the potential for improving achievement and narrowing the “gap,” the increases may also have substantial negative consequences, such as the diversion of effort from teaching deeper conceptual understanding and subject areas not tested to teaching test-taking skills and content knowledge that closely imitates the tests.
Assessments take many forms, with different purposes and consequences. These differences are often obscured or ignored, leading to policies that fail to achieve the most important educational goals. The difference between formative and summative assessment is one such key distinction. Research showing dramatic improvements in learning from formative assessment has inspired policies in the name of formative assessment, though the assessments implemented often lack critical “formative” features.
A literature review by Black and Wiliam in 1998 found that “formative assessment” has dramatic effects on student achievement, with effects typically 0.4 to 0.7. The instructional approach was most effective with low achievers, suggesting that it might help both to raise overall achievement and narrow the “achievement gap.”
“Formative assessment,” as defined in the literature review, has several key features:
It operates on a relatively short cycle.
It stays close to the classroom (rather than reporting for school or district).
It is connected to change in practice.
It is directed at both adjusting instruction and enhancing pupil motivation and self-regulation.
The logic model explaining the effectiveness of formative assessment is based in psychological theory, empirical research, and common sense. If teachers and students regularly monitor frequent student progress towards instructional goals, making links to ongoing classroom activities, teachers and pupils will both be able to modify classroom work in ways that address gaps in learning and build on successes. Their academic engagement can be based on an understanding of what has been accomplished, when learning has gone off track, and where work still needs to be done. Moreover, if students understand the learning goals and gain proficiency in judging their own progress towards those goals, they are more likely to direct their energies toward reaching the goals.
Like many other appealing ideas in education, however, formative assessment faces formidable implementation challenges. The studies in the Black and Wiliam review, while mostly set in classroom instruction, typically were of modest scale and duration. Subsequent intervention studies by Black and Wiliam and other investigators have required teams of researchers to help teachers learn formative assessment principles and adapt them for their own practice. To date there have not been well documented studies of efforts to incorporate formative assessment into the ongoing, long-term practice of school systems. For taking formative assessment to scale, salient challenges are:
(a) making information about pupil achievement readily and frequently accessible to teachers and pupils,
(b) reducing the burden of gathering and reviewing the information, and
(c) linking the insights gained to specific suggestions for what to do next.
In the current context of intense pressure to raise scores on external accountability tests, an additional challenge for implementation of formative assessment is the competing investment by school districts in interim and benchmark assessments. These quarterly summative tests provide information about which students may not be on track to pass the end-of-year tests. They are sometimes called formative assessments but typically do not provide real-time information in the course of instruction about what students are and are not understanding. Interim and benchmark assessments are easier to implement than the teacher professional development needed to support real formative assessment, but they do not meet the criteria for effective formative assessment identified in the research.
Hewlett’s Grantmaking Work On Formative Assessment
With the policy context’s emphasis on summative assessment, the substantial promise of formative assessment can only be realized if some organization is persistently supporting and promoting formative assessment, with explicit attention to the differences between formative and summative assessment. Among the major education funders, Hewlett has been alone in making the case for formative assessment and supporting its development. The major testing firms have made occasional contributions (e.g., the ETS project led by Dylan Wiliam), but they have more often renamed their assessments, “formative,” adding to the difficulties with implementing true formative assessment.
The Hewlett Foundation has funded development of tools for use in formative assessment (FAST-R, STEP, Agile Mind), associated studies of these tools (Murnane, Education Matters, MDRC), work with policy makers (ACHIEVE, Aspen Institute, CPRE), literature reviews (CPRE, SRI), and implementations of approaches (New Teacher Center, work supported by New Schools Venture Fund).
The work done by Hewlett grantees has resulted in the creation of formative assessment tools now beginning to be used to give professional developers and teachers easier and quicker access to displays and reports of pupil achievement, with connections to informative IRT-based scales for early literacy. These tools address the first two challenges for implementing formative assessment listed above, making information quickly available to teachers (and often students), without adding greatly to teachers’ workload. Initial experience with these tools, expecially Agile Mind, FAST-R, and STEP, has demonstrated that teachers and students can make regular use of assessment information when it is closely connected to a curriculum in a single subject. The next step to increasing the value of these tools would be linking assessment information with recommendations for action built on research on learning progressions or on the practical experience that curriculum developers gain from the teachers using their curricula.
Hewlett work has also added to the field’s understanding of the system-wide attempts to implement formative assessment. The Study for Instructional Improvement studied the formative assessment components of America’s Choice and Success for All, finding that teachers and students could sometimes sustain formative assessment processes in their writing program by creating classroom routines for reviewing students’ work in comparison to shared criteria for good work. O’Day’s research on San Diego reform has described the ways that some schools have built up routines for monitoring student progress in literacy, with regular opportunities to display patterns of student achievement and discuss ways to address the needs of particular groups of students.
Hewlett’s support of work with policy makers played an important part getting them to embrace formative assessment in ongoing initiatives such as the CCSSO’s conferences on Formative Assessment for Students and Teachers (FAST). That embrace is a step forward, but continued work is needed to help policy makers understand the differences between formative and summative assessment. Policy makers will be drawn to simple solutions, and to traditional assessment systems.
Recommendations for Future Investments
We recommend that the Hewlett Foundation continue to support work on formative assessment, with strategic investments in three areas.
1. Efforts that enhance the development and use of formative assessment in the context of exemplary research-based curricula and teacher professional development projects. The Hewlett Foundation should use funds strategically to build upon on-going work supported by other agencies.
a. Because formative assessment is most likely to be fruitful when conducted in the context of particular content and curricula, Hewlett should encourage and support work to build additional, explicit attention to formative assessment into curriculum development projects, such as Agile Mind, Connected Mathematics, FOSS science, and BSCS. With additional support targeted at improving assessments and assisting teachers and students in using them, assessments could be built on the learning progressions in the curricula and could be built into the routines teachers use to work with the curricula.
b. Hewlett should bring together scholars from areas where research related to formative assessment is underway, under a variety of names. Science educators working on learning trajectories, cognitive scientists working on tutoring systems, special education researchers working on adaptive instruction are all working to link real-time assessments of students to changes in instruction. Hewlett can push research on formative assessment forward by bringing these scholars together to learn from one another.
2. Research to study the implementation of systems for formative assessment. School systems are beginning to adopt some curriculum packages and tools that attempt to support formative assessment (e.g., Writers’ Workshop, Agile Mind, FAST-R) but NCLB and the general press for quick fixes are likely to shift attention towards summative assessments. The work should attend both to teachers’ use of assessments to change instruction and to students’ use in learning to monitor their own progress. Research is needed to understand the conditions under which educators are able to sustain the formative emphasis, in the face of accountability pressures. Continued support for research that highlights formative assessment will be crucial to taking advantage of the potential for achieving the large possible effects of formative assessment at scale.
3. Work to engage and inform the policy community. Hewlett’s past efforts have played an important part in persuading the policy community that “formative assessment” has great promise. Policy makers, however, seem likely to translate their enthusiasm into a simple increase in the amount of testing, without attention to student motivation and to teachers’ need for real-time information to inform instruction. Unless continuing efforts are made to educate policy makers about the particular features and benefits of formative assessment, the initial enthusiasm may do more harm than good. The Foundation can make important continuing contributions, such as helping policy makers see that increasing the use of formative assessment will require work with content experts around particular curricula, rather than simply increasing the frequency of district-wide testing.
References
Black, P., & Wiliam, D. (1998). Assessment and classroom learning. Assessment in Education: Principles, Policy, and Practice, 5(1), 7-73.