New and Better Assessments

Following up on an earlier post, the debate on assessment policy turns on the question of how much information we really need or even want to have about student performance. At the level of rhetoric, of course, we want as much information as we can get, and we want that information to be as accurate as possible. As a practical matter, though, information has costs, and we are prepared to settle for a lot less than we would ideally like to have. The real question for policy is how much less.

A new PACE report, co-published with the Rennie Center for Education Research & Policy examines new approaches to assessment that promise more and better information about what students know and are able to do than current “snapshot” tests can provide. (See also the podcast from “21st Century Assessments” conference in July 2010). The report includes papers on computer-adaptive assessment (by Mark Reckase from Michigan State), on the assessment of English learners (by Robert Linquanti from WestEd), and on virtual performance assessments (by Chris Dede and Jody Clarke-Midura from Harvard). The assessments foreseen in the report would greatly increase our knowledge about what students are learning, and where they’re struggling.

From the point of view of educational piety this is unalloyed good news. Better assessments can provide a much clearer account of students’ strengths and weaknesses, and guide instruction in ways that better respond to their needs and interests. The papers in the recent PACE/Rennie Center report give good reason to think that addressing them would produce substantial improvements in the performance of students and schools.

But from the point of view of policy, sadly, the case is less clear. Having access to more and better data about student performance would seem to imply some responsibility to make good use of the new information, and doing so would require significant changes in the way schools work. Vast new quantities of information would have to be compiled, organized, and presented to teachers and parents in ways that make them both comprehensible and useful for instructional planning. Curricula would have to become more flexible, to support teachers’ efforts to adapt instruction to the different learning trajectories of individual students. Most teachers would have to be trained to interpret data on student learning, and many would need help to expand their portfolio of teaching strategies to take account of detailed information on each student’s strengths and weaknesses.

All of these challenges are surmountable, but overcoming them would require a lot of hard work by lots of different people. The real policy question is whether the gains from better assessments would justify the cost of all that work, or whether the flawed assessments on which we now rely already give us enough information to accomplish our core educational objectives.