It’s widely accepted that teaching to the test is a bad thing. The phrase itself summons images of robotic children reciting factoids and filling in bubbles with their number 2 pencils. The practice is said to suppress creativity and critical thought, and crowd out subjects and topics that aren’t tested. Classrooms become joyless.
But what if test prep didn’t mean memorization drills and a stripped down curriculum? What if students couldn’t tell the difference between assessment and a really interesting lesson? What if the test was something teachers wanted to teach to?
That is one of the challenges the “next generation assessments” are attempting to meet. Two state consortia (PARCC and SMARTER Balance) are currently at work on tests aligned to the Common Core that will “measure the full range” of the new standards. This will include the abilities to think critically and creatively, reason mathematically and communicate effectively — skills that aren’t typically associated with standardized, multiple choice tests.
The consortia greatly benefit by the involvement of the country’s leaders in performance assessment, that is, tests designed to mimic rich “authentic” classroom instruction. The $346 million in federal dollars doesn’t hurt either. Still, developing these tests while making sure they meet high psychometric standards for validity and reliability is an enormous task. Even if they succeed at creating the instruments, will the tests have the desired effect on teaching?
A new report from the RAND Corporation seeks to answer the question, can standardized tests be worth teaching to? The authors of New Assessments, Better Instruction reviewed over 130 studies examining the relationship between high stakes testing and instruction. They offer a qualified “yes”— as long as certain conditions are met. These conditions don’t only relate to the test design, but also to policies governing implementation and accountability that lawmakers at all levels should heed.
First, the tests themselves. According to the report, the assessments should “mirror high-quality instruction.” So-called “deeper learning” isn’t easily modeled in multiple-choice format. The authors caution that producing rich assessments may require sacrificing some reliability related to scores, but add that the resulting impact on instruction may be worth it.
The RAND authors also recognize the importance of finding the right balance for accountability. They write:
Research on the ways that high-stakes tests influence instruction suggests that if the test is unimportant or irrelevant to students, teachers, administrators, and parents, it is unlikely to have an effect on instruction. On the other hand, if there are very high stakes attached …. there may be severe “teaching to the test” that does not promote real deeper learning but focuses on superficial features of items.
The solution? Multiple measures for teacher and principal evaluation , accountability based on growth as well as status, and an integrated system that includes low-stakes, formative assessments for tracking student progress along with the high-stakes, end-of-year summative tests.
Already examples are emerging from the consortia that illustrate what tests worth teaching to might look like. The CPE team has been having a lot of fun trying out the SMARTER Balanced practice tests. We admit some of these items had us momentarily humbled, but we still maintain we’re smarter than a 5th grader. Try them out for yourselves here.
More importantly, we think that assessments like these could drive deeper learning and instruction especially if combined with sensible accountability policies.