Challenges Envisioned for Next-Generation Science Tests
Amidst growing fights over common state standards and tests for mathematics and English/language arts, the separate set of common K-12 science standards have been quietly gaining steam, but some experts predict that measuring student progress under these standards may require even more testing innovation than the common core.
Less than six months after the final Next-Generation Science Standards were released by a consortium of 26 states and several national groups, seven states have adopted and started to implement them: California, Delaware, Kansas, Kentucky, Maryland, Rhode Island, and Vermont. Researchers and education officials from those and other states met in Washington last week at a summit sponsored by the Education Testing Service, the Council of Chief State School Officers, and the College Board to discuss new ways to measure what students learn from them.
The standards ask students to take a much deeper approach to understanding science concepts and processes, and to apply their knowledge through scientific experiments, investigations, and engineering designs. As a result, the standards will require educators to measure not just how well students can recite facts or predict experimental outcomes, but also to gauge how well students develop and use conceptual models on their own, develop and follow lines of investigation, and communicate research findings.
“We’ve got a lot of experience assessing content” in science, Jonathan Osborne, a science education professor at Stanford University in Palo Alto, Calif., said during the summit. “We have a checkered history [of] assessing more than that, and particularly practice.”
Teachers may have a tough time gearing up for complex assessment tasks because more than a decade of focus on reading and mathematics has in some areas squeezed the time in the school day for science, according to Kathleen Scalise, an associate education professor at the University of Oregon in Eugene.
“Learning time for science is greatly reduced, and opportunity to learn this type of scientific thinking is often absent,” Ms. Scalise said. For example, she noted, performing proficiently in 8th grade science on the National Assessment of Educational Progress has been highly correlated with the amount of out-of-school science enrichment students receive, which, she says suggests that schools already have little time to dedicate to complex science tasks, and less time to prepare for similarly complex test items.
“Students are often asked to look at something once and come up with a claim,” noted Alicia C. Alonzo of Michigan State University, in East Lansing, who analyzed the test requirements of the science standards. “Rarely are they asked [in the test] to go back and revise their claims in view of the evidence.”
To assess these sorts of skills, tests likely could not use multiple-choice or short-answer problems, Ms. Alonzo noted, and the tests would have to measure both students’ answers and the process by which they arrived at them.
“What does it mean to think scientifically, to reason scientifically, and use community-normed discourse?” said Karen Lionberger, the curriculum director of science for the College Board, which administers the Advanced Placement tests. The board recently overhauled its science courses and tests to meet more sophisticated science mastery. “You have to give students the opportunity to fail in conceptualizing a model, then go back to the drawing board and think.”
High- and Low-Tech
Other large-scale science assessments, such as NAEP, the Advanced Placement science test, and the Program for International Student Assessment, are already changing their test items in ways that may inform the next-generation tests.
NAEP and PISA, for example, are rolling out new hands-on and interactive computer tasks which give educators a better understanding of how students work through a science problem.
New problems in NAEP science and technology and engineering literacy assessments allow students to work through game-like scenarios, such as finding and fixing a village’s broken water pump or laying out art in a gallery exhibition for the best traffic flow. The test software automatically captures what, when, and how frequently the student clicks different parts of the scenarios to re-create how he or she worked through the problem.
“It’s about capturing the actions, the ‘how’ of student learning,” explained Peggy G. Carr, an associate commissioner for the National Center for Education Statistics, which administers NAEP.
The increasing use of technology could be a problem for districts with less capacity, but test designers Angela DeBarger and Christopher Harris of SRI International, in Arlington, Va., said more multifaceted questions are being created for pen-and-paper assessments, too.
For example, Ms. DeBarger described an incomplete drawing of the ocean floor where two plates are moving apart. Middle school students were asked to complete the drawing to show the motion of the earth’s mantle, explain why that motion affected the plates, and mark the oldest parts of the land created by that motion. The problems require students to understand multiple processes and convey them in writing and pictures, with points given for different levels of sophistication at each part of the problem.
Regardless of how states assess the new science standards, researchers and educators alike agreed that the new tests should support and improve science instruction, and they will have to be paired with teacher professional development.
“There’s an opening of thinking about these sorts of things. ... Current assessments can’t cover the range and depth of what is in the NGSS,” said Mark R. Wilson, an education professor at the University of California-Berkeley. He said that tests must match instruction, rather than the other way around.
“We need not to rush into this,” he said. “We don’t have a federal law pushing us to do this, so ... we need to take our time. I’m worried that the assessment will get ahead of the instruction.
Vol. 33, Issue 06, Page 6Published in Print: October 2, 2013, as Challenges Predicted for Next-Generation Science Tests