Designing Common Core Tests For All Proving a Challenge
Although more students with disabilities than ever are included in state testing programs, the task of giving these students high-quality assessments in the future that measure how adept they are at mastering the Common Core State Standards seems to have an endless number of hurdles to overcome before students face these new assessments in the 2014-15 school year.
And one of them has less to do with the test than with instruction, said Stephen N. Elliott, a professor of education at Arizona State University. Elliot spoke Tuesday at a U.S. Department of Education meeting addressing the challenges that remain in preparing new tests that all students are scheduled to take in 2014. This was the fourth meeting about the assessments.
In his research looking among several states, Elliott found that the most time any state was able to spend on teaching the standards was 81 percent of the time students were in school, and special education teachers covered even less of the content and standards.
"We get that test score and we make that big inference that kids have been taught this," Elliott told a collection of special education and testing experts, including members of the consortia that are designing common-core assessments and alternate assessments for students with significant cognitive disabilities. "Many students with disabilities need 30 to 40 more days of class time to get an equitable opportunity to learn."
That problem may only grow as the common-core standards are implemented.
"It's not that we can't improve assessments, [but] that can serve as a distraction from the critical need to improve instruction, said Lou Danielson, a managing director at the American Institutes of Research who focuses on special education policy and evaluation.
Progress, But More Work to Do
The major hurdle of increased, improved instruction aside, the technical and content issues with the exams are numerous.
While students with disabilities have become a bigger part of state accountability systems, albeit gradually, over the last 20 years, now even students with the most significant cognitive disabilities are included in state testing programs. "It's remarkable how far we've come," Danielson said.
One fundamental advantage to designing tests with students with disabilities in mind from the beginning is that for the most part, tests won't have to be retrofitted and adapted to work with students with disabilities after the fact, a situation that is typical with many current state assessments.
But there's still a long way to go, he said.
One big issue lies with computer-adaptive tests, in which students are given a question that's harder or easier depending on whether they answered the previous question right or wrong, pose a challenge for students with disabilities. Some students (with and without disabilities alike) may shut down if they miss the first question. And there's the risk that the computer will throw a student a question that's below his or her grade level because of one or multiple wrong answers, a situation that concerns special education advocates.
Another issue: Though it's not a universally popular idea, Danielson said the time has come to put an end to portfolio-style assessments, used in particular for students with severe cognitive disabilities. There are questions about how reflective the collections of work are of students' ability and how heavily influenced they are by teachers, who may be evaluated on their contents in some cases. Proficiency rates that are too high may indicate expectations for students with disabilities who are assessed this way are too low, he said.
"It's too tempting for teachers to help a lot with their [students'] work," Danielson said.
Accommodations Require Agreement
Yet another issue is that states using exams developed for most students by one of the two consortia working on those tests will have to agree on a common set of acceptable accommodations on those tests. One particular sore spot is whether, or how much, students should have test instructions or test content read aloud to them.
"I feel when reading reaches a point where it's about comprehension and they still have trouble decoding it, it becomes a test of decoding," Danielson said. "In earlier grades where decoding is being tested, it makes sense not to read aloud."
Students may waste a lot of their time when they hit unfamiliar proper nouns, reducing their fluency and comprehension, he said research shows. And students using digital text in class where read-aloud features are common may be stumped on tests where these features aren't allowed.
Keeping the principles of Universal Design for Learning in mind as the tests are developed is critical, said Sheryl Lazarus, a senior research associate at the National Center on Educational Outcomes, which is leading one of two groups of states in designing alternate assessments for students with severe cognitive disabilities. This will help ensure that test items measure what they're intended to measure, not students' test-taking or other abilities. Pictures and graphics should be clear and only used when essential. Text should be concise and clear. Test items could be designed to appear one at a time on a computer screen, and test items should be consistent in their format from one to the next. Navigating the exams shouldn't be confusing.
"This is an opportunity," Lazarus said. "I beg of... the consortia to take advantage of this opportunity and make the most of it."
Alexa Posny, assistant secretary for special education and rehabilitative services at the U.S. Department of Education, said one proposed accommodation has already been the subject of concern: sign-language avatars. If students are used to live sign--language interpreters in class, the avatars could be startling and awkward to use. "Students shouldn't suddenly encounter an accommodation they haven't used in the classroom," Lazarus replied.
And Dan Wiener, administrator of inclusive assessment for the Massachusetts Department of Elementary and Secondary Education, asked about situations in which meeting the needs of one type of student with disabilities conflicts with what another student needs, such as colorful graphics for a student who is deaf or hard of hearing that would be difficult to work into a Braille version of the test. Will every test item have to be scrutinized for its accessibility, he wondered. And would this time-consuming process have to be repeated yearly? "It really does sound like pie in the sky," he said.
Lazarus mentioned the Accessible Portable Item Profile Standards, or APIP, which can help to tailor tests to different kinds of students
Of note: The meeting made several references to a format of tests allowed to an extent under No Child Left Behind: alternate assessments for another group of students with disabilities, tests that are expected to disappear when No Child Left Behind, the current version of the Elementary and Secondary Education Act, is reauthorized.
There was an entire presentation by Shelley Loving-Ryder of Virginia's education department on this version of exams her state has developed. She showed examples of how test questions have been modified, such as by providing only three choices instead of four on multiple-choice questions, providing graphics when students have to calculate something about a figure such as a cylinder, providing formulas required to solve some problems, color-coding some test items, and so forth. Presumably these are the kinds of modifications to exams we could see on future common core assessments, even if these tests in particular—so-called 2 percent exams—are obsolete by then.