Computer-adaptive tests are designed to adjust their level of difficulty—based on the responses provided—to match the knowledge and ability of a test taker. If a student gives a wrong answer, the computer follows up with an easier question; if the student answers correctly, the next question will be more difficult. Considered to be on the leading edge of assessment technology, computer-adaptive tests represent an attempt to measure the abilities of individual students more precisely, while avoiding some of the issues often associated with the “one-size-fits-all” nature of standardized tests.
For students, computer-adaptive testing offers a shorter testing session with a smaller number of questions, since only those questions considered appropriate for the student are offered. On the other hand, test developers have to create a larger pool of test items so that testing systems have enough questions to match the varied abilities of all students taking the exam. The most current forms of computer-adaptive testing are typically administered online, and because the scoring is computerized, teachers and students can get test results more quickly than with paper-and-pencil tests.
Computer-adaptive tests can be used for a wide variety of purposes, including large-scale, high-stakes testing; formative assessment, which provides teachers with in-process feedback on student learning that they can use to modify instructional techniques; and summative assessment, which educators use to determine what students have learned at the end of a unit, term, or year. They are also used to identify students who may need specialized academic support in a specific skill or subject area, such reading, writing, or math.
Since computer-adaptive testing systems select questions that are intended to be appropriately challenging for each student, most students will get about half the questions right and half wrong, so a score based on the total number or percentage of correct responses will be meaningless. Therefore, computer-adaptive scoring is based on both the number of correct answers provided and the difficulty of the items completed. Before the tests are administered to students, test questions are typically field-tested with representative samples of students to calibrate difficulty levels.
While computer-adaptive technology is a relatively recent development, its use appears poised to grow significantly in the United States over the coming years. For example, two major national assessment initiatives, the Smarter Balanced Assessment Consortium and the Partnership for Assessment of Readiness for College and Careers (PARCC), intend to use the technology. In general, computer-adaptive testing is being adopted by states and schools based on the following rationales:
- By using more precise and efficient assessments that take less time to complete, teachers and students will have more time for instruction and learning, while still getting test results that are either just as accurate as traditional tests or more accurate.
- The tests tailor each question to the knowledge and abilities of the test taker, removing the need for students to struggle with questions that are too difficult or spend time on questions that are too easy.
- The tests can provide more precise and quickly available information on student learning needs, which teachers can use to adapt instruction and improve academic support for students.
- Test security is improved because not all test takers see the same items.
Since computer-adaptive tests are still relatively new, debates about their use, reliability, benefits, and shortcomings are just beginning to emerge. With many states planning to use new computer-adaptive online tests in coming years, the technology will likely become the object of increasing scrutiny, discussion, and debate.
In addition to the potential benefits described above, the following are a few representative arguments that may be made by advocates of computer-adaptive testing:
- The tests can help to identify a student’s learning level more precisely than fixed-question exams, especially for students at the lower and higher ends of the learning spectrum.
- Adaptive tests give teachers more precise information about students who are exceptionally adept or exceptionally far behind in their mastery of expected knowledge and skills.
- The tests may increase student engagement in the testing process, and possibly lead to more accurate results because the tests are shorter, less tiring, and better calibrated to a student’s individual abilities.
- Computerized scoring of open-ended and essay-style questions is becoming more accurate, and may even become more reliable than human scoring, which could increase the efficiency and reduce of costs of large-scale standardized testing.
The following are a few representative arguments that may be made by critics of computer-adaptive testing:
- The sophisticated technology needed to score open-ended questions and essay sections on computer-adaptive tests is not yet ready for widespread use in schools. Systems may not have been sufficiently tested and others may be prone to glitches and errors, which could lead to inaccurate results that can disadvantage the students taking the tests, resulting in the need for human scoring.
- The use of computerized tests could disadvantage students with lower technological literacy and less access to digital technology, such as students from lower-income households and students in rural areas with less-reliable internet access.
- Transitioning from paper-and-pencil exams to computer-adaptive tests entails, in many cases, significant logistical challenges and financial burdens, particularly for cash-strapped states, districts, and public schools. The sophisticated software required for the tests—whether it is custom developed or an off-the-shelf product—can be expensive and potentially cost-prohibitive.
- For schools with few computers or inadequate computing networks—or both—it may be prohibitively difficult to allocate the time and computers needed for all students to complete a test.
- Computer-adaptive testing typically requires robust technical support because broken or malfunctioning systems can derail test administration and significantly disrupt school scheduling and operations.
- For districts and schools that still rely on paper-based processes, transitioning to online, computer-adaptive testing might be burdensome or infeasible because a school may not have sufficient resources, devices, and on-staff technical expertise.
The Glossary of Education Reform by Great Schools Partnership is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.