When a test is not using a fixed form or fixed set of questions but instead adjusts based on input from the test-taker, it is computer-adaptive. The idea is that the questions change based on the students’ responses as they are taking the test.
In practice, there are many different types of computer-adaptive tests. The SBAC, a state standards proficiency assessment used by many states for accountability testing of their students, is computer-adaptive. It uses adaptive logic to decrease overall test time by reducing questions in certain areas.
Other assessments may use computer-adaptive logic in order to improve their overall objective. However, just because a test is computer-adaptive doesn’t mean it is better. Short assessments, for example, may be computer-adaptive, but if they are defined as “screeners” or “standards-proficiency” tests, they will not be able to diagnose a student in a particular area or subject with greater accuracy.
For over 20 years, Let's Go Learn has been offering best-in-class, personalized assessments in reading and math. See for yourself!
Assessment scores indicate the type of test
Look at the end scoring of an assessment to judge what it does.
Examples:
67% ← Percentile scores rank students relative to others. This is NOT diagnostic. Relative scores don’t tell a teacher what to do next with a student.
780 ← Scaled scores provide a ranking within a subject (such as reading, numbers and operations, total math, etc.) that is typically a summative measure. These are NOT diagnostic.
340 Lexile ← This is a readability score. It reveals the level at which a student can read and answer questions correctly. This is not diagnostic because it doesn’t tell you why.
3.5 gls ← If this is assigned to a broad category such as “Math” or “Reading,” it is NOT diagnostic. It doesn’t tell you why a student is at this level. However, if tied to a specific sub-test of a subject such as “Phonics,” “Place Value,” “High-Frequency Words,” or “Multiplication of Whole Numbers,” then it is diagnostic, since each grade-level score will be linked to specific instructional content. For example, at mid-third grade, a teacher would typically teach “thousands, ten-thousands, and hundred-thousands place value” within the math sub-test of Place Value. Grade-level scores are typically criterion-referenced and tied to externally fixed skills or concepts.
Early pioneers of computer-adaptive testing in education
The DORA assessment by Let’s Go Learn launched in September 2001. It was arguably the first true computer-adaptive assessment that was also diagnostic. It operated over the Internet using a 28.8k modem. To put this in perspective, today we measure our high-bandwidth Internet in MB or megabits, which is 1000 times faster. So 28.8k is 0.028 MB.
DORA assessed students in reading. It started with decoding and examined high-frequency words. If a student’s High-Frequency Word grade-level score was low, it then transitioned to a lower level for Word Recognition and then into Phonics. Ultimately, based on earlier scores, a comprehension passage level was selected. In this manner, even students who struggle with reading are not overwhelmed by the DORA assessment. DORA uses computer-adaptive technology to reduce test time and decrease student frustration. This leads to a better and more accurate diagnostic assessment.
In 2005, the DOMA Pre-Algebra and DOMA Algebra assessments were launched by Let’s Go Learn. And in 2010, the ADAM assessment was also launched by Let’s Go Learn in the area of K-7/8 foundational mathematics. These assessments are all highly diagnostic, using computer-adaptive technologies to reduce test-taking time and student frustration and improve diagnostic accuracy.