How Standards Are Used in Language Assessment
Language proficiency assessment can seem complex or even mysterious – a black box that just spits out a score. However, it isn’t that complicated once you get familiarized with a few basic concepts. I will be describing some of these basic concepts in the next few blog entries. Here is the first.
Standards-Based (or Criterion-based) vs Norm-Based (or Norm-Referenced)
Today, most proficiency assessments are standards-based, meaning that the test measures what test-takers can do against a set of fixed standards. Avant STAMP (STAndards-based Measure of Proficiency) assessments use standards that are aligned with the ACTFL Proficiency Guidelines. The world language standards for virtually all of the states in the US describe proficiency levels based on the ACTFL Guidelines. Often the standards are operationalized by testing organizations into benchmarks or rubrics. Avant has done this by adding some granularity to these standards to make it possible to use them to score learner responses to test questions. Here are the Benchmarks and Rubrics for Avant STAMP. Other standards-based language tests use a rubric as well to identify specific standards. Here for example, are the rubrics for the AP Spanish Language and Culture assessments.
Norm-based tests measure how students perform against other learners taking the same test. The SAT and ACT are examples of norm referenced tests. With the large number of test takers in these tests and the years of experience of the testing companies who develop and deliver these tests, the results are quite consistent and very accurate measures of test taker skills.
In the classroom both standards-based and norm-referenced tests are used. Generally, however, teacher developed tests in class are norm-referenced and often graded on a curve. This provides a distribution of grades that can be evenly spread across the class. What this approach can often result in, however, is a distortion of the actual level of skill of some learners. For example, if all the test takers scored between 60 and 75, some students would likely receive A’s even though they hadn’t demonstrated a high level of mastery.
The advantage of standards-based tests over norm-based tests in the classroom is that learners receive a score that measures their skills in absolute terms – can they do what the standards describe or not. In a standards-based classroom test, if learners all scored above 90 on a test based on a set of agreed upon goals linked to standards, all the students would receive an A. if they all scored in the 60’s they would all receive a D. Using a truly standards-based grading system in the classroom may seem uncomfortable, but we believe it is the best way to encourage effective language learning. See the Tell Projects materials on effective language teaching criteria
Avant delivered the first online, adaptive, standards-based language proficiency test in the world. We have always believed that tests should bring out the best in learners by showing what they CAN do, not what they can’t. We believe that this approach benefits both the learner and educator in powerful ways.