Accessibility of tests for individuals with disabilities within a validity framework

There is a great need to ensure that language tests are accessible to individuals with disabilities. Yet accessibility features can sometimes conflict with the validity of test scores. In some cases the nature of the conflict seems obvious, yet in other cases there is controversy, such as that concerning the use of a “readaloud” accessibility feature on tests of reading. What is needed is a more rigorous approach for reasoning about the validity implications of accessibility features. The approach described in this article seeks to integrate thinking about accessibility, task design, and validity - all in a framework of sharable terminology, concepts, and knowledge representations. We believe that such a framework can allow one to more accurately and quickly identify the validity-related consequences of design changes that are intended to improve accessibility for individuals with disabilities. Such a framework may permit greater inclusion of individuals with disabilities or other sub-populations without invalidating test results.