One respondent, who coordinates adult basic education programs for a school district that uses CAAT, expressed these concerns with older tests:
In addition to these three points, older tests are usually based on outdated reading theories. CAAT, for example, is based on the text-based model of reading, rather than on a social constructivist or new literacies model. In fact, in spite of changes in reading theories, there has been little change in either the basic content or the format of standardized assessments since the 1930s (Bainbridge & Malicky, 2004).
Due to the diversity of students attending adult basic education programs, instructors want to use assessment tools that are "fair" and without "bias." The respondents stressed that all tools need to be geographically and culturally sensitive, with respect to First Nations populations and visible minority groups who have taken English as a Second Language. Many students reside in remote areas, which means that they experience test items that are geographically biased. For example, consider a CABS test item that asks questions about paying parking tickets. Would this be relevant to students who live in isolated hamlets in the territories or in rural areas where parking tickets are nonexistent? However, according to Johnston (1998), bias is always embedded in assessments. Johnston writes that "because of the cultural nature of literacy, it is not possible to create an unbiased literacy test; tests always privilege particular forms of language and experience" (p. 98). Despite Johnston's claim, test developers are not off the hook when it comes to developing culturally sensitive assessment tools. Test developers have a responsibility to reduce bias in tests by analyzing item data separately for different populations and then identifying and discarding items that appear to be biased.