According to a review of this test, there is no validity data for CAST, and "users must rely on their judgment of the face validity of the items" (Daniel, 2001). If the skills that a test measures do not align with what is being taught and learned in the adult basic educational system, caution should be exercised if the test is being used to determine placement.
Reliability refers to the degree to which test scores are consistent over time and different test situations. The length of a test is one of the factors that influences reliability; tests with larger numbers of test items produce higher reliabilities. One of the respondents who used the Brigance Diagnostic Reading Inventory of Essential Skills stated that "as an initial placement tool I don't feel that it gave a very accurate assessment. The sampling of questions was too small." Indeed, in the past, the Brigance has been criticized by a review in the Mental Measurements Yearbook for its lack of psychomteric data on reliability (Matuszek, 1985). The reliability of CABS is also questionable because the directions do not specify how many demonstrations should be completed in each content domain. This can result in inadequate sampling. In a worst-case scenario, an educator might engage the learner in only a few demonstrations to determine placement. One of the respondents stated that some of the CABS demonstrations may not "depict the actual LBS (Literacy and Basic Skills) level."
Fifty-one respondents (13 percent) who administer assessments to determine placement are using instruments that have been standardized and normed on youth, rather than on an adult population. For example, the Canadian Achievement Test (CAT), the Canadian Test of Basic Skills (CTBS), and the Gates-MacGinitie Reading Tests were normed on students in the K-to-12 public school system. In a nutshell, tests normed on children and adolescents are not an accurate measure of adults ' literacy and numeracy skills. Despite their obvious limitations, these tests continue to be used for placement. In fact, only a small percentage of the respondents (15 percent) who use tools for determining placement ranked the norming population as an important factor in test selection.