elementary and secondary school classrooms, undergraduate and graduate college classes, and participants in a local summer institute for artistically talented students. These groups were administered test items and studied in terms of their reactions as well as the test results. In this process, the CDAT was adapted to its present form as four drawing tasks and a complex set of scoring criteria were adapted as 13 criteria scored on a scale of 1-5 points per item.
In 1989, results of an experimental use of Clark's Drawing Abilities Test were reported and analyzed; demonstrating that the test was highly effective and reliable in both screening and identifying students who are talented in the visual arts (Clark 1989). Those early results, for instance, showed that test scores consistently correlated with teacher's ratings of student success in various art classes in a summer institute for artistically talented students at a greater than .01 significance level and the test was shown to be both reliable and valid (Clark, 1989). Split-half reliability also was calculated using a Guttman Split Half formula and yielded a correlation at the .01 level of significance.
These findings obviously were sufficiently high and accurate enough to encourage further explorations of use of the test with additional subjects. In the ensuing years, the test was used to assign students in the Indiana University Summer Art Institute to various levels of classes. Each applicant, his or her parents or guardians, and an art teacher, classroom teacher, or principal at his or her school filed application or nomination forms; each applicant also filed a letter of application explaining why he or she wanted to attend the Institute. Clark's Drawing Abilities Test was administered to all those who were selected to attend the Institute on the night students arrived, before they attended any classes.
Each year, teachers at the Institute were asked to rank order their students in categories of superior, above average, average, and below average in regard to successful classroom performance. This measure was