ERIC Number: ED206717
Record Type: Non-Journal
Publication Date: 1980
Reference Count: N/A
A Comparison of Two Methods of Interactive Testing. Final Report.
Nicewander, W. Alan; And Others
Two methods of interactive, computer-assisted testing methods for multiple-choice items were compared with each other and with conventional multiple-choice tests. The interactive testing methods compared were tailored testing and the respond-until-correct (RUC) item response method. In tailored testing, examinee ability is successively estimated during the course of test administration, and optimal items, based on the most current ability estimate, are administered to maximize the test's information for the examinee's ability level. The RUC response mode requires examinees to respond repeatedly to muliple-choice items until the correct answer is discovered. RUC tests consist of a fixed set of items for all examinees. The comparisons among testing methods--tailored, RUC and conventional--were analytical, rather than empirical, and the primary index for the comparisons was the average test information. The analytic comparisons of the testing methods indicated that in certain situations, RUC tests can provide greater average information than tailored tests, and furthermore, there are circumstances in which conventional tests provide average information that is essentially equal to that of the tailored tests. However, in terms of providing adequate information across nearly all ability levels, tailored tests have no competitor. (Author/GK)
Publication Type: Reports - Research
Education Level: N/A
Sponsor: National Inst. of Education (DHEW), Washington, DC.
Authoring Institution: Oklahoma Univ., Norman. Dept. of Psychology.