1. Effects of Item Exposure for Conventional Examinations in a Continuous Testing Environment.
- Author
-
Hertz, Norman R. and Chinn, Roberta N.
- Abstract
This study explored the effect of item exposure on two conventional examinations administered as computer-based tests. A principal hypothesis was that item exposure would have little or no effect on average difficulty of the items over the course of an administrative cycle. This hypothesis was tested by exploring conventional item statistics and Rasch estimates of ability and difficulty of four separate groups of candidates who took a licensing examination in a continuous testing environment over the course of a 6-month administration cycle. Subjects were 1,001 candidates for a state license in clinical social work (LCSW) and 1,660 candidates for licensure as a marriage and family therapist (MFT). Taken together, results suggest that even if candidates took the examinations later in the cycle, there was no clear indication that information obtained form candidates who took the test early in the cycle improved performance. The most important implication for small testing programs is that they can enjoy the benefits of computer administration without having large item pools and candidate populations or using computer adaptive delivery systems. (Contains 4 figures, 2 tables, and 22 references.) (SLD)
- Published
- 2003