Cluster-branching methodology for adaptive testing and the development of the Adaptive Category Test. |
| |
Authors: | Laatsch, Linda Choca, James |
| |
Abstract: | Adaptive testing involves the adjustment of a set of test items, in accordance with an individual's characteristics, to minimize items that do not yield useful information. The best known methodology used to develop adaptive tests, item response theory (IRT), cannot be used with most psychological instruments. The authors propose using cluster analysis to develop a branching logic that would allow the adaptive administration of such instruments. The proposed methodology is described in detail and is used to develop an adaptive version of the Halstead Category Test (W. Halstead & P. Settlage, 1943) from archival data. Real-data simulations show the Adaptive Category Test to yield scores that are not significantly different from the scores actually obtained on the original version of the test. (PsycINFO Database Record (c) 2010 APA, all rights reserved) |
| |
Keywords: | |
|
|