Open AccessCONTROLLING ITEM EXPOSURE RATES IN A REALISTIC ADAPTIVE TESTING PARADIGM
Author(s) -
Stocking Martha L.
Publication year - 1993
Publication title -
ets research report series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.235
H-Index - 5
ISSN - 2330-8516
DOI - 10.1002/j.2333-8504.1993.tb01513.x
Subject(s) - computerized adaptive testing , computer science , context (archaeology) , probabilistic logic , reuse , test (biology) , machine learning , artificial intelligence , statistics , mathematics , psychometrics , paleontology , ecology , biology
ABSTRACT In the context of paper and pencil testing, the frequency of the exposure of items is usually controlled through policies that regulate both the reuse of test forms and the frequency with which a candidate may retake the test. In the context of computerized adaptive testing, where items pools are large and expensive to produce and testing can be on a continual basis, new strategies are required. This paper discusses the popular randomization strategy for controlling item security and a less well known probabilistic approach due to Sympson and Hetter. Extensions are developed to the Sympson and Hetter approach to make it more relevant for modern adaptive testing. Examples are given of the application of the randomization approach and the extended Sympson and Hetter approach.