Open AccessA SURVEY OF EQUATING METHODS BASED ON ITEM CHARACTERISTIC CURVE THEORY
Author(s) -
Lord Frederic M.
Publication year - 1975
Publication title -
ets research bulletin series
Language(s) - English
Resource type - Journals
eISSN - 2333-8504
pISSN - 0424-6144
DOI - 10.1002/j.2333-8504.1975.tb01052.x
Subject(s) - equating , variety (cybernetics) , stability (learning theory) , econometrics , nonlinear system , mathematics , statistics , item response theory , computer science , machine learning , physics , psychometrics , quantum mechanics , rasch model
A variety of equating methods based on ice theory are surveyed and compared with nonlinear conventional methods. In many cases, the ice model replaces a nonlinear equating problem by a linear one. This helps the stability of the equating near the extremes of the data. Strictly speaking, ice methods are the only methods appropriate when the tests to be equated are given (with an anchor test) to groups differing in ability. It seems likely that ice methods can reduce drift in equating chains involving many links. Research into the practical value of preequating is recommended.