The Influence of Item Calibration Error
on Variable-Length Computerized Adaptive Testing
The goal of this paper examine the effects of item calibration error on important testing outcomes in VL-CAT. The simulation study was conducted（ Two IRT models: 2PLM or 3PLM;Two termination criteria: CSE or ACI; and Four calibration sample sizes: N = ∞, 2500, 1000, or 500. ）.The r esults confirm s that capitalization on chance occurs in VL-CAT and has complex effects on test length, ability estimation, and classification accuracy ： The test length under the CSE termination rule was sensitive to the magnitude of item calibration erro r，i n contrast, test length under the ACI termination rule was clearly robust to the magnitude of calibration error .
Some concepts:
VL-CAT ： A test that is adaptive in terms of both item selection and length is commonly referred to as a variable-length CAT, or VL-CAT.
The calibration error ： refers to the magnitude of sampling variability in item parameter estimates as determined by the size of the calibration sample for a given method of calibration and distribution of the latent trait.
Two different stopping criteria:
CSE criterio n：
the asymptotic standard error of based on test information (i.e., treating item parameter estimates as the true values) was compared to a threshold; if the SE fell below the threshold, the test ended
ACI criterion ： the 95% asymptotic confidence interval for was compared to
I have to say it is a long and difficult paper.