組合研討會

主講者: Dr. Yi-Hsuan Lee (Educational Testing Service, USA)
講題: Statistical Applications of Logistic Regression Model in Educational Testing
時間: 2009-02-16 (Mon.)  14:00 - 15:00
地點:
Abstract:  Differential item functioning (DIF) is an important issue in large scale standardized testing. Its presence could seriously affect the validity of inferences drawn from a test. In computerized adaptive testing (CAT), a DIF item may be more consequential and more detrimental because fewer items are administered in a CAT than in a traditional pencil-and-paper test, and because the remaining sequence of items presented to examinees depends in part on their responses to the DIF item. In this talk, we will illustrate the use of logistic regression model in DIF detection. Two bias correction procedures using asymptotic expansion and conditional score theory are developed and incorporated to adjust for the covariate measurement errors resulting from the use of estimated latent traits. Extensive simulation studies show that the resulting methods perform well in terms of Type-I error rate control, accuracy in estimating DIF, and power for detecting both unidirectional and crossing DIF.
  || Close window ||