Abstract
The present study developed the item characteristic curve method and testlet characteristic curve method for linking the testlet-based tests under testlet response model. Then the effectiveness of the characteristic curve methods and the concurrent calibration method as regarding to testlet response model and IRT were compared under different conditions in this simulation study. The results indicated that the newly developed methods were shown to perform well in the recovery of the generating item parameters. Larger testlet variances led to greater equating errors. The study contributes to a better understanding of the effectiveness of different equating methods for testlet-based tests and helps clarify the influences of the local dependences, testlet variances, and number of common testlets on equating results.
Original language | English |
---|---|
Publication status | Published - 2011 |