Paper Summary

Investigation of Fixed-Item Parameter Calibration for Mixed Format Tests

Sun, April 15, 12:25 to 1:55pm, Vancouver Convention Centre, Floor: First Level, East Ballroom B

Abstract

IRT linking for mixed format test has become one of the major psychometric issues recently. Among various linking alternatives, the fixed item parameter calibration (FIPC) method is technically preferred when two test forms have dissimilar formats. The current study aims to investigate the performance of an alternative FIPC method which updates the prior distribution in the calibration cycles when linking mixed format tests. Under a common item nonequivalent groups design, factors such as format effect and group differences will be considered. The results of the alternative FIPC will be compared with those of other linking methods. The findings will inform practitioners about the appropriateness of the procedures of implementing FIPC in mixed format test linking.

Authors