The Kernel Method of Test Equating

Comparisons of the Kernel Equating Method With the Traditional Equating Methods on Praxis™ Data

“…The KE function with large bandwidths is approximately linear, in which case the estimated KE function is mainly based on the means and standard deviations of the continuized distributions. Because the mean and variances of the continuized distribution in the KE exactly match those of the presmoothed discrete distribution (von Davier et al, 2004b), it is not surprising that the KE function with large bandwidth is very close to the linear equating function in the EG design. For similar reasons, the KE with large bandwidth should produce very similar results to its analogue, the PSE-linear method, in the NEAT design using the PSE method.…”

“…The traditional observed-equating methods for this design include linear equating and equipercentile equating (Kolen & Brennan, 2004 (Kolen & Brennan, 2004;von Davier et al, 2004b). KE methods include approximations to the frequency estimation and the PSE-linear methods.…”

“…The kernel method of test equating or kernel equating (KE) was introduced by Holland and Thayer (1989) and extensively developed by von Davier, Holland, and Thayer (2004b).…”

“…The mathematical and statistical details of KE are reported by von Davier et al (2004b) and are briefly described in Appendix A.…”

“…

ETS and the ETS logo are registered trademarks of Educational Testing Service (ETS).

AbstractKernel equating (KE) is a new approach to observed-score equating and is described in detail in von Davier, Holland, and Thayer (2004b). Over the past months, several evaluation studies of KE have been designed and carried out.

…”