An Illustration of the Epanechnikov and Adaptive Continuization Methods in Kernel Equating
Gaussian kernel continuization of the score distributions has been the standard choice in kernel equating. In this paper we illustrate the use of both the Epanechnikov and adaptive kernels in the actual equating step using the R package SNSequate (González, J Stat Softw 59(7):1–30, 2014). The two new kernel equating methods are compared with each other and with the Gaussian, logistic, and uniform kernels.
KeywordsKernel equating Epanechnikov kernel Adaptive kernel Continuization
The first author acknowledges partial support of grant Fondecyt 1150233. The authors thank Ms. Laura Frisby, ACT, for editorial help.
- N.J. Dorans, M.D. Feigenbaum, Equating issues engendered by changes to the SAT and PSAT/NMSQT. Technical issues related to the introduction of the new SAT and PSAT/NMSQT (1994), pp. 91–122Google Scholar
- P.W. Holland, D.T. Thayer, The kernel method of equating score distributions. Technical report, Educational Testing Service, Princeton, NJ, 1989Google Scholar
- Y.H. Lee, A.A. von Davier, Equating through alternative kernels, in Statistical Models for Test Equating, Scaling, and Linking, vol. 1, ed. by A.A. von Davier (Springer, New York, 2011), pp. 159–173Google Scholar