Selecting Parameters in Cubic Spline PostSmoothing Using Cross-Validation

• In Event: Scaling, Linking, and Equating

Abstract

The purpose of this study is to propose a new approach to selecting a smoothing parameter in cubic-spline functions in equating using a cross-validation technique. Results from the proposed method are evaluated in comparison to various levels of parameters typically used in cubic-spline postsmoothing and log-linear presmoothing through simulation studies.

Authors

• Stella Kim, University of North Carolina at Charlotte

• Won-Chan Lee, UNIVERSITY OF IOWA

• Hwanggyu Lim, Inha University

• yeonwho kim, Seoul National University