Paper Summary
Share...

Direct link:

Development of an Online Scale Linking Design Using Random Item Selection and Content Balancing Procedures

Fri, April 12, 4:55 to 6:25pm, Pennsylvania Convention Center, Floor: Level 200, Exhibit Hall B

Abstract

For decades, the balanced incomplete block (BIB) and nonequivalent groups with an anchor test (NEAT) have been used for scale-linking design in large-scale assessments. Both designs can assemble multiple booklets and build a common scale, but they were mainly developed for paper-and-pencil tests. Since computerized testing has become popular nowadays, whether the procedures in this field can help implement online scale linking design is of interest. This study proposes an innovative design, (online scale-linking design, OSL) that applies random item selection and content balancing procedures, which were critical components of computerized adaptive testing. The performance of OSL was compared to BIB and NEAT through a series of simulation studies. The results and practical implications of OSL are discussed.

Authors