Please use this identifier to cite or link to this item:
Authors: Pierre Foy
Liqun Yin
Issue Date: Dec-2015
Publisher: 教育研究與評鑑中心
Center for Educational Research and Evaluation
Abstract: Large-scale assessments in education, such as IEA’s TIMSS and PIRLS, rely on sophisticated assessment instruments, elaborate sample designs, and leading-edge item response theory to meet their analytical objectives. Both assessments provide a rich and complex database intended to support and promote secondary analyses. This paper describes the complex international database structures and the statistical methods and procedures for analyzing TIMSS and PIRLS data, with examples provided for illustration. Three essential elements must be considered by researchers in any statistical analysis based on data from the TIMSS and PIRLS international databases. The first is the use of sampling weights in order to produce accurate and reliable results. Second, both assessments apply the Jackknife Repeated Replication technique to derive proper estimates of sampling variance. Finally, with student achievement reported as sets of five plausible values, statistical analyses are performed five times, once for each plausible value, and the final results aggregated across the five plausible values. Researchers and users of the TIMSS and PIRLS international databases who conduct their analyses as described in this paper should feel confident in the results their analyses will yield.
Other Identifiers: 53C061E9-CB6F-DF70-CD8C-002FB89210FB
Appears in Collections:當代教育研究

Files in This Item:
File SizeFormat 
ntnulib_ja_K0101_2304_041.pdf906.91 kBAdobe PDFView/Open

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.